Emerging Research in Computing, Information, Communication and Applications

Blockchain is an emerging technology, which offering numerous opportunities to develop decentralized and distributed digital services by ensuring privacy and transparency. It has mainly concentrating on the legal and technical issues rather developing advanced digitized services. In this article, we make use of the smart contracts with Blockchain to design the secure electronic voting system. The aspect of privacy, authenticity, transparency and security is a threat and challenging in the traditional voting systems. In general, mostly elections is based on the centralized infrastructure consists of central entity that maintains over all the voting process. The major pitfalls in the existing E-voting infrastructure is with an entity that has full influence over the system, it is feasible to modify with databases of considerable opportunities. In addition, the paper based voting systems are assisted by Electronic Voting Machines (EVMs) have multiple vulnerabilities, which can be caused to election rigging, fraudulent intent of the third party entities and government. The decentralized public Blockchain technology might offers a scalable solution to current voting systems by providing trust based and fraud proof digital voting.

Glioblastomas are evaluated by neuroimaging techniques like Magnetic Resonance Imaging (MRI) and Computed Tomography (CT). Each imaging technique represent a different characteristic of the Glioblastoma. Although, a single imaging modality is not enough to confirm the presence the tumor. Experts analyze each of the images independently and then make decisions on the location of the tumor. Early detection and accuracy of diagnosis depends on expertise of doctors and quality of imaging technique. Multimodal Image Fusion is a process of fusing images obtained from different modalities to a single image. It supports in early diagnosis, treatment planning or image guided surgery. However, it introduces blocking effect, noise and artifacts in the fused image, thereby reducing the image quality and making it difficult to assess the tumor location. There is a need for an enhancement technique capable of improving contrast, structural information, entropy, peak signal to noise ratio. Contrast Limited Adaptive Histogram Equalization (CLAHE) is a widely used enhancement technique. The major drawback of CLAHE, it only enhances the pixel intensities. Moreover, CLAHE requires operational parameters like clip limit, block size and distribution function, and initializing them is always a challenging task. Particle Swarm Optimization (PSO) helps to choose the CLAHE parameters, based on a fitness function representing entropy and edge information of the image. The proposed technique provides improvement in visual quality of the fused images.

With the technological advancements in the domain of computer science, Medicinal and engineering fields, Intelligent Assistive Technologies (IAT) have been developed to mitigate the burden of caretakers and patients of Alzheimer’s. A wide variety of IAT has been developed. Caretakers and patients of Alzheimer’s will be much relieved if there are tools that assist them in performing their daily activities like bathing, cooking, hand wash, brushing and medication aids etc. Further, a safety system that ensures the safety of the patients like fall detection, gas leakage detection and fire detection etc. is also needed. This paper mainly reviews various existing IAT for medication aids, brushing and fall detection. Thus, this provides new research challenges and future directions for other researchers working in this area.

Nowadays, providing security for the data is the major concern especially in the cloud storage unit. Therefore, a security aware mechanism has to be developed to improve the security using ontology concepts in the cloud storage most of the existing servers used in the networks are used for storing and accessing the data, but these servers are built in a centralized manner, providing security for the data is a key requirement for the servers. The major threat in the existing servers is the data stored in the servers are subjected to hacking. This paper presents a technique to efficiently store the data in cloud by using concepts of ontology and Secret Sharing Scheme (SSS) has been introduced in the proposed framework to protect the data. This paper presents a model whereby the data can be securely stored for the longer duration in the distributed environments. Ultimately the proposed security scheme has achieved better performance and outperformed the conventional technique.

Driver assistance technologies such as self-driving, automated parking, cruise control have improved exponentially and have much more relevance today but this inadvertently leads to negligence and inattention so it’s critical to have the tools to combat this cause. Along the same lines our paper presents a method to monitor the driver’s activity and continuously look for red flags such as distracted driving, overuse of mobile phones while driving, drowsiness, sleeping. Our objective was to detect these red flags in real time alert the driver about the same. This is achieved by using a camera-based system and the MobileNet neural network which has been fine-tuned on our self-made dataset. Further we draw comparisons against common neural network architectures for this task on the basis of compute efficiency and processing speed.

