New Trends in Computational Vision and Bio-inspired Computing

In this chapter, a new 3-dimensional real and discrete chaotic transformation technique is introduced in order to encrypt the information. The encryption algorithm involves two stages per round: Transformation and Substitution. In the transformation phase, the host image is subjected for three dimensional multi linear transformation (3D-MLT) along with the secrete image. The transformed image is subjected for substitution using S-box created by secrete keys. The cipher image is subjected to security tests which involves different statistical and analytical performance analysis. The obtained results prove high degree of security compared to existing techniques.

Object recognition is one of the essential Computer Vision techniques. The success of object recognition lies in identifying features that strongly represent the object of interest. The manuscript comes up with a hybrid feature descriptor that combines the properties of HOG, ORB and BRISK feature descriptors. Linear SVM is used to classify the feature vectors of the object of interest and other objects in the scene. Occlusion, Orientation and Scaling are some of the limitations in existing approach. From the experimental analysis, we infer that the proposed framework handles partial occlusion and is invariant to scaling and rotation. The framework has been tested with a manually built data library and the classification accuracy of the proposed framework is 0.91, whereas the standalone performance of the HOG, ORB and BRISK are 0.85, 0.87, and 0.89 respectively.

The search for the digital image from the repository is challenging since the volume of the image created and consumed is growing exponentially with respect to time. This makes the image retrieval an ongoing research problem. Rather than relying on metadata, analyzing the content of the image is proven to be a successful solution for retrieval. Since manually annotating the growing images is considered to be impossible. Thus the need for the hour is a framework that is capable of retrieving the similar images with less time complexity. In this paper, an image retrieval framework has been proposed to retrieve similar images from the repository. The motivation of the work is to leverage the smart nature of building, in such a way that when an asset inside the building is captured and given as query image, the user of the building should be provided with all the relevant assets inside the building. The proposed framework is built with color, shape and edge retrieval system that works in cascade approach. Since the framework works as a cascade, it is observed that the results get fine-tuned at every layer of the cascade thus increasing the precision. The novelty of the work lies in the third layer where the XOR operation is performed to check the magnitude of dissimilarity between the query and database images. Based on the above dissimilarity, the threshold has been fixed to differentiate the image of interest from other images in the repository. The performance of the proposed approach is evaluated with the manually built dataset. On evaluating the performance, it is inferred that the precision of the proposed framework is 99%.

Geospatial data is processed using a geographical information system, geospatial data is mainly of two kind’s vector and raster data. Vector data largely includes point data and arcs whereas raster data is the used to represent surfaces. This data is processed and the insights are used for demographical analysis, geological studies, map route optimizations, creating hydrological models among other uses making GIS important for both academia and industry. The suggested system is aimed to further the cause of data standardization as well as web-based GIS tools. The system explores the uses of open source application in order to provide the base for a system that enable the system to provide the GIS functionalities over a portal in form of WPS. The proposed system intends to provide services that enable the user’s access standardized methods of publishing geospatial data and use various services.

India has huge potential to generate power from renewable sources, but still most of the regions are denied to use renewable sources due to its poor reliability. This problem can be resolved by micro grid, integrating several renewable sources, as a main source for that particular region. A smart condition monitoring system and power management is essential to maintain the system reliability. Condition monitoring is the process that collects the various electrical parameters and analyse the performance of the system or its components such that remedial action may be planned in a smart manner. In this planned work, an Internet of Things (IoT) based condition monitoring system is developed for standalone micro grid. This standalone micro grid consists of solar, hybrid generator and battery. In this work, an IoT based Web Architecture will be developed with intelligent controller to take optimal decisions for efficient power management of micro grid. A dedicated web page and android application for remotely accessing the data will be developed for continuous access of production and usage of power, condition of loads, availability of battery power, voltage and load power factor. The status of the system can be monitored online through this smart network. The fuzzy logic controller will be developed to perform the control action on micro grid. The single fuzzy based controller will perform both load and source power management to improve the reliability. This controller facilitates the management of distribution of loads by properly scheduling loads and source management by optimally utilizing the sources. The control actions are based on the total power both from generating units and battery storage, the load requirement and the battery status. The designed controller will facilitate as a communication channel between the sources and load to avoid the deep discharging of battery at heavy load condition and prevents the overcharging of battery at light load condition. The proposed system will work efficiently for accessing the data without any specific control room with computers.

Cloud computing infrastructure is combination of software and hardware resources for performing the efficient computing. In this context a number of contributions are established for optimizing their performance more. The resource scheduling and the workload balancing is the similar directional effort for optimizing the computational performance of the cloud infrastructures. The proposed work is intended to measure the performance of cloud scheduling approaches that are claimed to optimize the performance of the cloud computing. Therefore two popular and frequently used scheduling approaches i.e. round robin and first come first serve techniques are implemented with the help of cloudSim simulator. The round robin technique allocates a fixed amount of time for all the resources to a given job therefore this concept works as the time shared manner. Similarly the FCFS technique allocates jobs according to their appearance or sequence therefore that technique is functions according to the space shared manner. Additionally the performance of both the approaches are measured and compared. In order to compare the performance of both the techniques the average processing time, average processing cost, average waiting time and the CPU utilization is computed using the simulation trace. According to the obtained performance the round robin technique found much efficient in all the parameters. Therefore it is acceptable for future extension of the proposed work.

The rapid development in multimedia and imaging technology, the numbers of images uploaded and shared on the internet have increased. It leads to develop the highly effective image retrieval system to satisfy the human needs. The content-text image retrieval (CTIR) system which retrieves the image based on the high level features such as tags which are not sufficient to describe the user’s low level perception for images. Therefore reducing this semantic gap problem of image retrieval is challenging task. Some of the most important notions in image retrieval are keywords, terms or concepts. Terms are used by humans to describe their information need and it also used by system as a way to represent images. Here in this paper different types of features their advantage and disadvantages are described.

Amid late years, the field of emotion classification of Speech signals has been picking up a huge consideration and a few systems have been built by various scientists for acknowledgment of human feelings in talked articulations with their local voice correspondence. This paper depicts a test which exhibits the acknowledgment of simulated human emotions from an artificial Marathi emotional speech database. The database comprises of the information tests that were gathered from five Marathi emotion pictures (Actors and Actress voice) recreated the feelings creating the Marathi expressions which could be utilized as a part of regular correspondence and are interpretable in every single connected feeling. The discourse tests were recognized by the different circumstances from the films. The information tests were classified in five fundamental classifications that are Happy, Sad, Anger, Afraid and Surprise. Our point is to explore three fundamental feelings. The Study explores Linear Predictive Coding (LPC) model for speech signal production based on the supposition that the speech signal is produced by a very precise model for emotion recognition.

Due to the gargantuan development of technology in medicine, transmission of medical data securely through multimedia has become possible. The foremost objective of transmission of secret medical data over the internet is, it should be difficult to access the secret data information for the attackers. In this research patient’s secret medical data is concealed in the bio-medical signal like EEG/ECG/PPG. The information that is transmitted generally contains biomedical-signals and patient data. Major apprehensions comprises of authenticity and privacy of the data being transmitted. A secret key is used which is flanked by the receiver and sender involved in the system and unknown to others. A novel steganography technique is introduced in this paper that guarantees (1) security of private information applying a key and (2) uniqueness of the bio-medical-signals. To exploit embedding, Fast-Walsh-Hadamard Transform is applied for conversion of the signals into a set of coefficients. The recommended procedure uses sensors such as pulse sensor, BP sensor, DHT-11 sensor to monitor the patient in POC system. To accomplish least distortion, least significant coefficients bit is considered. The impact of the algorithm on the bio-medical signal is less and recovery of the signal at the transmitting side can be done with a smaller amount of distortion. Proposed technique is implemented in POC.

In the recent days mainly focus on cloud security in different areas. Now a day’s many businesses cloud computing technology is growing at higher level but the main issues is maintaining the security and privacy of confidential data. Cloud stores various types of data for example data sheets, digital media object and it is very important to maintain data security and confidentiality while accessing in cloud environment. Data sharing on cloud need extra approach to secure path for data confidentiality. While sharing resources at distributed level has severe implications when sensitive or privacy-relevant data is concerned. This paper aims to ensure the data security in public or untrusted cloud servers by constructing strong data access control frame work. We propose an encryption-based schemes which provide on a public cloud server. We apply the scheme to provide data access control.

Agriculture is a vast field with variety of species with different attributes which are specific for species. First step in all identification or recognition is image pre processing. In recent world, this recognition has become active in the field of medicine plants and flora species. Leaf features are extracted and based on those features, the neural network is trained to recognize the leaf. In this paper Artificial Neural Network (ANN) is used for leaf classification and Graphical User Interface (GUI) is developed to identify the leaf automatically. Six types of leaf images used for identification and achieved 90.9% accuracy. As well as result is evaluated with other performance metrics like Recall, Precision and F1 Score.

Cloud computing is an integral part of the IT architecture of any modern-day business enterprise, where the user can access the cloud pool of software resources and business data on a demand basis. The resources and data analysis in the cloud where security is a paramount concern as the issue of cyber security gains momentum day by day. Here we report the RSA and GNFS algorithm based integrated approach to generate cyber security key for the protection of data and transfer of the same in the cloud. Through this proposed model, the data from the database can be retrieved efficiently in encrypted form with best execution time and speed. We believe this model of integrated cryptographic key system generated through RSA and GFNS algorithms has wide implications on cloud based cyber-security applications.

The missing data and skewed values in data perhaps imply the wrong interpretation in the analysis. The accurate values generated from the sensor necessary accumulate into the Internet of Things (IoT) based environment either at the web server or the Cloud-based system or combination of both in the case of the Cloudburst. It should handle systematic programming either thread enable or simultaneous processing. This work shares the lessons learned in the data collection and statistical analysis in the context of a Generalized Physical Location Monitoring system, for further applying the machine learning techniques.

Applying the popularly known technologies to solve real world problems are common practice among student researcher community, as it brings deeper understanding of the underlying technology for its further study and improvement. This paper aims at applying the Genetic Algorithm (GA) to solve the course allocation problem of educational institutions. The course allocation problem comprises of p number choices given by n numbers of students for m number of courses. Assigning the maximum number of students with their first or second choice of their courses is a cumbersome task. It is a typical optimization problem, which can be solved in ease by the Evolutionary Algorithms (EAs) such as GA. This paper proposes an automated system which uses GA (with five different crossover operators and three different mutation operators) to solve the course allocation system. A comparative study on the results obtained for different crossover operators is performed. The obtained results are verified with a real time data set collected from our University and validated the superiority of the proposed system.

Now a days detecting anomalies has become wide domain and it is considered as one of the main problem in many applications. From the standard, normal, or expected behavior something that varies from these are called as anomalies. Anomaly detection is identifying or finding anomalies from various applications. There are some kind of problems that arises in many applications such as structural defects, frauds or errors, the anomalous items have the potential of getting converted into such problems. Many techniques or methods are developed and are used for detecting anomalies. In this paper implementation using K-means, Support vector machine techniques for detecting anomalies in credit card transaction dataset have been described and accuracy is calculated to determine which technique is efficient in detecting anomalies.

The recent cloud computing sector has introduced multiple cloud service provider, which provides cloud storage to the consumers through a wireless or device less method of storing data. However, it has been difficult for the consumers to choose providers who are more secured and cheaper at the same time. Recently, there have been lot of security issues found in cloud storages by different consumers and their respective providers. However, calculating the security issues and security level for any provider plays a crucial role for the consumer to select the cloud service. It has been found that data breaching is the most common security issue faced by cloud storages which has swayed to the concept of cloud computing technology. This paper suggests a method to detect data breaching in any cloud service which can detect the breach with the help of a decision making tree. This decision making tree will have many constraints which include from the password to the location of access. Once a breaching is detected it will take further steps with terminating the connection. The tree based method is exploited. It can be measure and find all the practicable attack structure. This is the main concept of data breach on any particular cloud service and will help in reducing the growing rate of this attack.

The paper deals with the break-ins that the modern-day biometric verification systems like Automatic Border Control, facial unlocking scheme as in many smartphones, and other photo-ID documents generation and verification systems face. One of the most prominent attacks is the facial morphing attack, wherein the system is fooled by asking it to do facial recognition and matching of a person with a photo which is morphed and has features of two persons overlapped. The proposed framework gives a deep insight into the concept of image morphing and the way to analyze the features and allocate them priorities. The system tries to integrate all the features of image that could possibly have an influence on the face image if morphed with another face image. The paper also presents an account of the advantages and disadvantages as well as the intuition of various approaches of face image morphing detection, especially we take into account the deep learning models that have been used previously and try to tune in the parameters and analyze their complexity in order to try various methods to reduce the overfitting of such models.

Hyperbilirubinemia is a condition in which the bilirubin levels rise above the normal mark and this causes discoloration of the skin and eyes among other symptoms. This paper proposes a framework which determines the severity of this condition using non-invasive methods and produces relevant results after proper analysis. The technique involved uses a specially designed camera module for capturing images of the sclera region of the eye and Digital Image Processing for analyzing the obtained raw data. The processes include Bilateral and Gaussian filtering of the raw images, detection of iris of the eye, masking of every color except shades of yellow and eventually, pixel by pixel analysis of the resulting image. The whole framework is designed in a way that it is capable of restricting unnecessary variations in raw data due to different working conditions while being cost-effective and reliable simultaneously.

Cloud computing is a prototype for usage-based network. It is an Internet-based computing in which large groups of remote servers are networked so as to allow sharing of data-processing tasks, centralized data storage, and online access to computer services or resources. Data Center refers to the hardware that stores data within an organization’s local network. They are typically run by an in-house IT department. Cloud computing allows users to access secure and scalable networks of Data Centers and enables availability of virtually housed data, cloud-native and enterprise applications. The challenging problems in cloud data Center is the management of the load of different reconfigurable virtual machines. A mechanism for efficient resource management will be very significant to suite the need. The data Centers comprises of thousands of servers to provide services. The cost of maintaining this cloud data canters is extremely high. This paper focuses on the review of optimizing task load of different zone of data Center and users in the cloud environment.

The human recognition is achieved easier and cheaper and the single modality employed for the recognition faces a lot of challenges due to the environmental factors. This paper proposes a multimodal recognition system based on the Multi-Support Vector Neural Network (Multi-SVNN). The algorithm proposed is Glowworm Penguin Search Optimization Algorithm (GwPeSOA), which is the modification of the Glowworm Optimization Algorithm (GOA) with the Penguin Search Optimization Algorithm (PeSOA). The proposed method employs ear and the palm vein modalities and the features of the ear image is obtained using the proposed BiComp masking method of feature extraction, whereas the features from the palm vein is extracted using the Local Binary Pattern method. The features obtained are applied to the Multi-SVNN classifier to recognize with good accuracy and the proposed BiComp Mask offers the robust features for the extraction. The experimentation using the proposed method attained a better accuracy, specificity, and sensitivity.