India is a nation which is known worldwide for its agriculture which helps in developing socio-economic conditions of the nation. This sector helps for the nation's Gross Domestic Product (GDP) as it is providing employment to half of the nation’s population. Due to the development in different sectors, the contribution of the agriculture sector towards the Indian economy is decreasing year by year which is leading towards the descending lifestyle of farmers of India. This predictive model for crop production using different machine learning algorithms is aimed to help farmers to grow different kinds of crops taking into account the weather, statistics of crop production in past years in specific regions such as state-wise or district wise. This model will be using different classification algorithms and showing which crop is most favorable to grow to give proper payback for the crops for the farmers. It will take into account the year in which the crop is to grow, season, area, state, district and will provide the result as to how many percentage of chances that in given crop production will be more that will help the farmers to decide that particular crop.

The proposed system, SPD [Speed Bump and Pothole Detection], aims at detecting and notifying a user of upcoming road anomalies such as speed bumps and potholes in real time. The system collects live sensor data from a detection hardware module which comprises a Raspberry Pi, GPS module and a 3-axis accelerometer. This sensor data collected simultaneously from multiple users is then pushed on to the Cloud. It is then pre-processed to get the required sequences of data points and then fed to a 3-class classifier which uses a Recurrent Neural Network Model (Long Short-term Memory with Adaptive moment estimation [Adam] optimizer). The system classifies the given sequence of data points as ROAD, SPEED BUMP or POTHOLE. The system was validated using 20 km of real-world data and achieved an average accuracy of 80%.

The extraction of embedded knowledge from data is the fundamental task of data mining (DM). This extracted knowledge should be understandable to the end user. Earlier, statistical methods were utilized for the purpose of knowledge extraction. Later, semi-automated DM methods were developed through technological advancements. As data increased, these semi-automated DM techniques became ineffective. Hence, currently for synthesizing knowledge from data, fully automated DM techniques are being utilized. The early yield predictions of important crops such as soybean benefit the agriculture stakeholders by increasing their profits. This work deals with the prediction of soybean yield as high yield or low yield using ID3 algorithm. The result of ID3 was compared with naïve Bayes (NB) classifier. Results demonstrated that the ID3 algorithm performed with improvements of 7% when compared to the NB classifier.

Missing values in datasets are caused by various reasons. These missing values in datasets adversely impact the performance of data mining (DM) algorithms. These values can be ignored if there are large number of instances in the dataset. However, deleting the records containing missing values in smaller sized datasets can lead to improper classification or predictions by the data mining algorithms. Several methods for finding missing values are present. Commonly used methods like replacing the missing values by the mean of the column, repeated value, are present. However, these techniques don’t approximately forecast the missing values. In our work, we have considered regression-based methodology for predicting the missing values. Results on stock dataset demonstrated that the polynomial regression (PR) model exhibited better performance when compared to linear regression (LR), quadratic regression (QR), pure-quadratic regression (PQR) and interactions regression (IR) models for prediction of missing values.

Intrusion detection is very much needed in today's scenario in many application areas of network security. The development of effective, efficient, and flexibly adaptable security mechanisms has become more critical in today's network. The traditionally existing security mechanisms including NAT, firewall, user authentication, and information encryption, etc., are not enough insufficiently covering today's network. In this research, proposed the improved stacked sparse auto-encoder (ISSAE) method for dimensionality reduction and classifiers used for intrusion detection. We have used the principal component analysis, support vector machine, random forest classifiers on KDD-cup 99, NSL-KDD, UNSW-15nb, and NMITIDS datasets for result analysis of precision, recall, F-score, and false-positive rate. With minimum cost function and training epoch can get better results compared to existing methods. In this research best result obtained through random forest method with a less false positive rate compared to other classifiers.

Mobile Ad-hoc Network (MANET) is a significant type of wireless sensor networks, executes on a resource restricted environment. MANET turns into an emerging technology for its special features. Secured data transmissions in MANET can be established by dynamic Node clustering algorithm. Selecting an efficient Cluster Head (CH) node is a challenging issue in MANET due to finite battery power supply. Cluster generation is a precious task in terms of power distribution of nodes to the clusters. Every cluster having one or many selected cluster head(s), all cluster Heads (CH) nodes are interconnected to form a networks and broadcasting the data. The Cluster Head (CH) node is authorized node for scanning a node performance and other security related functions. A Cluster Heads should be capable of functioning under a limited energy mode and resource constraint environments, malicious nodes can drain energy rapidly and reduces the total lifespan of the networks. In this paper proposed an efficient algorithm for selecting a Cluster Head (CH) nodes by using K means algorithm and Weighted clustering Algorithm (WCA).This algorithm aims to elect a node having high Sustainable Cell (SC) rate as a cluster Header (CH) A Sustainable Cell (SC) rate will be computed by all nodes individually based on the important quality of a node, for instance energy utilization, degree of a node, remaining energy, mobility of node and distance between the CH and base stations. The proposed method also implements an algorithm for removing malicious nodes from the network and increasing the MANET security. By using high sustainable cluster head, this proposed algorithm will improve the stability of a network and network throughput.

Liver diseases turn out to be lethal when detected late. This creates a dire requirement of an efficient and robust system to defy death with early detection. With machine learning setting to grow its roots in the healthcare industry, it would be a good choice to opt for it to overcome this problem. This study aims to tackle the same by applying selected machine learning algorithms on the Indian Liver Patients Dataset (ILPD), the accuracies of which are finally compared to arrive at the algorithm yielding the best results. For robust results, extensive preprocessing involving sampling, normalization, and particle swarm optimization (PSO) is performed on the data. This proceeds in J48 yielding the best result in classifying patients and non-patients of the dataset. Further, an etiological survey depicting the effect of various attributes on the dataset is presented.

Big data is accumulate and development with the facilitate of disseminated encoding structures like Hadoop which handles the new encoding prototype. Big data can handle huge quantity of information with thousands of computers in clusters. Through the condition of compute of different possessions in public cloud, people and institute are clever to outsource their statistics storage and services. Conversely, the information owners are disturbed with security of their outsourced statistics. The Cloud Security Alliance has recognized and announced different security confronts with reverence to big data security and safe estimation in disseminated encoding structure veracity and immediate security. There are important investigate assistance establish in the survey. Conversely, a complete approach that provides to the security wants of the cloud-big data environment to stimulate of the different afore reveals dispute is lost. Consequently the plan of this proposal is to examine, design, and execute method for safe computation, big data privacy, big data security, and efficient statistics administration. Individual challenges are big data privacy, infrastructure security, big data security, and data management. Authentication using username and password which contains the user data such as identity number, data provider name, date and time of uploading the data is also used to recognize the person who uploaded the data into the big data cloud. It identifies the origin of data and someone who downloads the data could know the person who created it. A variety of cryptographic methods used for encryption are Attribute-Based Encryption as Scheme 1 , Advanced Encryption Standard as Scheme 2 , hybridized Attribute-Based Encryption and Advanced Encryption Standard as Scheme 3 , and the proposed algorithm named Secure Dynamic Bit Standard as Scheme 4 which offer high security for the data stored and access by the end user. The Secure Dynamic Bit Standard algorithm provides the security during two dissimilar keys such as master key and session key produced by cloud service provider. The Secure Dynamic Bit Standard algorithms enclose the three disparate key lengths such as 128 bits, 256 bits, and 512 bits. The key length of the master key and session key randomly creates through the encryption process. This method will minimize the maximum number of unauthorized and unauthenticated users in the big data cloud. It also minimizes the maximum number of assault in the big data cloud environment such as collusion, brute force attack, and Structured Query Language.