In recent years, government has provided a lot of support for development of agriculture. Farmers also have a lot of interest in adopting new schemes and technologies, where they can expect to increase their income and boost their yields. Decision support systems for Agriculture can help farmers to handle and manage complex problems in crop cultivation and production, utilizing the best available data and knowledge to frame the strategies for better yield and profit.This paper titled “E-Agriculture” presents an application to help farmers to get information to increase returns. The application provides complete information on crop selection, soil constituents, fertilizer selection The Application also provides information on current weather, existing market prices of crops in the nearest areas, direct selling and buying of products/crops etc. E-Agriculture is an application that helps farmers cut the middleman out.

The paper presents a compact and low cost monopole Ultra Wide Band (UWB) antenna for wireless communication having a operating range between 3.1 to 10.6 GHz. The proposed antenna has a 8 sided polygon configuration with dimension of 22 × 32 × 1.6 mm3 and design on a low cost FR-4 dielectric substrate. Design has a trapezoidal shape ground plane to provide better impedance matching. Also it has a trapezoidal shape microstrip line which is used to feed the patch. The simulated results shows a uniform radiation characteristics over entire frequency bands of interest.

Green inventory policy of Green chemical Green warehouses is a very important issue because a Green chemical supply chain is directly related to human’s lives. Variables order quantity (VOQ), economic order quantity (EOQ), Time order quantity (TOQ), Disposal order quantity (DOQ) and Environment order quantity (EOQ) of inventory policies is widely used to manage the Green chemical stock in the present trend. However, effective management of Green chemical inventory is a difficult problem due to characteristics of the Green chemical. We proposed ABC (Artificial bee colony algorithm) based inventory policy for Green warehouses dealers in Green chemical supply chain. We also simulate the designed model to measure a performance of management policies. The proposed artificial bee colony algorithm (ABC) determines an optimal volume of products at the specific order time by using current stocks. Simulation results demonstrate that artificial bee colony algorithm ABC is an efficient method for Green warehouses inventory management and Green chemical supply chain.

Diagnosing of heart disease is one of the important issue and many researchers investigated to develop intelligent medical decision support systems to improve the ability of the physicians. Data classification can often be applied to medical data helps to detect the prevalence of disease. Many tools and algorithms are proposed by researchers to develop an effective medical decision system. Feature selection refers to the problem of Selecting relevant features to produce the most predictive outcome is called as feature selection. This paper proposes a new feature selection method based on rough set theory with modified dynamic quick reduct algorithm.

Threat characterization is critical for associations, as it is an imperative move towards execution of data security. Vast majority of the current threat characterizations recorded threats in static courses without connecting risks to information system zones. The aim of this paper is to represent each threat in different areas of the information system the methodology to solve the problem. Data security is habitually represented to different kinds of threats which may cause distinctive types of harms that can prompt to critical monetary losses. Data security problems can go from small losses to entire data framework destruction. The effect of various threats vary extensively: some manipulate the integrity or confidentiality of information while others manipulate the accessibility of a framework. At present, associations are trying to comprehend what are the threats to their data resources are and what are the ways to get the significant intends to combat them which keep on representing a challenge.

Dominating set problem is a famous problem that finds applications in many fields. In this paper, an algorithm based on DNA Computing sticker-based model is proposed for finding dominating set in a given graph. Here, initially solution space containing all possible dominating sets of a graph is constructed and then the solution space is filtered iteratively until the desired solution set is obtained. Along with the operations in the sticker-based model, operations defined for DNA computer sticker-based model ALU is used.

Different evolving technologies surround humans today. Among the various technologies, Assistive Technology has still not established itself firmly because there is an absence of proper integration of this technology with human life. However, in the future, it will become one of the most important and vital phenomena in everyone’s life. Because humans want to make their life easier and longer and these are the reasons for the rapid growth in demand for Assistive Technology. Therefore, improvements in the technology and the way it is applied are essential and, for this reason, there is a requirement of a detailed study of the technology. This paper demonstrates the different milestones achieved in assistive technology by using different techniques to attempt to improve intelligence in assistive systems; and also, it describes the gaps that are still present even after such extensive works and, which are required to be either resolved or bridged. This study is done to understand where the assistive technology is today and in which direction it needs to get directed.

The AFE with 4-channels is depict, lofty impedance squat control utilization of bio-medicinal electrical motion. The defy in acquisition exact accounts of biomedical flag for instance ECG to ponder the person cadaver in research work. This article is to intend Multi-Vt circuit configuration fell by CT modulator. A novel engineering is proposed with four diverse info sign separated from channel to ∑ΔM. In this line, the speaker is little fueled multi-VT Simple starting sense path circuit which expends less control by apply double edge voltage. Sort-I classification 4 channel sign of main mode: 50, 100, 150 and 500 Hz intensified as of AFE be sure to second CT ∑ΔM ADC. Show the Sign NR and Sign NDR as 64 and 61 dB individually, serious consumed the intensity of approx. 11 mW. The circuit configuration was reenacted phantom instrument in a 1.8 V to 0.18 μm typical UMCCMOS process. AFE configuration is faculties and stately measure the recurrence reaction starting <50 to 360 Hz, portray the Sign NR and Sign NDR saw as 63 and 60 dB correspondingly, extreme expended the intensity of <11 mW. “The AFE stately recurrence reaction beginning <50 to >360 Hz, programmable addition on upto 72 dB, input alluded commotion would be an estimation of 3.5 μV in the enhancer scope of data transfer capacity, NEF of 3.”

In this chapter, a modified Tinkerbell chaotic generator based substitution method is proposed for an image encryption algorithm. The correlation between the adjacent elements in the sequence generated by modified Tinkerbell generator is very less and gives high degree of randomness. This sequence is introduced in the substitution stage of encryption process to diffuse the pixel value of cipher image after transformation. The cipher images are subjected to various security analysis and the results obtained are compared with many existing techniques to prove high degree of security.

Spectrum-Map-empowered Opportunistic Routing (SMOR) systems have been created to accomplish dynamic opportunistic links and dependable end-to-end transmission in Cognitive radio ad-hoc networks (CRAHNs). However, only delay has been considered in the mathematical analysis of SMOR in both regular and large-scale networks which results in degraded routing performance. This work examines the transmission delay and the network throughput is evaluated and the relationship between them to develop modified SMOR algorithm by incorporating the concept of acknowledgment (ACK) for each node in the routing link. The Modified SMOR for regular CRAHN utilizes Diffusion approximation based Markov chain modeling and queuing network theory while for large-scale CRAHN utilizes sparse approximation based stochastic geometry and queuing network theory for examining delay and throughput. The Modified SMOR-1 and Modified SMOR-2 are proposed for satisfying the opportunistic routing mechanisms. The experimental results illustrate that the modified SMOR improves the reliability and dynamic routing performance.

It has always been noticed that the excuse given for any work delay are traffic. Urbanites face a lot of trouble due to traffic congestion. These congestions are caused by multiple reasons like traffic violators, traffic collision, inadequate green time, obstacles and many other reasons. Sometimes this can be due to traffic signal going out of sync due to system malfunctioning. Various interesting and engrossing advancements in this sector can be made using the field of IoT. It can bring many visions true to the real world. This paper focuses on reducing the reasons for traffic congestion by tackling them and providing the traffic handlers a technological relief, by giving them a centrally tracking system controlled via a network. For the gathering and processing of the real-time data, IoT is used. This processed information is sent to Cloud for further processes. This stratagem is generalized and can be deployed anywhere efficiently without much manual efforts.

Testing in a cloud environment involves a range of tests to analyze various aspects of the system in various cloud adoption scenarios such as public, private, or hybrid. Performance testing is a key aspect of the cloud testing strategy and it evaluates the application/software for various issues such as speed, stability, and scalability under varying load conditions. The key issues that must be considered in a cloud set-up as a part of the performance testing strategy are collection of statistics on the load, conducting stress test, monitoring the memory, checking for elasticity and scalability. This paper focuses on testing the resource utilization in cloud infrastructure in general and memory utilization in particular. The memory utilization is evaluated in public and private cloud environment with simulation of 10 users to 10,000 users.

Air pollution is a subject widely studied around the world, mainly due to its impacts on human health. The assessment of air pollution impact on health is often carried out using statistical regressions. This work proposes to analyze the performance of Echo State Networks with and without regularization coefficient to predict the number of hospital admissions due to air pollution and climate variables. This procedure can help the governors to take decisions in situations of high concentrations of pollutants. The results showed that the Artificial Neural Networks with regularization coefficient reached best overall results.

Biosensors are invented by Michael Clark and Lyon in 1962. The main motive of invention of Biosensors is to calculate Bio electric potentials present in human body and also to detect pathogens present in human body. Pathogen refers to a virus or fungi which cause a disease in human body. In my work I used biosensor to detect cancer at early stage with the help of Nano Technology. The experimentation involves the Direct current cyclic voltammetry method to detect the cancer at what stage it is. The work involves electro chemical type of Biosensor which involves AMPEROMETRIC, POTENTIOMETRIC and CONDUCTOMETRIC biosensors. The actual work involves Amperometric biosensor whose output is current order of milli amperes to femto amperes. The work involves Optical techniques which yield lithography and is used in detection of cancer. The nano technology components used in this work are Graphene, Nano wires. In this paper I am put forwarding the detection of cancer through optical techniques. This is similar to photo diode principle. When light input increases the current output increases. AMPEROMETRIC BIOSENSOR output is current similar to that of photodiode. The details of cancer detection is done through photodiode principle and fabrication of the device through VLSI techniques.

Nano fabrication of carbon based molecular domain called electro-magnetic nano domain is identified here as an extension of our earlier work on carbon optimized domain in biological systems. Identification suggests that nano fabrication of silicon chip can be replaced by carbon domain where carbon value works out to be both least conductor and conductor. Carbon rich portion attenuate electrons while the carbon domain with defined value highly conductor of current. Demonstrations are here to show how carbon value possesses a dual character of electrical conductance and least conductance. It is hoped that the principle of 31.45% of carbon around individual atoms is individually fabricated for the domain. The results are in agreement to the above value. The carbon fraction distributions in viral protein reveal that it follows an abnormal carbon fraction distribution than that of normal one. It is demonstrated that the carbon fraction distribution plays an important role in deciding the structure and assembly of the proteins. It will be helpful in understanding the molecular structure and nano fabrication of electronic devices. Clue on fabrication technology has been introduced in order to fabricate information technology for tomorrow’s computer and allied nature.

Development in usage of internet for sharing data over internet leads to some risk on privacy, authenticity and confidentiality of information. To overcome the problem of security and authenticity, biometrics and cryptography technology are separately used due few drawbacks in both system, but because of their similar characteristics these two technologies are combined and Bio-crypto system has been designed, to satisfy the needs of user who transmit their data through internet for enhancing the security and authenticity of data and the user. In this proposed work bio-cryptosystem based on fuzzy extractor using face attributes, the user face feature points are extracted and bio code can be generated by bio hashing technique to facilitate the user to access the key that are already stored on the database server. By using face attribute for retrieving the key there is no need for the user to remember the pass code which does not corresponds to the user moreover biometric features cannot be stole and forgotten. Using bio-crypto system the user can be authenticated by enrolment and verification process and encrypt the key along with the own face attribute of the user to make the system more secure and authenticated. The robustness of the data is prevented and there is no cause of bug or intrusion occurs during the interval of data transmission. By this proposed work security, privacy, confidentiality and authenticity can be increased and provide authority only to the valid user to access the data using bio-crypto key.

This paper presents test logic insertion and pattern generation for RTL designs. Test logic is the circuitry that the tool adds to improve the testability of design. Some of the memory elements in the design do not have controllability on clocks and resets. Our proposed work implements scan logic to have controllability and observability on each and every node of the design and adopted EDT technique to generate patterns with improved compression of scan test data and reduction in test time by controlling a large number of internal scan chains using small number of scan channels. Experimental results confirm that the proposed approach can significantly reduce test cost and test time with maximum possible fault and test coverage with ATPG effectiveness.

Various approaches have been proposed for Datacentre to balance the data traffic load recently. Both network topology and the routing approaches can affect the smoothness/quality of the network broadcast. Recently a new system/network architecture called fat-tree becomes one of the trend approach/architecture which is widely used topologies for data centre networks. The routing algorithms such as global load balancing (GLB) and dynamic load balancing (DLB) approaches, are the proposals and both use fat-tree (which are triangular interconnected topologies) topology for the Datacentre using SDN (Software Defined Network) approach. Basically GLB and DLB will be having limited of link information storing and path finding between node edges. So this work proposes an efficient framework dynamic cluster-topology with triangular model and also dynamic sub topology load balancing (DCLB) for fat free to schedule the network flows by taking some IP header path information addition dynamically in data centres networks. The DCLB approach will overcome the existing limitations of both GLB and DLB approaches. This approach not only uses less storage information about the link attributes in the controller, but also choose the path with lowest and neutralized parameters dynamically to have the best routed path and will be updated in the current node in the data centre’s host IP dynamically. Some sub approaches are considered to overcome the load balancing traffic load.

This paper reads on the possibility of using morphological operations to specify out a text from image using the basic mathematical grayscale morphology operations to make the text conspicuous and to use the artificial neural network methods to recognize the text from the image without any loss of texts from image even if the surface of the foreground-background combination is not properly defined. So as to make sure even in cases of unshaped background surfaces, light flares upon the foreground text, blurry or low-quality text scenes, the possibility of recognising the text from the natural scene is high and viable.

Co-existence, that’s the word to describe our relationship with environment and people around us. We are dependent on each other for our daily needs. As we exploit the things around as per our needs and requirements, we are endangering ourselves and the earth we live in. We have been exploiting our world for thousands of years now and we haven’t looked back at the havoc we created. Now in the twenty-first century it’s a major concern climate change, global warming, pollution you name it we got it. People have turned a blind eye and we need to educate the people and let them know of the adverse effects of their own actions. Aim of our project is develop a simple device which is of low cost and ease to use. With this gadget we can quantify Carbon Monoxide (CO), Temperature, Relative Humidity, Particulate Matter 2.5, Noise and UV radiation. Using the concept of Internet of Things, we measure these values using sensors and the data is store in a MySQL database. With help of Geo-tagging we can locate which part in the city where the pollutants are high and send a report to the people of that neighborhood.

In this research paper, the design and development of the algorithms for the automatic detection of glaucoma for healthy cases and unhealthy cases is being presented.

w.r.t. this research article, the design and development of algorithms for the automatic detection of glaucoma for healthy cases and unhealthy cases is being presented.

In this work, we discussed the solutions of viral dynamics in hepatitis B virus infection. Antiviral treatment for patients infected with hepatitis B virus is just somewhat proficient. The field is popular for understanding the associations between the virus, immune responses, short-term and long-term drug efficacy and the general strength of the liver. Furthermore, we investigated the Homotopy Perturbation Method (HPM) to construct the approximate analytical solution of the nonlinear differential equations arise in this model. Analytical results are compared with MATLAB simulation results and satisfactory agreement is noted.

In this proposed work we look at the analytical and numerical simulations of Dengue fever and recognize its best highlights alongside their execution for different situations. Dengue is a mosquito-borne viral disease. There are four particular, related, serotypes of the infection. Reviving from contamination by one gives long lasting invulnerability against that precise serotype. We are finding the solutions of nonlinear differential equations utilizing homotopy analysis method (HAM). We will utilize numerical simulations for MATLAB.

In this research, a vision-based wall-climbing robot is developed and discussed. In an electric power plant, a boiler is one of the most critical components, and the damage to this component can lead to a disaster in a power plant. Currently, boiler tubes are manually inspected to prevent damage and failure. The main damage in Cr-Mo tubes in a boiler plant includes fireside and internal corrosion that cause wall thinning. However, low carbon, overheating, creep, and hydrogen embrittlement are considered as the additional and primary damage mechanisms. It is observed that the creep occurs in low carbon steels for temperature over 400–440 °C, which is very common in boiler tubes. In Kazakhstan, the creep damage is the main reason for the tube failures. A regular inspection is crucial to prevent the damage to the boiler. In this paper, a robot is developed, which climbs vertically up to the end of the pipe and inspect the external surface of the tubes and pipes with the help of a camera. It consists of electronics and mechanical key components such as motors, controllers, sensors, and metal frames.

Automatic detection of cancer by digital images using data mining algorithm is required for next generation medical data analysis system. Classification of mammogram images for cancer detection is carried out with different machine learning algorithms as SVM (Support vector machine), KNN, Naive Bayesian. Analysing these algorithms are carried out to identify a technique that can provide the best possible result for all variant of images. This paper presents a qualitative study on these algorithms. The result of the analysis declares that SVM with bagging provides the better performance in the classification of mammogram images with accuracy of 99.467%.

Recently, big surge can be seen in big data technologies which requires companies to migrate their relational data into big data for improving the performance of their system. Odoo is an Enterprise Resource planning Software and it manages all kinds of business application that makes it a complete package for enterprise management. It stores its data in PostgreSQL relational database. Every audit log entry creates a very large database for Odoo application and it cause the problem of data handling. Also, it needs to read all the log entries whenever the problem occurs in the system which consumes lot of processing time. This research work will identify a NoSQL database for Odoo software and to design a framework for data migration of the Odoo ERP postgreSQL database into NoSQL database. The experimental results on different database of Odoo software shows the successful migration of data from a relational database to NoSQL database with an improvement in processing time.

Machine Learning and Data Mining have been used extensively in the field of medical science. Approximately 2% of the world population, i.e., 3.9 million people are infected by Hepatitis C. This paper is an investigative study on the comparison of classification models—Support Vector Machine, Random Forest Classifier, Decision Tree Classifier, Logistic Regression, and Naive Bayes Classifier—modeling Hepatitis C Data based on various performance measures—Accuracy, Balanced Accuracy, Precision, Recall, F1-Measure, Matthews Correlation Coefficient and many more using R Programming Language. On normalizing the numerical attributes using Z-score Normalization and using the holdout method for the Train Test data split of 80–20%, the result shows that Random Forest outperforms the other classifiers with an accuracy of 90.7%, followed by Support Vector Machine, Logistic Regression, Decision Tree Classifier, and Naive Bayes Classifier.

The worldwide demand for electrical energy is growing incessantly, twice the rate of primary energy consumption. Smart grid is one of the solution for the requirement of present scenario. Power electronics can be the key knowledge to build the following group of the more electrical power structures to support the major developments in energy efficiency, renewable energy integration and smart grid. The unpredictable nature of renewable energy resources due to varying weather settings results in voltage and frequency variations at the interrelated power grid.This paper describes the use of a buck converter for controlling the variable input voltage. Some DC-DC converters for photovoltaic necessitate that the input voltage be controlled while the output voltage is constant. This control is not so obvious and requires converter and regulator design. This paper presents the usage of a buck converter that are appropriate for connecting the solar power plant to the grid. Even though all power controllers have high controlling capability, not all achieve the identical in a bus voltage condition.

Most of the through hole and SMD electronic components have the technical information about the value and other technical specifications printed on the body of the component. This paper proposes Machine Vision system for Recognition of used throughole printed electronic components into different categories using optical character recognition technique. In this paper the method of classification of printed electronic components into different classes viz. Polar Capacitors, Disc capacitors and printed resistors, transistors and integrated circuits have been discussed. As the components to be classified are manufactured by different manufacturing companies it is a very challenging task to read the printed information with different fonts, different color background and different illumination level. Methodology for isolation of lines, words and character using segmentation has been discussed. Extracted characters were recognized using template matching method.

Semantic similarity and definition of words are very crucial part of the data. Therefore there are various online dictionaries and thesauri available on the web. But wordnet is the most famous online lexical system. It is an ontology-based system which organizes english nouns, verbs, and adjectives into synonym sets, each representing one underlying lexical concept. We have analyzed all important features of online dictionaries and thesauri. We found that some important features are missing into existing wordnet. Therefore in this paper, we have developed a modern wordnet which provides classification of the sentences into part of speech. In order to achieve our objective, we have developed a new ontology for modern wordnet.

In this growing technological era, path planning has become extensively useful application in the fields of robotics, surveillance and planning, gaming, animation, and bio-informatics. The act of path planning is the way to identify a collision free path from defined source to destination. In this paper, an analysis on the existing path planning concepts based on computational intelligence techniques is presented. Computational intelligence techniques have the capability to deal with uncertainty and approximation that makes these techniques more efficient to work on path planning. In this paper, the selected latest quality contributions of researchers are discussed along with their obstacles handing information, system type, and workspace environment. Moreover, some research questions also proposed and answered for the considered research contributions.

Yoga can significantly contribute to giving physical and mental relaxation as high-frequency brain waves (Gamma) generated according to specified yoga techniques. Yoga techniques like Pranayama (AnulomVilom, Kriya Yoga, etc.), Sudarshan Kriya, Super Brain Yoga, Meditation is getting huge admiration as a most feasible solution for healing stress, anxiety and depression-related brain disorders and increase in the brain performance after yoga sessions. Incorporating various yoga techniques has positively affected the human mind which leads to better social behavior in daily life and helps in relaxing mind. Various techniques which are utilized for analyzing the electroencephalograph (EEG) signals which are taken from the different subjects are reviewed in this article. Therefore, the objective of this article was to inspect and study the already present written works on the various outcomes of different yoga techniques on mind waves with the use of electroencephalography.

The effectiveness of the newly developed multiobjective nature-inspired evolutionary algorithm is proposed and discussed in this paper. It includes the integration of stochastic type particle swarm optimization and deterministic type simplex search method. As PSO based algorithms are quite successful in various engineering applications, therefore it is used here as base search to find the global optimum solution which is further refined by local search using simplex search method. The validity of the proposed method is tested by considering certain benchmark mark functions and the results are compared with conventional PSO method. The practical applicability of the method is checked by applying it on the engineering problem of economic emission dispatch in thermal power plants and the results obtained are compared with other traditional optimization methods. The results confirm the potential and superiority of the proposed combination of two different type of methods.

In recent days, user perceived latency has become a significant performance problem in the World Wide Web. The poor performance of the website is the important reason for the visitors to quit from the site. This reason may lead to loss of revenue for the e-commerce websites. In this paper, an attempt has been made to minimize web user perceived latency by predicting and prefetching the users’ future requested pages. The present work shows the Enhanced Monte Carlo Prediction (EMCP) algorithm by examining and including a rank for each predicted pages through feedback process from the most recent user navigation. Here, Webpages are predicted dynamically, (i.e.) the graph constructed in the work will refresh automatically whenever new pages are added to the website. Experimental results shows that better accuracy has been given by the rank based feedback prediction algorithm.

Statistical learning of molecular structure properties is gaining interests among researchers. These methods are faster compared to traditional QM based methods. In addition, the physical properties can be incorporated as feature sets and a properly trained model can predict the desired properties of a molecular system. For this, a number of machine learning regressors are used to predict molecular energies of Si − n (n = 1, 2, ⋯25) clusters, water, methane and ethane molecules. For the Si n cluster, six out of eight regressors seem to predict the energies accurately. For other data sets, Decision Tree Regressor prediction resulted fairly good, in general, compared to others. However, through the addition of atomic charges as an extra feature improved the performance of other regressors, this did not cause any improvement for the Decision Tree regressor. Since calculating atomic charges in itself is an expensive task, we summarize that decision tree regressor is suitable for predicting molecular properties compared to other regressors tested here.

Clustering is an essential step to discover the actionable information from complicated search space. In the era of digitization, the need to identify and structure this actionable information has made clustering one of the potential research areas. The traditional clustering models results into local optima, as clustering results confines to selection of initial seeds. Therefore, the computational models with heuristic search approach are required to get optimal clusters.This paper presents a review of the various approaches for research in data clustering. It describes the advancements achieved in the area of data clustering using evolutionary approaches and briefly traces the progress made to the clustering problem. Analysis of existing approaches is presented with critical remarks. Summary and comparison of related work are discussed. Finally, paper closes with a summary that leads to the issues and challenges for future research.

This is an era of data explosion. The amount of data being created and stored on a global level is almost unimaginable, and it just keeps growing. Immense information can be obtained from this data by applying various techniques of analytics. It has been observed that in the present-day scenario, there are many constraints associated with data acquisition, storage, analysis, search, sharing, transfer, visualization, querying updation, privacy, security etc. Data Lakes are emerging as one of the promising possible solutions to tackle these issues. The key feature of the data lakes is that the structured, unstructured and semi structured data can be stored in their raw format. Azure Data Lake Store (ADLS) provides optimized and best solutions for a wide range of Big Data analytics. It holds its base in Hadoop distributed file system (HDFS). It is quite scalable, secure and agile. ADLS supports multiple storage tiers at exabyte scale. This paper discusses the concepts of big data and data lake in addition to the architecture of Azure data lake.

Alzheimer’s is a disease of the brain that causes problems with memory, thinking and behavior. Using advanced electronic technology, the research work aids the victims to perform their daily scheduled tasks. The research work involves solution to Alzheimer’s patients by providing them a personal assistant. The assistant consists of an event reminder, location detector, a fall sensor, a Panic or SOS button for emergency. Basically, the assistant is a wearable module, which has a display, built in power supply; also it comes with all the required modules that does the above described functions. The patient’s caretaker is responsible for input data that is, the time during which the pills have to be taken must be provided initially by him. This information is stored in the database. The main aim of the project work is tracing path as well as navigating along with emergency contacts and also health monitoring.

The OCR technology is gaining more and more importance in the digitalization of the documents, this is because of its functionality to convert the text data in the image to the machine-encoded text and this machine-encoded text can be further used for processing. The orientation of the digitized document is important for the OCR to recognize the data in the document veraciously. Sometimes due to manual error, the scanned document may not be properly oriented, this condition is called skew of an image. Deskewing is a procedure to align the image properly, before further processing the data in the image. There are many existing approaches for deskewing the image such as mathematical morphology, principal of connected components, projection profile technique, Fourier transform, Hough transform, Radon transform and KL Transform. These methods for deskewing have their own constraints with respect to font style, font size and are not rotation invariant. In this paper, we propose a method which can deskew an image with any degree of skewness using warp-affine transform, Hough transform and feedback of the OCR output. The warp-affine transform is used for adjusting the shape of the background image, Hough transform is used for checking the vertical symmetry of the text and feedback from OCR is used for checking the skewness of 180° and flipped document cases. The proposed method was evaluated on 40 images with the various skew angle and the performance was comparable with the existing techniques in the literature.

Blindness is the complete or partial loss of vision. It has dramatic impacts on individuals experiencing such disabilities, affecting the quality of life and holds backs them from performing many of their day to activities. It can last for years or lifelong. There are different forms of surgery can help, but sometimes no treatment is available. This paper presents a system concept to provide a smart electronic travelling aid for blind people. The system is intended to help blind people with a navigation assistant. This consists of obstacle detector, voice commands, location tracker and power backup unit. The assistive system helps the blind to find obstacle free path based on the voice commands. With the push of an emergency button, the blind person can inform his caretaker in case of emergencies. The main aim of this research work is to provide a better solution regarding navigation system for blind people with all the mentioned functions.

Breast cancer detection can be done using mammography which uses low-dose X-rays to obtain images of the breast in order to detect the abnormalities present in it. Analysing the mammogram can be done by using a Computer Aided Diagnosis (CAD) system to detect abnormalities. Designing a CAD system involves pre-processing, segmentation, feature learning and finally classifying the types of abnormalities using an efficient classifier. This paper compares the various steps and techniques used to design a CAD system and the results obtained are analysed based on their performance. By comparing various methods for pre-processing, segmentation and feature learning, it was found that different methods work well in different contexts. Some of the drawbacks in a few systems are also discussed and future works for further improvements have been suggested.

Accurate prediction of traffic makes it easy to make decisions of travelling route, travelling schedule, travel vehicles choice for a commuter. The surveillance systems, GPS system installed on road way are the abundant source of traffic data. This huge amount of traffic data and increased computing power definitely motivates researchers to analyze the data to solve the road traffic and transportation problems. Deep learning always proved to be a good solution for prediction problems such as audio classification, signal processing, image classification and Network traffic prediction. This research shows that Long short term memory network can be used to design a traffic forecast model. Many researchers have proved that deep sigmoidal networks could be trained to produce good results for many tasks such as audio classification, signal processing, image classification and Network traffic prediction. In this work comparison of LSTM with Linear Regression, Logistics Regression and ARIMA was done.

The main cause of atherosclerosis is stroke. To evaluate this disease, the assurance of the surface density [Intima-Media Thickness (IMT)], portrait about atherosclerotic carotid plaque, assessment about width of common carotid artery (CCA), along with their classifying of stenosis to be done. This proposed paper presents the analysis that considers the design and automated arrangement article achieving automated or semi-automated bisection in the ultrasound image and the CCA. These automated systems are developed based on the techniques are edge detection, active contours, level sets, dynamic programming, local statistics, Hough transform, statistical modeling, neural networks, and also with assimilation of the raised approach. To identify the best segmentation method, the techniques proposed in the literature are implemented in the different stages and CCA segmentation approach, forthcoming viewpoint, and more expansion of the approach is ultrasound carotid bisection and wall tracking of CCA.

Minutiae-based fingerprint recognition system is an important technique for person identification. Yet spurious and false minutiae are often occurred and it can be removed during the post-processing phase is passible. False minutiae will affect the accuracy of fingerprint matching process. Hence it is essential to decrease the false minutiae and increase the fingerprint authentication. Besides initially at the pre-processing the input fingerprint image is subjected to adaptive histogram equalization (AHE) followed by Gabor Filter (GF). Using Gabor filter to improve the given fingerprint in wavelet domain and restructure the fingerprint. Further the enhanced image subjected to binarization and thinning process. After successful pre-processing the minutiae set extracted by crossing number method while the existence of spurious minutiae lying on the boundaries of the fingerprint image. To overcome this problem in post-processing stage, the Graham’s Scan Algorithm (GSA) based closed hull filtering technique (CHFT) is successfully to remove the border minutiae. The studied closed hull filtering can be simply superposed to the fingerprint template, the accuracy of filtered minutiae evaluated by Goodness Index (GI) value calculated through manually. Thus the post-processing method, the Graham’s scan algorithm applied on fingerprint minutiae points(vertex) proved mathematically, The investigational results of the projected algorithm was verified on FVC2002and 2004 fingerprint database DBI.

Online Social Networks (OSN) offers a dais for the individuals to create a social affiliation or network with people of common interest. These sites plays a vital role in people’s lives ranging from entertainment to various business. Application Program Interface developed for OSN can be used to avail services offered by other third party applications. OSNs obliges as a consistent identity for web services. Due to the open nature of Social Networking Sites, they are more susceptible to malicious attacks. One major attack is creation of bogus identities called Sybil attack. This attack leads to outflow of private information, forged reviews and various fraudulent activities. Different methods such as graph based techniques, machine learning based techniques are there to detect and prevent Sybil attack in OSN. This paper provides a perspective view on various research works implemented on this area.

The overall recognition rate of a biometric authentic system is completely depends upon the input of an image quality. Nowadays the image qualities are besmirched by some reasons such as apprehending the image of input devices, underprivileged quality of an image, large enlightenment dissimilarities of the inside and outside environments, pixel intensity level, etc. In biometrics based recognition system various physical and behavioral traits are used for recognition purpose. Here considering minor finger knuckle print trait is secondhand for authentication process. The main issue of the MinorFKP system is the quality of an image is relatively poor level and matching level of the image is difficult due to poor image quality. Therefore the CLAHE method is applied to MinorFKP image enhancement. The values of an MSE, RMSE and PSNR that the CLAHE technique considerably improves the better quality of MinorFKP images and reducing noise and artifacts.

Providing personalized learning path according to the individual learning characteristics poses great challenge. In the personalized learning environment, the content and the sequence of contents into path vary with individuals depending on their needs. Researchers have been proposing several variations on swarm intelligence and evolutionary algorithms since last decade. Their intention is to improve the performance of these algorithms on various optimization problems such as travelling salesman problem, scheduling problem. In this paper, Genetic Algorithm (GA) and Ant Colony Optimization Algorithm (ACO) have been combined for constructing personalized learning path. Firstly, experiment was conducted to choose the best possible value for crucial parameters that influence the efficiency of the algorithm. Secondly, proposed method was compared for its execution time and quality of results to state the better performing algorithm. The experiment results indicated that the proposed method performs better produces good quality result with smaller to medium solution space.

Chest X-rays are widely used for diagnosis of diseases such as pneumonia which affects the lungs. This paper provides an approach to detect pneumonia and classify the chest X-ray images into two classes pneumonia or normal using convolutional neural networks. This is done by training the convolutional neural network to differentiate between the normal and pneumonia chest X-ray images using a deep learning platform Pytorch. Image preprocessing technique has been applied in order to enhance the image. Python and OpenCV have been used.

Advancement in technology has made systems integrated, connected and communicating with each other, so amount of data generated in day to day life has been increasing in leaps and bounds. IOT is the best example of such kind of systems and it has opened new gates of research on such data generation and analysis. Data generated by non-stationary system are fast, huge, and continuous in nature. These data are termed as data streams. Mining of these kind of data has inbuilt challenges as they possess different characteristics. Traditional algorithms are not well suited for these kind of data. Also, Mining data streams to classify outlier attribute becomes a more tedious task as data arrives continuously. Also, multiple scans of stream data is not possible due to its huge size. Hence, to address above said issues, changes in the structure of algorithm needs to be done. In this paper, a modified approach on outlier detection method MCOD has been discussed and proposed which gives improved results in terms of outlier attribute detection.

The capability of a gesture recognition using input device Kinect is explored. There is a continuous need to communicate using sign languages, such as interacting with people who have hearing and speech impairment. The situations are when silent communication is favored. For the sample, during a surgery, a surgeon may signal to the nurse for assistance and requirement of medical tools. It is difficult for most people to communicate with deaf/dumb people who are not acquainted with a sign language without an interpreter. Therefore, a combination of software and hardware that converts sign languages into plain text and then to speak can help with real-time communication, and it also provides interactive training for people to learn a sign language and communicate with the people having speech and hearing impairment. Here in our project Hand gesture recognition system can be divided into two parts according to its processing steps: hand detection, and gesture recognition and we will achieve this detection and recognition using depth sensor of Kinect (a line of motion sensing input device). The Kinect has an infrared blaster which is used for depth measurement. The paper is not only aimed at converting the sign language into voice but also can be useful for controlling different appliances using the gesture. To overcome these disabilities this project can be used efficiently.

The semantic interoperability is the communication and understanding in between user and semantic web. The user gives a query to the web and gets the response from the search engine with the help of semantic web. The semantic interoperability gives idea of semantic web which need a system to retrieve data and reuse the retrieved data with their precise meanings. It is very difficult to achieve such semantic interoperability in different systems. The problem of semantic interoperability arises due to semantic heterogeneity in the World Wide Web. The QOMS Query Ontology Matching schema is proposed to produce accurate results of the given query. The extracted semantics help to reduce the ambiguity.

Machine learning and data mining methods plays major role in biosciences. Now-a-days data mining methods are used to intelligently transform the information available into valuable knowledge. Gestational Diabetes Mellitus (GDM) is a kind of diabetes that occur in women during pregnancy. Some women develop high blood glucose levels during their gestation. Gestational Diabetes Mellitus if ignored and untreated can result in permanent medical problems to the baby in the future. This is identified as a serious open problem of research which calls for good prediction algorithms to predict the GDM at an earlier stage of gestation. Literature shows a wide range of machine learning algorithms employed for the prediction of GDM. This paper proposes novel prediction framework for gestational diabetes based on Logistic Regression. This results of the framework show promising results of better prediction at an early stage of gestation.

As the majority of the people uses Liquefied Petroleum Gas (LPG) as a fuel for cooking, but the technology applied in this field is very less.A major issue in usage of LPG cylinder is that the cylinder gets emptied and we are unaware when it is going to finish. Liquefied petroleum gas is a flammable gas, which has the potential to create a hazard. Therefore it is important that the properties and safe handling of LPG are understood and applied in the domestic and commercial/industrial situations. The proposed system is aimed at providing a method that constantly monitors the gas pressure with the help of a transducer (Jarašūniene. Transport 22: 61–67, 2007) and sense gas leakage with the help of mq2 sensor (Om Prasad. “Event-driven information integration for the digital oilfield.” In SPE Annual Technical Conference and Exhibition. Society of Petroleum Engineers, 2012). This data is made available in real time through real time feeds over the internet. Mongo DB [11] database can be used for storing the data that is generated by the transducer and mq2 sensor. The transducer monitors, and detected the gas pressure and sends an alert to the client when 20% of gas is left inside the cylinder through IFTTT [13] (web sms, mail services) server. The device also raises an emergency alarm through a buzzer. Based on the real time data feed connected to Mongo DB. Data analysis can be done in future on usage of LPG cylinder.

The simplicity of digital media alteration and diffusion necessitates satisfied shield ahead of encryption. In decrypted form information’s are hidden in the form of watermarks. By embedding owner’s information, their property is protected and copyright protection is strengthened. To improve robustness and security multiple image watermarking is applied for the copyright of owner’s. Several literatures were assured about Discrete Wavelet Transform (DWT) watermarking schemes for data protection. On the other hand, DWT based watermarking schemes are establish to be fewer robust beside image processing attacks and the shift variance of Wavelet Packet Transform causes erroneous extraction. The multiple images watermarking technique based on hybrid Steerable Pyramid with Discrete Wavelet Transform (SPDWT) is proposed and it is compared with Integer Wavelet Transform (IWT). In this work, Input image is transformed by steerable pyramid results in various sub-bands and DWT is applied to these sub-bands. For watermark extraction, Pearson Independent component Analysis (ICA) is applied as it attains the new trait is that it not entails the renovation procedure in watermark extraction. The robustness of the proposed scheme is validated by different attacks to be applied on watermarked image. The performance measures like PSNR and Normalized Correlation are calculated to check the imperceptibility and robustness over integer wavelet transform.

Ultrasound imaging is one of the most generally & frequently used techniques in the medical field due to its low cost, portability and non-invasive nature. Kidney stones & tumors are the two form of the disorder that are accurately & correctly detectable by employing ultrasound imaging. This article provides an understanding of automation which could be embedded with the ultrasound system resulting in spatial as well as dimensional analysis of various kidneys disorders in ultrasound images through image processing. Usually, human visual perception is utilized for detecting the exact position & spatial dimension of these artifacts. The advent of new image processing era has enabled the new ways for detecting these kidney related artifacts accurately. With these new image processing techniques, the beginners/experience radiologists, doctors’ paramedics, etc. are able to draw correct conclusion regarding the presence of artifacts through these techniques and hence opt for a correct course of treatment for the removal of the identified disorder. A fast and accurate way of diagnosis can be achieved which leads to better therapy and confidence in patients. The aim of this review paper is to understand & review the various techniques of image processing which are frequently employed in the detection of disorders in ultrasound images. A brief comparison of these image processing techniques is also conducted where the comparing parameter is complexities & accuracy.

Uncertainty is almost part of every field of scientific readings and it conveys our life processes. Uncertainty can be due to many reasons like incomplete information, ambiguity, lack of knowledge and no specific information on a specific topic. Broadly two categories of uncertainty have been defined, namely, aleatory uncertainty and epistemic uncertainty. Aleatory uncertainty arises due to natural reasons and randomness like random show of a flipped coin, while, epistemic uncertainty is the result of lack of knowledge on a subject like calculating distance between two places. Also, Different methods have been designed to handle different kind of uncertainty. This paper analyses handling of uncertainty activities belonging to multiple disciplines in diverse fields. Different uncertainty handling studies related to fields of engineering, economics, ecology and information sciences have also been explained. Hopefully this study will provide understanding to public on how uncertainty is handled in other disciplines and what are the challenges and issues faced by them. It will also give and instigate database researchers to create better data management tools and techniques to resolve many uncertainty issues in today’s world.

IoT devices are part of our daily life, and organizations allow us to use them to connect to their secure network. In this regard we analyze the security challenges and proposean architecture for hetrogenous IoT environment to make sure that only authenticated users are connected and allowed to communicate. Designed a system which automatically detects the presence of suspicious IoT devices in secure perimeter of an organization. Also we designed and implemented a new light weight hash-based digital signature algorithm for authentic access in a resource-constrained environment. The simulation of our idea result in the algorithm providing authenticity and integrity with a small foot print and reduces power consumption.

Placement is a critical step in FPGA physical design. Placement determines the locations of the logic and I/O blocks on the FPGA Proper placement reduces the wire-length and routing time and in turn, increases the overall efficiency of the FPGA. Here a modified adaptive inertia weight Particle Swarm Optimization (PSO) algorithm is applied for placement problem in FPGA. The convergence behavior of the modified adaptive inertia weight PSO algorithm is analyzed and fast convergence is observed. The algorithm is implemented in the VPR tool and the performance is evaluated based on the wire-length and compared with that of VPR placement algorithm using the MCNC benchmark circuits. The modified adaptive inertia weight PSO algorithm gives more optimized results for FPGA placement problem.

Collagen Gene is a protein which is otherwise called alpha-1 type I collagen is found in human beings in type of COL1A1 encoded. The real encoding segment is type I collagen which is the fibrillar collagen. This febrile collagen includes the cartilage which is mainly found in the connectivity tissues. The diseases caused by this collagen are Osteogenesis imperfecta, Chondrodysplasias, Ehlers-Danlos Syndrome, Alport syndrome, Osteoporosis and Knobloch syndrome. These syndromes are associated according to the kinds of collagen which are from Type I to XVIII. Each kind of syndrome has respective impacts over human in which these five syndromes have real impact over the body. Clustering is one of the information mining techniques which is utilized to gather similar articles. The proposed framework is mainly utilized to assemble the significance of disease according to the phases of the collagen gene compose. This procedure of clustering is finished using the Dynamic Path Selection Clustering algorithm to assist in the prompts over the recurrence, conditional changes and its clinical significance which demonstrates the analytical factor in the gene behaviour.

Water is the key for life to sustain and to guarantees people’s quality of life. The water resource management plays an important role in checking the unnecessary wastage of water. In water resource management the demand forecasting plays the key feature in the planning of the distribution of water. There are traditional methods for forecasting the demand, but these methods lack in accuracy. To predict the demand it is possible to use the multiple linear regression, which offers more accuracy than the traditional method. In this paper, we propose a system which uses a ultrasonic sensor, NodeMCU and a progressive web app to collect the water usage data from the user. The NodeMCU and the ultrasonic sensor constitutes the hardware of the system which is a IoT Level 3 system. Using the collected data with multiple linear regression it is possible to predict the water required for the next month for the user. The proposed system was evaluated for 10 data sets and the results were comparable.

Decision making is a process of reaction against organizational hazards and opportunities. It includes the process of collecting and processing the information gathered and selecting the alternative from the set of alternatives based on their values using different tools, techniques, and insights. Regression analysis is one among the most dominant techniques used for decision making in business by the management. To make better decisions, regression analysis helps the managers to understand the data to model dependencies and helps to understand the relationship between the expected output and the input features to predict the values. The main application of regression analysis is to find how strong an independent variable influence the dependent variable. There are many areas where regression analysis can be applied in organizations for better prediction, mainly in financial forecasting, marketing, understanding inventory levels, supply chain, trend analysis, and time series prediction. In this paper, the application of regression analysis in different organizational decision making and the different types of regressions used in organizational decision making are discussed.

This paper describes blockchain based architecture for the Human Organ Transplantation, to provide a trustworthy mechanism that will make sure a fair distribution of organs available. Transparency and auto tracking features of blockchain will make sure that no single party controls the priority of recipients in the waiting list. Every transaction is visible to all those who are having access to the system. It can also ensure a control on the organ trafficking and provide a reliable and secure exchange mechanism for recipient and donor health records. It can greatly revolutionize the transplantation network with greater efficiency by allowing every entity to directly verify the transaction records by eliminating many of the intermediaries as we know them today.

In recent times, Melanoma has become one of the most dreadful type of skin cancers with mortality rates being high. Although there exist state of the art methods for identification of melanoma, the usefulness of automated approach such as deep learning proves to be very much appealing. This paper deals with Convolutional neural network framework, which has been evaluated for non-dermoscopic images of melanoma and benign nevi for early diagnosis and efficient classification. The image dataset has 70 images of melanoma and 100 images of benign nevi, which was augmented to 1020 images and then split into two groups. These groups are trained, and a two-fold validation is done for achieving better accuracy.

Accident data is often recorded more for the sake of updating record books and for creating statistical information rather than, as a source of intelligence. As most prior studies have been persistent on a few risk factors, some specific road users, or certain types of accidents, many significant factors affecting injury or sternness of the collision have not been completely identified as yet. The fundamental stipulation for recuperating road security is to attain and analyze a comprehensive road catastrophe database. An advanced road accident analysis system is needed to help develop a road safety initiative strategy as well as to instil a better understanding of road traffic accidents. Data mining has the potential to abolish paucity related to road traffic accidents as well as statistical constraints. In this paper, we analyze data extraction methods, which can be applied to arrive at some new, intangible, and reasonable facts from road traffic accident data.

Separating cells from the background in microscopy images is the critical step in image processing pipeline for the study of single cell life cycle. Live cell imaging experiments involve thousands of cells and images taken for a few days, which results in huge data generation. Automatic analysis of such images is essential rather than performing analysis manually. The challenges involved are non-uniform illumination of the image, different types of cell lines to be studied, large curation time required and analysis of large data to name a few. In this work we present a image processing pipeline using a convolutional neural network (CNN) model followed by thresholding and morphological operations for segmenting the NIH 3T3 cells in microscopic images. The segmentation results are evaluated by comparing them with the ground truth images. The proposed methodology gave a Dice index of 0.93 on a stack of 238 phase contrast images. Further, we show that CNN based approach performs superior to conventional image processing segmentation methods on phase contrast images of NIH 3T3 cells.

Server farms over Cloud portrays with load adjusting is a crucial advance in processing by offering shared computational intensity of the assets on interest. Being grounded on the major idea of virtualization, it has essentially changed the way of conveying the IT administrations with limited infrastructural prerequisites. The virtual condition includes the production of numerous VMs (or virtual servers) with legitimate load adjusting on a solitary physical hub. In genuine setting, the numerous working frameworks (OSs) can keep running on a solitary OS as for server farms hidden the relocation recompense benefit stage. The running of virtual servers limits the asset sit out of gear time over the server farms with legitimate load adjusting utilizing an approach brought virtualization transient over the cloud with load offsetting with threshold (VMOVLBWT) as for characteristics, along these lines keeping the asset under-usage. Furthermore, the decrease in the measure of required equipment brings down the power required for task which thusly chops down the vitality request. The decentralized administration server farm with appropriate load adjusting) allocator with virtual machine relocation over unique server farms.

Rice (Oryza sativa L.) is susceptible to a number of diseases. Among them, sheath rot disease which is caused by Sarocladium oryzae (Gums & Hawks.) is the most devastating diseases and major challenge to rice cultivation. Use of Plant Growth Promoting Rhizobacteria (PGPR) for biocontrol viz., Pseudomonas fluorescents is an another disease management approach as it is the growth promotion and reduces disease in crops. Fuzzy Expert System with the algorithm Fuzzy Verdict Method is used to find the disease severity of rice. The Fuzzy expert system has three phases; they are fuzzification which is followed by Fuzzy Verdict Method and defuzzification phase. The fuzzification phase helps to change the crisp value into fuzzy value. The proposed algorithm helps to diagnosis the disease severity of rice crop with the input parameters Number of discoloured grains/panicle, Number of chaffy grains/panicle, Lesion Number/tiller, Lesion size (mm)-Length& width and Number of panicles infected/tiller, becomes simpler for farmers and scientist. Algorithm uses triangular membership function with mamdani’s interface. The fuzzy values are changed into crisp values using defuzzification phase. The algorithm was tested using Fuzzy tool box in MATLAB to diagnosis the disease severity of rice.

Getting appropriate data is the greatest need of today’s world, particularly for visually impaired, and when the data is in form of online media, its representation in different format is a major issue. We can solve this issue by commenting on (naming) the pictures using different content, to make the image simpler and more semantic for visually impaired individuals. Semantic annotation of pictures provides a way for semantic search of image with less difficulty, and gives ease in representation. This paper presents a user-friendly tool, called HELPI VIZ, for annotating, visualizing and navigating images using ontologies. It is designed to be used by everyone, including ontology experts and users not familiar with ontologies. It provides an accessible and understandable user interface that follows the “design-for-all” philosophy. It is a responsive web application for the visualization of annotated pictures and ontologies using sound and text. It provides a platform to increases tourism by making historical pictures of monuments and sculpture self-talking by annotating with text and sound. It can be used as a image visualization and search tool for online media content, it can used as educational tool kit for students and numerous all the more such applications.

This paper reviews on detection and classification of multi-abnormalities occur in small bowel region. Multi-abnormalities like ulcer, bleeding, polyp and tumor are caught by utilizing Wireless Capsule Endoscopy (WCE). WCE images are utilized to diagnosis the infections in the stomach related tract. To identify and detect the diseases which occur in small bowl is difficult for human to recognize the exact type of disease. To overcome this problem, various Image processing and machine learning techniques are utilized over the decade to detect and identify the diseases more accurately are discussed in this paper.

Alzheimer’s Disease (AD) is a serious neurodegenerative and progressive disease. It annihilates memory power and other mental functionalities. Development of Computer aided diagnosis will provide a second opinion to the radiologist in the detection and diagnosis processes. So, herein a novel method has been developed to study the brain MR image to classify the normal and Alzheimer’s. The proposed method consists of two stages. In the first stage image pre-processing like image enhancement, skull stripping and region of interest extraction has been performed. In the later stage, different graph structure based methods like local graph structure, extended local graph structure and hybrid local graph structure has been experimented and reported. The proposed method is experimented using publically available database OASIS and the average accuracy is observed to be 74.84%. This proposed method will floor the approach for developing devices for detection of AD in the future. Alzheimer’s Disease (AD) is an progressive neuro degenerative disease.

Object Tracking Wireless Sensor Network [OTWSN] are used for sensing in various application associated with smart city. Sensing involves monitoring and tracking using advanced technologies and intelligent devices. In this paper recent trends and technologies have been reviewed in OTWSN pertaining to a variety of applications. Architectural design, terminologies, technical challenges along with solutions provided by the researchers have been studied and summarised. Future direction for this research has been discussed in context of OTWSN for smart surveillance.

Living creatures especially human, always aimed to commune every process or incident that take place within the environment, to lead an easy and luxurious life. Everyday a person has to execute certain basic tasks to control their body movements or particular parts of the body. Paralyzed people do not have control over some of their body parts. However, there are persons who are severely paralyzed and they cannot move themselves. They need some assistive technologies to fulfill their needs. A person with disabilities, mainly total paralysis is often unable to exploit the biological communication channels such as voice and action. One such condition was massive Brainstem Lesions, Stupor, Guillain-Baree Syndrome and Traumatic Brain Injury. In these conditions they cannot move their muscles, but they can able to control their eye movement, which leads to a condition called locked in state. In this state the person were unable to control all the motor neural activity which leads to other communication technique to convey their thoughts with others using eye movements. To solve this problem eye controlled interfaces are needed. Human Computer Interfaces help individuals with disabilities to communicate through a computer using a digital channel and make life more prosperous for the paralyzed patients and further enhance their quality of life with the support of bio-based HCI.

Machine learning is one of the advance topic in research areas. In this machine learns itself to do certain predictions and works. Machine learning is mainly divided into different types of learning methods and each method is using different algorithms. Using these algorithms and methods the manual work is reduced and Automatic way of working is done or work is done automated. Agile systems using machine learning is one of the developing technology in current and mostly liked y the people. Agile systems uses algorithms of machine learning and their types of learning are explained in this paper. Machine learning along with artificial intelligence is said to be one of the prevailing topic which are used, using and under researching.

The argument between the open source and licensed forensic tools has always proved that the civilization is based on fundamental issues like reliability and security. This paper shows the comparative study of different forensic desktop recovery tool on different parameters and functionality. Comparative study gives the result of the best suitable forensic tool for a Desktop Forensic application and Live Forensic application. It also has a brief description of the tools which are available for the benchmarking process and also a brief description of the parameters used in this research. So this study of different types of forensic tools will save our time in correct verdict for the range of the tool.

A movie recommendation is vital in our social life because of its quality in giving improved excitement. Such a framework can recommend an arrangement of movies to users in light of their advantage, or the popularities of the movies. Despite the fact that, an arrangement of movie recommendation frameworks has been proposed, the vast majority of these either can’t prescribe a movie to the existing users productively or to another user by any methods. This paper proposes a movie recommendation framework that can extract from data and recommend similar movies to users, based on users input using Item based Collaborative Filtering. First, user item rating matrix is examined to identify relationships among various items, then finally similar movies were recommended based on user’s input. A part of this recommender system is execute using Apache Pig and Hadoop Distributed File System is used as data storage.

The general aim of our study and this research is to find out better classifier for musical device identification with great accuracy. This is one of the most popular topics for study. In our research paper, we present the idea to identify the musical instrument from a monophonic audio signal. For this purpose, we have used Cepstral features (i.e. MFCC features) extraction technique for extraction of features and there is the number of classifiers out of which, we have used SVM and KNN classifiers for sorting purpose. We have compared the results from both classifiers. In our work, we have made a catalog of different music samples from various musical instruments. We use this catalog for both training and testing purpose.

The efficient feature matching algorithms are used to improve the quality of object instance search from videos. A trajectory is created based on a sequence of bounding boxes that track the object instance in each frame. The goal is to track the trajectories in high amount of video files. Although the traditional methods of object instance search works well on large image dataset but it fails to produce accurate result in time on videos, which concerns about locating instances of the query object with various changes like color, shape and background. The proposed algorithm was tested with NTU database and it achieves an overall accuracy of 94%.

In this paper, we propose a modified multichannel Kalman speech enhancement algorithm for the enhancement of noisy signal from the effect of colored noise. Compared with other multichannel speech enhancement algorithms, the projected algorithm requires lower computational resources with satisfactory noise reduction and lower signal distortion.

The study has proposed a hybrid watermarking method. It consists of numerous leveled thresholds, in which pixel-inferred and pixel determined watermark information are conveyed through the slightest significant bits of all pixels. It basically masks specific location for buying that region watermarked, and additionally may be applicable for content retrieval. The content may be watermarked without intermixing the records pixel and provide a green result. The experimented end result is commonly based totally on dynamic threshold. Considering assumptions, it is proposed that data pixel is absolute impartial from every different in order that if threshold activated over a probe picture dynamically, the records pixel tends no longer to combine collectively and attain in efficient way. The paper consists of different theory blended together to carry out. This is greater like content based watermarked technique.

Cardiovascular diseases are commonly found all over the world. Patients having Cardiovascular risk (CVR) should not stop doing their daily activities without any fear or risk. This is achievable by continuous monitoring of the cardiovascular system to diagnose and avoid cardiovascular traumas such as cardiac arrest, in the minimum time. A little awareness and expert’s based decision support system would help patient to analyze the symptoms of cardiac arrest. This would help patient to get medical help as soon as possible and avoid the risk of cardiac arrest. In this paper, we designed a decision support system using fuzzy logic which allows us to represent the expert’s knowledge in terms of mathematics, accepting some level of uncertainties which lies within experts. Interval Type-2 based fuzzy logic system is designed and implemented using MATLAB. Developed system is tested on ten patients out of which eight patients diagnosis has validated with the test results. As this system is completely based on expert’s expertise, accuracy of the developed system depends on expert’s skill.

Retinal disease often refers to retinal vascular disease is currently growing tremendously in the field of ophthalmology which is to be diagnosed early to prevent from blindness. In recent days, many retinal diseases that cause damage to retina due to abnormal blood flow. In this paper, multiclass classification is been proposed for four classes of diseases such as Arteriosclerotic Retinopathy, Background Diabetic Retinopathy, Choroidal Neovascularization, Hypertensive Retinopathy. The proposed model uses convolutional neural network which extracts its own features and classify them when compared to other classification methods. The system uses ALEXNET architecture which is deeper with more filters that can extract the features of the image automatically and classifies them to predict the class of disease it belongs to. The model is trained over the data which is been collected from STRARE database. As a result, the model is able to achieve the prediction of test case and classifies the disease which brings betterment in diagnosis of Retinal Diseases and avoids from blindness.

Accurate quantification of cartilage is useful for diagnosis and treatment of osteoarthritis (OA) affected knee joints. Image processing techniques are required for clear visualization and quantification of cartilage degradations in different regions in OA affected knee joints. In this work femur articular cartilage were segmented from MRI of knee joint using two semiautomatic methods namely canny edge detection based method and radial search based method. The thickness and volume of cartilage were measured region wise using segmented images. The cartilages were also segmented using standard method under the supervision of radiologists for comparison and to find the accuracy of quantification results of two semiautomatic methods. The results of measurements using semiautomatic segmentation methods shown good accuracy and the errors are limited to less than 5%.

Machine Learning is used in many data analytics problems to predict the future with more accuracy. Trend of stock and index price are important issues of this arcade. Stock is great option of attracting investors and financial indexes of country. The target of this paper is discovery progression of Facebook stock observations using s and p indexes using numerous machine learning methods. We can use numerous machine learning algorithms to achieve the results. Moreover, we can predict weather arcade of Facebook stock is positive or negative. The result proves that Facebook stock exchange can be finding with machine learning methods.

Rice (Oryza sativa L.) is susceptible to a number of diseases. Among them, sheath rot disease which is caused by Sarocladium oryzae (Gums & Hawks.) is the most devastating diseases and major challenge to rice cultivation. Use of Plant Growth Promoting Rhizobacteria (PGPR) for biocontrol viz., Pseudomonas fluorescents is an another disease management approach as it is the growth promotion and reduces disease in crops. Fuzzy Expert System with the algorithm Fuzzy Verdict Method is used to find the disease severity of rice. The Fuzzy expert system has three phases; they are fuzzification which is followed by Fuzzy Verdict Method and defuzzification phase. The fuzzification phase helps to change the crisp value into fuzzy value. The proposed algorithm helps to diagnosis the disease severity of rice crop with the input parameter Number of discoloured grains/panicle, Number of chaffy grains/panicle, Lesion Number/tiller, Lesion size (mm)-Length& width and Number of panicles infected/tiller, becomes simpler for farmers and scientist. Algorithm uses triangular membership function with mamdani’s interface. The fuzzy values are changed into crisp values using defuzzification phase. The algorithm was tested using Fuzzy tool box in MATLAB to diagnosis the disease severity of rice.

Cloud computing mainly deals with the cloud services and cloud storage for its cloud users as well as to deliver all the data effectively from multiple cloud data centres. The cloud server plays an important role in cloud load balancing. As the number of cloud servers are increased day by day and we are continuously searching for optimal data and more reliable services over the cloud. We propose a new Bio-inspired fuzzy models in meta-heuristic algorithm named Firefly search algorithm that optimizes the load balancing of tasks among multiple virtual machines (VMs) in the cloud server thereby improving the energy efficiency in cloud servers. The proposed algorithm shows marked improvement in terms of throughput, response time, etc., when compared with existing cloud based load balancing algorithms.

Cloud-Forensic-as-Service is the one which common person’s daily accessible service which is centrally reposted with all the criminal’s forensic data in cloud environments. The continuous updating with all the records as faces with proper trained infrastructure with exact mean and variance values which will be mounted as digitized data with novel method of accommodation where client can use to cross check the forensic analysis in higher accuracy result with finite precession. The neural association algorithm will be used in the SaaS (Software as Service) where the logic is maintained with dynamic self-updating of the digitized data with deep learning model and anonymous clients concurrently can access this service over the cloud. This forensic cloud will not be accessed by all common persons and also lack of awareness where to check this type forensic check up with easy user friendly support. So to address this problem this work proposed a new framework NAMAFDS (Neural Association with Multi Access Forensic Dashboard as Service) which will give continuous rendering services to clients who wants instant criminal forensic check-up which is globally indicated as Cloud-Forensic-as-Service.

A set S ⊆ V of a graph G is said to be restrained step dominating set, if <S> is the restrained dominating set and <V − S> is a perfect matching. The minimum cardinality taken over all the restrained step dominating set is called the restrained step dominating number of G and is denoted by γrsd(G). In this paper we have discussed its application and extend the study of this parameter for Cartesian product of graphs.

In agriculture, most of the works are labour intensive. Farmers are suffering from lack of work force that struggles the industry to cope up with market need. In developed countries smart farming became a boon to increase efficiency and quality of agriculture production. Smart farming is a technology that involves various advanced technologies like IoT, robotics, automation systems, precision agriculture into existing farming methods. This improves the quality of life for farmers by reducing labour and tedious works. Robots or automation systems developed for agriculture has limited computational capability, memory and storage. Cloud computing enables users to access scalable pool of resources on-demand. We can use conveniently high-performance computing and storage through cloud with low cost. In this paper, we introduce the use of cloud computing in the smart farming and precision agriculture. We believe our paper creates innovation and scope for designing Cloud based Agri-Robots which helps to improve agricultural production efficiently.

The high quality and resolution of the time-frequency representation of Bio signals is one of the important area to depict geological structures which are used to reveal the valuable information hidden in them. In this paper, we proposed time-frequency analysis of Bio signal EMG using the empirical wavelet transform (EWT) which is a new approach for analyzing multichannel Bio signals. Comparative results of the EWT, STFT (short time Fourier transform) and CWT (continuous wavelet transform) were analyzed in MATLAB. Overall comparison of the results based on EMT with those obtained from STFT and CWT signals, revealed that PSNR values has been reasonably increased for EWT. In addition, it was observed that presence of EWT-based instantaneous frequency spectra can produce much sparser representation and much higher time frequency resolution than the traditional CWT approach. This emphasizes the feature of empirical wavelet transform.

In recent days, the development in the Internet plays a major role in all types of activities that have been driven specialists to consider research frameworks to help the clients and applications in getting the directions by conveying nature of administration in systems. Some sorts of strategies are appropriate for giving security in correspondence to seized conditions, for example, portable registering, internet business, media transmission, and system administration. In such cases, service providers are focusing more on enriching the service access to end users. Since, there is an intermediate data theft or attack that was occurred during the broadcast process. Hence, an overview on different detection schemes are considered for successful determination and arrangement for attack recognition with various neural systems and some swarm computations has been proposed. The proposed strategies have been valuable for adequately identifying the system interruptions with the objective to give security to the Internet and to upgrade the nature of administration.

To estimate the blood flow of skin employing infrared light, Photoplethysmography (PPG) is widely utilized. PPG has a lot of inherent advantages like inexpensiveness, non-invasive in nature and acts as a versatile diagnostic tool. It aids in the measurement of blood pressure, oxygen saturation levels, cardiac output and for managing various other autonomic functions of the body. For the effective screening of different pathologies, PPG serves as quite a promising technique. The movement of blood in the vessel which propagates from the heart to the toes and finger tips is reflected by the PPG signals. In this paper, PPG signals are utilized for a single patient who is suffering from cardio vascular problems. For the PPG signals, the Correlation Dimension (CD) and Detrend Fluctuation Analysis (DFA) concepts are applied and then Hilbert Transform is computed to it. After computing Hilbert transform, the signals are later classified with the help of three classifiers such as K Nearest Neighbours (KNN) Classifier, Firefly classifier and Trace Ratio Criterion based Linear Discriminant Analysis (TR-LDA) classifier to classify the cardiovascular risk of the patient. Results show that an average classification accuracy of 95.83% is obtained when TR-LDA is utilized, 93.75% is obtained when Firefly is obtained and 83.59% is obtained when KNN is utilized.

Retinal fundus images are important for the identification and detection of vision- related diseases such as diabetes and hypertension. From an acquisition process, retinal images often have large luminosity, noise and low contrast which seriously affect the automated system of deriving diagnostic parameters. In this paper, a new faster method of correcting luminosity by de-hazing is applied. This method corrects the non-uniform illumination in the intensity domain and then the contrast enhancement is performed along with filtering. Experiments were performed on the publicly available retinal image dataset DRIVE AND DIARETDB1.

Data Mining is the procedure of extracting information from a data set and transforms information into comprehensible structure for processing. Clustering is data mining technique used to process data elements into their related groups or partition. Thus, the process of partitioning data objects into subclasses is term as ‘cluster’. It consists of data objects with un-unified proposition of high inter similarity and low intra similarity ratios. Thus reflecting the quality of cluster depends on the methods used. Clustering also called data segmentation, divides huge data sets into several groups based upon their similarities. This paper discusses a literature study of various clustering techniques and their comparison on key issues to give guidance for choosing clustering algorithm for a specific research application. The comparison is based on computing performance and clustering accuracy.

In a day to day clinical practice a multimodal medical imaging becomes a general part in medical field. Multimodality image fusion is playing a significant role in disease diagnostics and post treatment analysis. An efficient image fusion system based on DFRWT—NSCT hybrid technique is proposed in this paper. The comparative analysis between the conventional and hybrid technique is presented. The proposed fusion system tested and evaluated with various performance metrics. Experimental results explained that the proposed technique achieves a superior visual quality for accurate disease diagnosis.

Recognition of emotions from human plays a vital role in our day to day life and is essential for social communication. In many application of human computer interaction non-verbal communication method like body movements, facial expression, eye movements and gestures are used. Among these methods body movements is widely used because it conveys the emotions of human from all views of camera. In this paper Multi Block Maximum Intensity Code (MBMIC) feature is proposed for emotion recognition from human body movements. The GEMEP corpus (straight and side view) videos are used for this experiment. The 36 dimensions MBMIC features were extracted from the motion of body movements of the human present in the difference frame. The extracted features are fed to the Random Forest classifier to predict the human emotions. The performance measure can be calculated using qualitative and quantitative analysis.

The modernized technology for gathering of data by distributed level collaboration of sensors nodules in the marine environments is termed as underwater wireless Sensor Network (UWSN). This UWSN is relatively rare in the area of research and it’s a trending areas, so it provides more attention to the researcher. The study of underwater environments is the enriching area in the world to explorer the things in the underwater areas. In this security systems like underwater security systems play a major role in getting of secured information from the distributed sensors which is placed in underwater to monitor the underwater environment known to be as UWSN. There are many applications are carrying in the underwater environments. The Applications ranges from oil industry to Habitat monitoring, surveillance security systems etc. The secure data communication in the underwater sensor networks will provide secured communication among the nodes. There are high probabilities for vulnerable of attacks in the UWSN system. The Underwater wireless sensor systems to remove the genuine information and security techniques are important. This could be a major challenging factor for the researchers as well as marine observers to prevent from severe attacks. As the properties and uses of UWSN, various procedures were designed to confirm protected feature of UWSN. But majority of security attacks mechanisms depend on the vulnerability of threats. In this paper, we analyzed the security measures provided by different researchers for vulnerabilities of attacks and also deals with the various security attack like jamming, wormhole, Sybil etc. Those attacks may occurs on various layers of underwater sensors network and we provided the view to minimize the vulnerabilities of the threads in the systems. This paper provides the view to secure an UWSN environment.

Cloud computing has gained status of red carpet in recent years. The only rationale behind achieving this huge applause for cloud is its accessibility in requisite personalized form without harming its effectiveness. Efficiency of cloud computing has became the outcome of scheduling algorithms applied to maintained its potential, high end hardware involved and networks that support this huge infrastructure. This article is focusing on tasks scheduling in cloud computing particularly when tasks are of independent nature. Various techniques are available for minimizing scheduling time of tasks still optimization has scope in this regards. Task scheduling is usually considered as NP-hard problem and meta-heuristic algorithms are treated as one of the best solution in dealing with this kind of problem. There are plenty of meta-heuristic techniques presented as Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO), Genetic Algorithm (GA), Language Championship Algorithm (LCA), Artificial Bee Colony (ABC) to mentioned a few. Comprehensive study and comparative analysis of these diverse types of algorithm in the region of user’s view and service provider’s view is articulated here. This article is focusing on tasks scheduling in cloud computing typically when tasks are of independent nature.

Facial expressions consisting of various emotions play a significant role in interpersonal relations. Emotion detection from various expressions of the face can be performed broadly in three major steps which involve face detection-normalization, extraction of features and classification. An automated facial expression detection methodology has been introduced by the authors in this letter. Here, after face detection and normalization we extract three different types of facial features: Geometric, Texture and Structural. Based on these extracted features we employ SVM classifier to separate the face expressions which includes Happy, Sad, Disgust, Angry, Surprise and Fear. We have applied our algorithm on two databases: JAFFE and COHEN. We have successfully detected over 80% expressions from JAFFE and COHEN database.

Finding and tracking the position of an Unmanned Air Vehicles (UAV) is an important research problem since they are increasingly being used. These devices are equipped with GPS and orientation sensors which are used for tracking. However, data from these sensors can be missing or inaccurate in case of signal outages or other calibration problems. In this paper, we present evolutionary optimization of a rule-base designed for predicting motion models for a Kalman filter that is used to track the position and orientation of a UAV. Results show improved performance in terms of filter accuracy.

With the data being immensely distributed and the need to analyze the data, data mining has gained importance over the years. The data is analyzed to make some strategic decisions and to derive some patterns out of it. Classification algorithms are classical data mining models to excerpt knowledge from bulk amount of data. The focus of the work is on comparison of various decision tree classification algorithms using WEKA tool taking contact lenses dataset. The methods used for classifier comparison are accuracy, mean absolute error and root mean squared error. The outputs are captured using training data set and then compared to understand the accuracy of the classifiers.

Retinal disorders are progressive in nature and remain passive for years together without causing any visual indication of disorder even to the subject themselves. Hence automated and intelligent methods of analysis of retinal scanned images are quite necessary to improve accuracy and detection time to aid in early detection and consequent treatment. This paper presents the findings of a vast literature survey done with respect to automated detection techniques by analyzing their underlying principles and obtained performance results. The entire survey has been done based on two main evaluation metrics namely detection accuracy and time of convergence. This is based on the underlying principle that migration from manual and conventional methods to automated systems is to improve the accuracy by overcoming the errors incurred in manual detection methods and at the same time to reduce the painstakingly long time required in the manual method of observation and detection. This paper proposes a computation complexity reduction mechanism in dehazing by utilizing the convolution properties of deep belief neural networks to train the data sets in the least possible time with improved image quality.

In this paper, we introduce the concept of Complementary perfect hop domination number of a graph. A set S ⊆ V is a hop dominating set of G, if every vertex v ∈ V − S there exists u ∈ S such that d(u,v) = 2. A set S ⊆ V is a complementary perfect hop dominating set of G if S is a hop dominating set and <V − S> has atleast one perfect matching. The minimum cardinality of complementary perfect hop dominating set is called complementary perfect hop domination number of G and it is denoted by CPHD(G). Here, we investigate this CPHD number for some mirror graphs and some special type of graphs.

Water is an important element to survive for humans and other forms of life. Due to rapid industrialization and urban growth, water bodies are shrinking and affected by anthropogenic activities. Remote sensed images serve as an important tool that captures historic data to understand changes in environment. This paper focus of identifying shrinkage that have occurred in lake water bodies over a period of time. Puzhal Lake located in Chennai city is considered for the study. Landsat image of 1999 and 2009 are used to identify the spatial and temporal changes that have occurred in the lake over a period of time. Normalized Difference Water Index (NDWI), Modified NDWI (MNDWI), Normalized Difference Vegetation Index (NDVI) are applied on the data to differentiate water bodies from land cover. Pre-processing of images are done in ArcGIS followed by unsupervised classification. Classification methods are applied on these images to extract water bodies. Topographic maps are also used as reference to validate the results. From the results obtained in the above process, lake regions are identified and pixel count is calculated. Pixel count from two images over a period is used to identify the change in the area of lakes. The results show that there is shrinkage in the area of Puzhal Lake which is chosen as our study area. The results help us to understand change in water bodies and take measures to conserve them.

Content-Based Image Retrieval task is still a challenging problem, Edge Detection method is an important role in image processing and big challenge task in feature extraction which can be fundamental problem in image analysis. Edge detection methods are a strong feature for characterizing an image. In this approach to make robust techniques for extracting edge pixels by edge feature detection, we propose a Multiple Edge Detection (MED) frame work for image re-ranking. In this approach multi-edge detection techniques like Roberts, Sobel, Prewitt, Canny, LoG (Laplacian of Gaussian) etc., of an image which are followed by combined edge feature using gray level as well as shape information of edges map. Our aim is to maximize the image retrieval combined edge detection methods which promote more relevant re-ranking image results. The experimental results proved the efficiency of the proposed method.

Every organization that interacts with some form of customers or users has a feedback system. Feedback provides a vital source of information to the organization about how the end users feel about their service. However, textual feedbacks are very subjective, and to be able to use them for rating or mathematical analysis, we need to quantify the textual feedback. Sentiment analysis can be performed on the textual feedback to attain a quantifiable output. In this paper we aim to design two systems; one that learns from raw text examples by clustering (unsupervised learning) the text samples and then assigning classes to these clusters, and another system that uses this trained data and classifiers to classify new textual data into one of the sentiment classes. For clustering we use K-means clustering method and discuss the performance of the same, and for classification we use two classifiers; K-nearest neighbors (KNN) and Naïve Bayes (NB). Finally, we compare the performance of the two proposed classifiers.

To provide the good range of image, the satellite image is used for road classification which consist very high resolution. A major challenge is that a road network has lot of topological irregularity. There are numerous applications to road classifications namely in the areas of designing emergency rescue systems, updating geographic information systems and roads navigation. This manuscript is deals with the extraction of the road network from the satellite perspectives that have high resolution. This proposed work deals with approximating whether the image pixel is the part of a road or which is not consuming a Convolutional Neural Network (CNN). This proposed work recommends a new innovative tactic for creating data sets for this intricate problem and has accomplished with a viable resolution for the given tricky problem. An attempt with passing Fourier Transform of the input image has contributed to better performance of the CNN in terms of small block size and thereby fast learning of the network.

Most of the complex cellular organisms are divided into genders. Genders are of two types. Gender of an organism would be a male or a female. Each Gender has its own behavioural and physical properties. By behavioural and physical appearance of an individual, one can easily identify the gender of a person. This project deals with the identification of the gender of an individual. Voice is used in this project as an input. The individuals’ voice can be useful only if it is taken in Acoustic form. An Acoustic form of voice is the numerical value for particular speech. These numerical values are used to find patterns of voice of individuals. Different insights are drawn between attributes of voices of people and Machine Learning techniques are applied to get the results from a person’s voice.

Recent Days, automated video surveillance is a major part of security in banks, streets, air ports, railway stations, and crowded areas with multiple cameras. One significant reason to choose automated video surveillance is that, it identifies suspects involved in suspicious activities which will give lead for further investigations. In this work the proposed system will re-identify the suspect face from various surveillance cameras which is been deployed in different locations or positions of street or building, etc. The proposed trained Convolution Neural Network model will extract the facial features. The facial features that are extracted from multiple cameras, are the given to feature ranking algorithm to identifies the frame with the maximum features. As a result, the model will be able to detect the person from multiple cameras which reduces the manual monitoring.

The Biosensors based on optical characteristics are used for different applications and are constantly used to give new technical results. Different types of medium in a Biosensor allows different characteristics with enhanced light reactions through it. In this paper, a class of label-free optical biosensors, are discussed. Biosensors built on label-free methods are widely used because for its identification and security features. There are different scientific methods for different kinds of detections, like proteins, viruses and DNA. The intrinsic adjustability of label-free Bio-sensing can be easily combined with Biomolecular technology. In this paper, the Biosensors are mainly used for WBC Quality Factor measurement in Blood Sample and also to find Cancerous cells present in Skin and also identification of Cancerous Cell from normal Cell based on QF.

In drug discovery, an efficient modelling of the synergies between the existing drugs/compounds and their targets is crucial. The application of In-vitro methods over millions of compounds is tedious and expensive. Virtual Screening, an In-Silico (Computational) technique has become as indispensable constituent of contemporary drug design. This technique executes efficient In-Silico searches over millions of compounds and drastically reduces the time and cost involved in drug design. This work intends to develop a Virtual Screening model for cancer drugs using Deep Learning. For this three Deep Learning algorithms were implemented on the dataset and their performance measures were recorded and compared to the performance measures given by the traditional Machine Learning algorithms on the same dataset. A significant gain in the performance metrics of the model was observed when the deep Learning algorithms were used. The activity of the molecules in the GDB13 dataset was predicted using this model to identify the potential anti cancer drugs.

Glaucoma is a retinal disease by increase in extra fluid in front of eye increasing the pressure in eye leading to total or complete blindness. The gangalion cells in retinal is affected due to increase in eye pressure leads to total or complete blindness. If not treated on time, one can go blind for lifetime. This is mostly found in population with age above 40 years. Glaucoma cannot be cured, but early detection of Glaucoma and proper medication can stop the problem. The proposed work represents different operators of segmentation of image processing method to find earlier detection of glaucoma. Preprocessing methods such as filtering, and image segmentation is used in the proposed work. The various operators of segmentation are sobel, canny, prewitt and Robert. The entropy values for set of 40 images using above operators is evaluated and the operator with highest entropy values gives optimal result. Manual determination and examination of ophthalmic images is time taking and tedious work. Automatic determination and examination of retinal or eye images gives accurate assessment. It aims at determination, diagnosis, and prevention of problems related to glaucoma. The segmentation operators used in the present paper can obtain the best operator that provides optimal result.

Two dimensional tiling of the plane are constructed using a distributed computing model, namely Tile Pasting P System (TPPS) where rules are used to glue the tiles together in a hierarchically arranged membrane regions. The present study focuses on the generation of Homeogonal family of tiling with the bio inspired computing model TPPS.

The abnormal growth of cells known as malignant tumors and also called as breast cancer. These types of tumors will affect other the entire body. Different types of cancer prevailing in the human body, the medical or scientific world are not sure about the exact cause of the disease. In this paper dataset are analyzed by the report by using screening mammogram, malignant neoplasm by using exam reasons and procedure codes. The classification rules are generated to represent the relationship between procedure code and exam reasons.

Nuclear power is one of the promising power sources in developing countries. Because of the disasters that has occurred in Nuclear Power Plants (NPPs), it has become a primary concern for the plant engineers to ensure the safe operation of the plant. The coolant towers are a subsystem of the NPP which is directly linked with the water sources. Faulty operation in coolant tower will degrade the ecosystem and environment around them. Deployment of artificial intelligence, machine learning and anomaly detection algorithms would reduce the human intervention in the plant and also eases the process of condition monitoring. This work gives a detailed empirical analysis of common machine learning classification algorithms with various performance metrics. This work would be of immense help to the plant engineers and the machine learning experts to share their knowledge, which would benefit each other.

Facial Expression gives significant information about emotion of a person. In this investigation area of machine vision, automated facial expression recognition is an important area because of its significance in Human Computer Interaction (HCI). To improve the state of interaction in man machine communication systems, extraction and validation of emotional information by facial expression analysis plays a major role. The proposed method in facial expression identification is dependent on Marker Controlled Watershed segmentation for Features. The granulometry of the image and the first derivative were used as texture features. The artificial intelligent system called Adaptive Neuro-Fuzzy Inference System (ANFIS) is used for finding the facial expressions. The performance of the proposed methodology is validated which yield promising performance showing the effectiveness of the recognition system.

Classification is the way toward finding a model or capacity that portrays and recognizes information classes or ideas, to be ready to utilize the model to foresee the class of items whose class mark is obscure. The objective of classification is to precisely foresee the object class for each case in the information. In sequence database having sequences, in which each sequence is a rundown of the exchanges requested by the exchange time. There is exchange time which is related to every exchange in the sequence database. The sequence classification can be characterized as appointing class marks to new sequences dependent on the learning picked up in the preparation organize.

In Disease Diagnosis acknowledgment of examples is so imperative for recognizing the disease precisely. Machine learning is the field which is utilized for building the models that can predict the yield depends on the sources of info which are related dependent on the past information. Disease recognizable proof is the most pivotal assignment for treating any disease. Classification calculations are utilized for arranging the disease. There are a few classification calculations and dimensionality decrease calculations utilized. Machine Learning enables the PCs to learn without being altered remotely. By utilizing the Classification Algorithm a theory can be chosen from the arrangement of choices the best fits an arrangement of perceptions. Machine Learning is utilized for the high dimensional and the multi-dimensional information. Better and programmed calculations can be created utilizing Machine Learning.

As the driver is the data beneficiary and essential chief in the driving procedure, this examination expects to research a driver’s hazard attention to survey a driver’s wellbeing. We built up a scale for evaluating a driver’s hazard mindfulness, which comprises of four scales: risk attitude, risk perception and risk behavior, and the sensation seeking scale. The markers are named below average records, for example, the general disposition towards obeying rules, forceful infringement and consciousness of safe driving, and so forth. In this investigation, with the end goal to build up a hazard mindfulness show, a study was directed in India. In view of the overview, exploratory factor investigation of the scale uncovered three hazard mindfulness factors (chance state of mind, chance recognition and hazard conduct), likewise named top of the line lists. Aftereffects of measurably breaking down the overview demonstrated that a few drivers in our investigation have high hazard mindfulness. Moreover, a graph was developed dependent on the relapse investigation of a driver’s sensation chasing and chance mindfulness lists. It created the impression that the higher the driver’s sensation seeking, the lower the driver’s hazard mindfulness.

Speech recognition is a major field among the fast growing technologies in engineering. It offers prospective benefits and has numerous applications in various domains. Around 20% of people on earth are affected by disabilities. Several such people cannot use their limbs effectively or are blind. In such situations, speech recognition provides required assistance. With this technology, they can use voice input and operate computer and share information. This project aims in providing assistance to this minority of people. The paper presents a technology that can recognize speech despite varied emotional state of the user. Speech recognition technology permits the computer to capture voice recording using RAVDESS and SAVEE dataset. These recordings are processed and recognized with the help of the speech recognizer. Further, emotions are recognized and provided as output by the system.

In this paper propose the method of image segmentation technique and introducing the classification of segmentation algorithms. We want to implement the part of execution time for this working with symmetric parallel computing using soft computing approach. Computation time is another factor of image restoration process. Restoration of image of body parts is risk in medical science. We want to reduce the execution time of segmentation techniques run with symmetric parallel computing using soft computing approach; using unsupervised segmentation techniques namely image segmentation through K-means, C-Means clustering algorithms and segmentation using histogram technique.

The most widely used procedure for perceiving lump in brain is Magnetic Resonance Imaging (MRI). Due to the increase in the need for an efficient classification procedure, a hybrid mechanism is proposed in this paper. The work discussed in this paper is a blend of Support Vector Machine (SVM) and Improved Kernel Prospective C-Means algorithm (IKProCM). The proposed approach is a collaborative work of SVM and IKProCM, a cross-breed procedure for perceiving brain lesion. The algorithm enhanced the image using enhancement techniques to improve contrast, resolution and removal of noise. Skull extraction is achieved by applying morphological operations and thresholding mechanism. IKProCM clustering mechanism is used to segment the image to identify the expanse of interest in the MR images. The Size Zone Matrix (SZM) is applied for pulling out required features found in the MRI image. Later SVM procedure segregates the MRI images as defective and non defective.

Day to day life security is very important aspect likewise one of the famous technique for security in present technology is steganography which will provide a new dimension to the data and that particular data will be in another form so the sender and receiver only knows the exact data others will be able to see the data but understanding of the data is very less compare to other technique. To strengthen the technique conversion of english language text into objects are another dimension which gives us more confident on data transfer by using proposed algorithm. This algorithm is applicable for both sender as well as receiver side and the system supports only english letters and special characters alone.

Reliability of software systems are less when compared to hardware systems. The optimization goal is to minimize the variance of the reliability estimator and to improve the accuracy of the software. Testing is a means of achieving reliability of the software. There are different kinds of testing to enhance the reliability. This research work improves the software reliability through adaptive testing. When software tested, the software tester uses different kinds of test cases for testing the software and each test case is obtained from the feedback of the previous testing process. Along with adaptive testing regression testing is also used to enhance different types of software testing for fixing the bugs in functional and non-functional areas.

It is a well-known fact in the world that the glaucoma is the second largest disease which is affecting the human beings in the world. Proper care has to be taken to avoid this at an early stage as this would result in the loss of vision in the humans. This occurs due to the increase in the pressure in the eyes, where it bursts the nerve fibres leading to the vision loss. If the patient goes to the doctor, it is an expensive treatment. Hence, we are devising a low cost module method of detecting the primary glaucoma in the humans using their fundus images. The images of the patients will be taken by the fundus camera, analyzed and a info is given to the patient that he/she is affected with the disease. Once the person comes to know that they are affected, then proper diagnosis can be done by consultations from the hospital experts. The method of detecting the primary glaucoma is being presented in this section using a revised fuzzy-c means algorithm clubbed with morphological operators. CLAHE concepts are being used for the pre-processing and the edge detection is done using canny operator’s method. The segmentation is done using fuzzy and finally the region of interest, i.e., the cup and the disc areas are found out from which the ratio is computed, from where the disease can be detection seeing the ratio. The simulation results shown the effectivity of the method proposed by us in this research work.

Apart from being a rich source of nutrients, medicinal treatment with medicinal plants hold a strong ground because these plants seem to be safe with least aftereffects. Medicinal Plants or herbs possess a special quality or phyto_property that enables them to combat multitude of health issues. This paper is a noble attempt to unearth these disease curing properties of medicinal plants from biomedical literature. The proposed architecture discusses a text mining based literatures mining technique to derive information between biomedical entities like properties of medicinal plants (e.g. anti inflammatory, antioxidant) and disease (e.g. arthritis). Unlike exiting heuristic attempts involving syntactic patterns, co-occurrence analysis, we propose a Verb Between Entities (VBE) algorithm which attempts to discover relationship between entities by analyzing the main verb between them. The framework also incorporates UMLs thesaurus to help identifying verb phrases which includes functional concepts in the course of verb analysis. Performance of the framework has been evaluated on multiple datasets and the outcomes indicate that the recommended framework is more effective in identifying functional semantic relations as compared with the other relevant methods.

Leaves can be of more importance in the context of recognition. Creating a model will help in recognizing them for different applications like Medicine and Herbal analysis. The leaves has features that can be statistical based or at high level. It can include edge and any features built over pixel level. Edge identification is used for data extraction, image segmentation and data compression. In this paper Statistical features for set of leaves are identified, Leaves are classified using multi class SVM and Edges of a leaf is identified by using canny, prewitt and sobel edge recognition techniques base on various Gaussian mask. Convolution Neural Network is used to classify the given image under 14 categories. From the trial results, it is seen that canny edge identification method gives preferable outcomes over prewitt and sobel edge recognition strategies. The paper also provides directions for using Convolution Neural Network for Leaf recognition.

India is one of the mega bio-diverse countries in the world harboring noteworthy ecological wealth and has a huge scope of vegetation types. Biodiversity data, especially in the vegetation types, stays isolated and extensively scattered among heterogeneous information systems. Semantic web technologies in artificial intelligence field incorporate Ontologies are standard knowledge representation mechanisms which are based on description logic. Given the diversity in biological species and the related rich wealth of information, it is required to build an ontology capable of representing and sharing this information in an automated fashion. To adapt these challenges for enabling semantic interoperability of vegetation type data, this paper presents the initiative of developing Indian Biodiversity Ontology. The primary thought process behind this development is to facilitate the computational mechanisms in support of a comprehensive knowledge base for the conceptualization of Indian biodiversity data at one place.

Information retrieval systems are those systems which work upon some set of searching algorithms that enables the system to retrieve the desired information from the system. There are several techniques that are already available like, algorithms based on directed trees, fuzzy clustering algorithm, and several divisive algorithms. Different algorithms provides partitioned results, but the ensemble clustering combines the multiple partitioned results and provide a better result to the user. The motive behind combination of multiple partitions is to enhance the quality of each every cluster and improving the service of the retrieval system. For this purpose, we introduce a new ensemble approach for effective information retrieval through clustering process over the documents or online contents. Ensemble clustering creates several smaller data clusters from a big data cluster and then transforms those clusters into consensus matrix form which results into very efficient and better performance.

Image compression finds an exhaustive application in the field of medical image file storage. Image files contain considerable amount of redundant and irrelevant data and suitable image compression algorithms can be used to eliminate this. In this paper, Coiflet wavelet is used to perform the compression for computer tomography image of coronel view. The CT images are encoded using the different types of encoding techniques .The performances of compression are measured in terms of PSNR, Compression ratio, Means square error and bits per pixel. The Image texture features are extracted, and it is proven that after compression and further it can be used for classification of images.

Image dehazing is one of the classification of image restoration methods in image processing and is gaining wide spread attention in recent times to dehaze the image/video in real time. They find an ominous utility in transportation systems and is of great aid to commuters in highly hazy urban as well as hilly terrains. They are also being widely researched to develop real time dehazing methods in aircraft to provide uninterrupted view through haze and mist. Dark channel prior techniques have been widely used in image dehazing with more research being done to improve the quality of image as well as to reduce the computation time. This paper proposes a computation complexity reduction mechanism in dehazing by utilizing the convolution properties of deep belief neural networks to train the data sets in the least possible time with improved image quality.

Change in opinion on a particular topic of interest of people can be studied by analyzing their activities on social media. This can be achieved by gathering user opinion and mindset from their activities in social media, and display results by mapping their emotions in a series of interactive data visualizations. The proposed work shows the sentiment and mindset of people varies with time. This idea would be extremely useful, or provide interesting insights. Analyzing the success of a marketing campaign, predicting the result of an election, market research are some applications of this idea. This is quite a new area, and many researches are going on in this field.

Smart grid plays a vital role in electricity supply network by using digital communication technology. It is capable of easily detecting and reacting to the local changes in the usage and thereby working efficiently. Against all the smart grid monitoring system there are many cyber intrusion which creates severe threat to power to power system operation and false data injection is one among them. In this paper we discuss about the different techniques. Kalman filter, state estimation, bad data detection test, Phasor measurement units (PMU) installation are used to detect and minimize false data injection attack.

The historical things and the way of traditional lifestyle as well as medicine secrets were available in palm leaf in order to get those details from the palm leaf through the trained data set has been used to identify the Tamil character from the palm leaf, then pre-process the character segmentation process follow and then feature extraction take place to identify the correct letter for that process make the comparison with the database. Finally, Identify the exact Tamil character from the palm leaf which gives support to damaged palm leaf also mainly concentrate in false acceptance rate as false rejection rate of Tamil character to give more accurate results.

Leaf species identification leads to multitude of societal applications. There is enormous research in the lines of plant identification using pattern recognition. With the help of robust algorithms for leaf identification, rural medicine has the potential to reappear as like the previous decades. This paper discusses Prewitt k-NN for leaf species identification from white background. Variations of the model over the features like traditional shape, texture, color and venation apart from the other miniature features of uniformity of edge patterns, leaf tip, margin and other statistical features are explored for efficient leaf classification.

Topic Models are representations of the given text. Topic models are unsupervised in nature because they totally depend on word distributions. Few generative topic models obtain the topics proportionate to the richness of the given text. Applying deep learning for identifying the worthy distributions in a rich text is quite helpful for generation of quality topics. This paper discusses the idea of learning deep topics of interest from scientific research articles. The learning is handled using deep stacked auto-encoder with three hidden layer stack coupled with generative topic models. The deep learning framework is explored excluding and including back-propagation and is tested upon both LDA and HDP as foundational topic model. Experiments conducted over the data-set of research articles from top bio-medical journals reveal better topic coherence for learning deep topics.

Biology is truly an area of unlimited possibilities for designing artificial intelligence systems that are capable of performing efficiently and independently in unfamiliar and dynamic ecology. It is challenging to refuse the enchantment of designing artifacts and showcase lifelike intelligence thus needing methods for prediction, design, optimization, control, security etc. and these needs can be addressed by using the Computational Algorithms that are motivated by biological mechanisms which are widely known as biological inspired algorithms. In this survey paper we have discussed working of some of such algorithms like Ant colony, Bee colony, Bat colony, Cuckoo and Firefly algorithms by using which many current world problem scan be addressed, adopted, designed, implemented and optimized.

The increased use of network usage for performing online shopping and globalization has created a necessity of credit card usage throughout the world. Credit-card fraud has lead to considerable loss to the merchants and card users of about billions of dollars. Machine learning algorithm development paved the way for finding the fraud in more sophisticated ways but practical implementations are rarely reported. This paper proposed a method for deploying the fraud detection system which can efficiently find the fraud in the provided transaction. SARSA algorithm is used here for prediction of threshold value to detect the fraud accounts. Reinforcement algorithm is used for detecting the frauds. Finally, the benchmark analysis is performed to find the effectiveness of the proposed algorithm. The weight calculated from SARSA reinforcement learning algorithm is given to the random forest method to increase the accuracy the prediction of credit card defaulters.

This work presents an investigation on the application of three deseasonalization models to monthly seasonal streamflow series forecasting: seasonal difference, moving average, and padronization. The deseasonalization is a mandatory preprocessing step for predicting series that present seasonal behavior. The predictors addressed are the linear periodic autoregressive model and an artificial neural network architecture, the extreme learning machines. The computational results showed that the padronization is the most adequate to deal with this problem.

Texture is a prevailing tool for feature extraction. This paper determines Local Painted Texture Pattern (LPTP) which gives the local chromatic texture data of an image. The data is extorted independently from the co-related pixel values of assorted channels. It affords the exceptional channel-wise data and it’s related with neighboring pixel data of opponent space. A feature level fusion framework is used to merge Colored Pattern Appearance Model (CPAM) along with LPTP in natural and face databases, which shows significant improvement. The experimental result by using this descriptor presents significant development from the related works for content-based image retrieval and face recognition.

The medical science is practising on the determination, treatment and aversion of the different disease. An AD is a type of neural dementia that causes a human body-brain especially memory, cognitive skills, and other parts of the brain. The motivation behind this examination is to propose an efficient algorithm structure using deep learning architectures methods and techniques for the perception of Alzheimer disease. In this study, the deep learning architecture structure is created using data normalization, generalized linear neural network (GLNN), regression techniques (softmax), K-means clustering. The detection of Alzheimer’s is done using the combined dataset of the spinal cord and brain. Compared to the previous workflows these methods are capable of detecting the Alzheimer at the minimal timestamp.

Blogging is a community-based effort and blogs carry rich information on a specific context. It serves as a platform for individuals to voice their knowledge, skills, thoughts, and feelings which become ingrained in our day-to-day lives. Improved blog classification using multi stage dimensionality reduction technique is proposed. A blog can be easily classified using the tags provided by the blogger. The blog contents can also deviate to a related topic as it can be written by novice content writers also. In the proposed method the blog contents are represented as a collection of text features. The blog representation and preprocessing stage transforms the blog contents into feature matrix. In the feature reduction stage the vital features which are pivotal in its class identification are recognized using term weighting technique and information gain. The pattern reduction stage uses leader algorithm to select only marker data from the data clusters and minimizes the number of patterns from the large amount of blogs. The reduced patterns are fed to compact pattern reduction and classification stage where a novel pruning algorithm N2PS is used to optimize the classifier by reducing the features of dataset. The multi stage dimensionality reduction technique improves the speed of the classifier. The proposed method is implemented and validated on a live dataset. The results outperform the existing methods in terms of accuracy and training time.

The proposed model uses Content Management System which allows user to share and communicate information under any platform. Based on the interest of the student the content can be created and shared inside the platform which works dynamically for providing the data to the end user. Student generated content provides a new way of understanding for both the students and teachers. This system adopts a Categorised way of providing the saved data to the students, which explains as what type of data are need to be provided to which set of students. Analysing the students based on the level of domain knowledge helps in evaluating student interest in that specific subject. The department administrator can update the news, events to their students and teachers. Articles can be written based on the topics which are categorized. Subject materials are uploaded by the teachers and stored as K-Store. The tf-idf is used to fetch the data from the database and can show the relevant data which the user asks. The uses of psychometric analysis helps us to understand the level of the users and show them what type of data should be displayed to the users based on high, medium and low level of understand of the subject.

This Paper Presents the Development of voice control and auto control home automation. The system has been design to control all home electronic devices. Amazon echo application has been used voice recognition and process the voice input from the smart devices. In this paper, the voice input has been captured by the android phone app and send to the Arduino micro controller using Bluetooth. Arduino Bluetooth module received the signal. And under the Arduino code processing the signal and control the home electronic application. This system understands Multilanguage. The proposed system intended to all electronic devices (TV, Projector, AC Remote controlling system). Another one is auto mode. This system has automatic mode, LDR sensor calculate the room light level and temperature sensor calculate the room temperature and control the room light and fan/AC. We have demonstrated up to 30 m of range to control the home application. This system has some home security process, main door ask the password user told the correct password then door is open, if user told the wrong password then this system take the photo and send the database. The paper also analyzes how Deep learning can be used for classifying the home automation data.

Recently Cognitive Radio Networks (CRN) are used for improving the efficiency of spectrum usage. It can select spectrum and operate in a adaptive manner. The spectrum sharing has to be allocated and used efficiently for knowing the performance of CRN. The paper analyzes the different spectrum techniques with its merits and limitations. The existing works namely centralized, distributed, cooperative and non cooperative schemes are investigated. From study it can be concluded that fuzzy based approaches in CRN are more efficient.

In the progress of 5G wireless cellular network the heterogeneous network plays an important role. In this paper energy consumption is investigated. Fırst, the placement of Small Base station with Macro Base station which reduces the power consumption then by applying Genetic Algorithm to reduce the power ingestion by switch off the Base station depends upon the load. Before that, the path loss (Okumura–Hata) which is an important parameter in the wireless network can be calculated using the different components like various parameters such as transmitting power, height and distance. Simulation results show that the optimization algorithm achieves nearly optimal performance.

The proposed Maximally Stable Extremal Regions (MSER) algorithm extracts stable connected component of an available set of gray levels in an image. The maximal intensity regions may appear, grow, and merge at different intensity value of thresholds. The Stability achieved through a finding of extremal regions whose support of region is virtually unchanged over a range of thresholds selection. Therefore the regional intensity variation of cyst makes the MSER algorithm perfect, to do cyst boundary extraction in digital dental x-ray images. The results of the MSER based cyst segmentation in dental x-ray images show that there is a significant correlation found between the cystic region extracted by the medical experts .The cystic area segmented by the proposed MSER method results in extraction of the dental cystic boundary very efficiently and accurately.

Salient object detection in an image attributes to finding an object which stands out from its neighbors. State of the art approaches in salient object detection have used learning-based methods for predicting saliency maps. Typically the features from the images are extracted using CNN architectures as they have become influential in computer vision tasks. In this paper, a bottom-up approach for salient object detection in images is described. Densely Connected Neural Network (DenseNet), a recent CNN architecture which has shown significant improvement in classification tasks, has been used for extracting features from the image. DenseNet has strengthened feature propagation, reduced training parameters and also has a lower error rate compared to other CNN architectures. Features from DenseNet have been used to predict the saliency maps of the images. The experimental results show significant improvements from previous works on saliency.

Natural Language Generation takes a key role in presenting data as text or speech. The translation into a natural language from semantic representation is similar to Neural Machine Translation. We use a similar methodology known as Seq2Seq modelling for generating natural language. The usage of common semantic representation such as Abstract Meaning Representation allows adding naturalness to the generated sentences while being domain neutral. Recurrent Neural Network based autoencoder learns a hidden representation from semantic input which is then used to generate natural language. Long-Short Term Memory while in theory being capable of learning long-term dependencies fails to capture the correct information required for generation. We introduce attention mechanism as a resolution to improve capturing contextually important information. The resulting model has a significantly improved accuracy.

Fundus camera captures the posterior portion of the eye. Extraction of Optic Papilla, Optic cup, Blood vessels and Macula from the captured fundus image is done by using image segmentation algorithms for diagnosis purpose. Diseases like glaucoma, diabetic retinopathy and retinoblastoma adds artefacts in the fundus images. These artefacts makes the segmentation a difficult task. When conventional segmentation algorithms are applied for the extraction of exudates, it also identifies the regions belonging to optic papilla. Even though optic papilla is a single region, the presence of optic nerve head occludes a portion of optic papilla that leads to generation of isolated regions. These isolated regions of optic papilla are false regions which have similar properties of exudates. Hence there is a need to exclude the false regions and segment the exudates for quantification which is used for diagnosing diabetic retinopathy. To achieve this objective, a region selection algorithm using Euclidean distance measure is proposed in this communication. After selecting the regions belonging to optic papilla, it is subtracted from the segmented image to retain the exudate regions. The discarded region can be used for segmenting optic papilla accurately which is an important for diagnosing glaucoma.

The reputation of the vehicle in the network plays a major role in providing reliability of the information being transmitted by the vehicle. Vehicles can earn a reputation score based on the trust worthiness of the message being transmitted and their level of acceptance in the network by the peer nodes. When the messages are transmitted along with the rating earned by the vehicle, the messages are accepted with a lesser rejection rate. These values of reputation can be made a continuous defining the performance of the vehicle over a short period of time. The proposed approach is to design a continuous rating scheme for the vehicles to offer better authenticity and reliability to the nodes in the network.

Reliable personal authentication system is essential for social, financial and political structures of today’s human life style. The advent of biometric technology has revolutionized personal authentication system to meet the current requirements through biometric modalities in a reliable, accurate, rapid and user-friendly way. However, there exist a number of unresolved issues for the biometric systems related to data, system design and algorithms. This work focuses on exploring features from dorsal side of the hand region known as finger knuckle surface for reliable personal authentication. This paper illustrates design and development of an integrated finger knuckle biometric framework using multiple units of finger knuckle surface and multi-algorithmic parameters for robust and accurate personal identification. This novel integrated approach known as Multi-Faceted and Multi-Algorithmic Framework (MFMA) for authentication using finger knuckle surface. This MFMA framework simultaneously acquires multiple instances of finger back knuckle surface, extracts multiple features using three different categories of algorithms, viz., angular geometric analysis, transform based texture analysis, statistical analysis and integrates the information derived from multiple algorithms using decision level fusion implemented based on Bayesian approach.

Attention deficit hyperactivity disorder is a neuropsychiatric disorder. It t can be seen as a disorder of life time that occurs in preschool age and continue fully or partially throughout the adulthood. Diagnosis of this disorder is usually not done in early age and this lead to delay in diagnosis and treatment. There are certain differences in condition as well as factors of paediatric male and paediatric female ADHD. Only few researches have done to point out the paediatric gender differences. So it becomes necessary for a good method to diagnose the ADHD in early years for overcoming the negative effects of this disorder. The foremost important part in the diagnosis of ADHD is to find out the exact area of the grey matter (gm), white matter (wm) and cerebrospinal fluid (csf). In this proposed paper, the differences between paediatric male and paediatric female taking cross sectional views of MR images are analyzed in detail. A novel hybrid technique, called Super Spectral Fuzzy C-Means (SSFCM) is proposed. In this new method, the optimal fuzzy is used for clustering and spectral segmentation is done. For the classification purpose, the Artificial Neural Network (ANN) technique is used. The experiment is performed on the cross sectional views of paediatric male and female brain MR images, collected from various Imaging centers. The proposed method is compared with existing techniques and results showed that this new one is accurate and faster than existing methods.

Hand gesture recognition is one of the most viable and popular solution for improving human computer interaction. In the recent years it has become very popular due to its use in gaming devices like Xbox, PS4, and other devices like laptops, smart phones, etc. Hand gesture recognition has usage in various applications like medicine, accessibility support etc. In this paper, we would like to propose on how to develop a hand gesture recognition simulation using OpenCV and python 2.7. Histogram based approach is used to separate out the hand from the background image. Background cancellation techniques are used to produce optimum results. The detector hand is then processed and modelled by finding contours and convex hull to recognize finger and palm positions and dimensions. Finally a gesture object is created from the input which is then used to recognise the count of fingers.

The main aim of the proposed work is to develop a three module face detection and recognition system. This system is mainly centralized on distinguishing the full frontal face present in the encompassed digital image or video. Our system assists the user to detect faces appearing in a video/webcam. In this process, the image fed as output by the user, is verified by referencing it with a federated database. The database assimilates a good deal of digital images from all the authorized users, who can pass through the system. The metadata is trained for effectuating the system’s purpose. This trained data is then used for recognizing the face present on the webcam or the image. If an image corresponds to the fed image, then the suitable data will displayed. This system is implemented with the help of openCV library. This library extends a FaceRecogniser class, which contains algorithms for performing the recognition process. Our system uses the local binary patterns histogram for achieving the result.