In India, most of the people are dependent on agriculture. The raw materials obtained from the agriculture are served as food for many people. The crop plantations are being destroyed because of the two main reasons: (i) The natural destructions such as drought, flood, famine, and earthquake. (ii) Pest and pathogens. About 98% of the destruction in crops are caused by pathogens and pests. The remaining 2% of the destruction is due to natural disaster in the surroundings. The rural farmers are severely affected by the crop production problems. In crop's life cycle, leaf plays a major role in getting the information about the growth and production of the plant. In this paper, the proposed system works on the preprocessing of the dataset. The leaf images are collected from the plant village dataset. The feature extraction is applied to the images during the data preprocessing stage. Convolution neural network (CNN) is used for the classification and detection of diseases. The recommendation of pesticides and fertilizers is done by using TensorFlow technique. The convolution neural network with various number of layers is used for training the model, and GUI screen serves as a user interface.

Groundwater has always been the primary water resource in arid and semi-arid areas. Monitoring water table fluctuations is essential for predicting the groundwater levels to outline the future needs. In this study, a detailed analysis is carried out on the prediction of groundwater levels in Karnataka. Nonlinear data-driven models (i.e.; random forest (RF) and gradient boosting (GB)) along with a variant of standard RNNs, Long Short-Term Memory (LSTM) was proposed to predict groundwater level variations. The prediction ability of these models was probed and evaluated using yearly groundwater level data scraped together from observation wells located in various districts across Karnataka in India. The statistical parameters: correlation coefficient (R), Mean Square Error (MSE), accuracy, precision, recall, f1 score, and support were used to assess the performance of these models. These evaluation measures emphasize the capability of these models to keep up with the shift in groundwater levels.

In recent times, social networks are actively emerging as an encouraging platform for democratic discussions on social, media, and political issues. However, the growing popularity of social networks leads to a surge of concerns on the quality of the shared data. Initially, bots are created with good intentions. Nowadays, bad bots are created and powered by AI to perform malicious activities in social network sites. Statistics say that 94.2% of the entire websites in the world are being affected by the Bots. They are more sophisticated, intelligent enough to mimic like humans and can change their signatures dynamically. It's hard to detect such kind of bots with the existing techniques as the new bots keep evolving. Social networks are growing exponentially. As a result, big data is growing unconditionally. So, we need more computationally efficient algorithms running at a higher detection rate to handle big data. For our study, we used the real-time dataset from the bot repository released by Indiana University. In our work, we propose a new custom bot classifier algorithm to predict the bot accounts. The algorithm detects the bots with an accuracy rate of 96.13%. Furthermore, we made a comparative analysis of our algorithm with existing machine learning algorithms using various classification metrics.

Our recent study using historic data of paddy yield and associated conditions include humidity, luminescence, and temperature. By incorporating regression models and neural networks (NN), one can produce highly satisfactory forecasting of paddy yield. Simulations indicate that our model can predict paddy yield with high accuracy while concurrently detecting diseases that may exist and are oblivious to the human eye. Crop Yield Prediction Using Regression and Neural Networks (CYPUR-NN) is developed here as a system that will facilitate agriculturists and farmers to predict yield from a picture or by entering values via a Web interface. CYPUR-NN has been tested on stock images, and the experimental results are promising.

Human Activity Recognition (HAR) is leading-edge in today's research field which has its applications in multiple research areas, some of those are Smart Health, Security and Ambient Assisted Living, etc. In today’s ubiquitous computing, HAR can be accomplished by espousing deep learning techniques that replace traditional analytical techniques that depend on the extraction of handcrafted features and classification methods. This work employed the Hierarchical Multi Convolution—Extreme Learning Machine approach for the classification of human activities. In the Hierarchical Multi CNN approach, the root CNN is employed to categorize the activities into static and dynamic activities. In the next level, two CNN-ELM are used to classify static activities into laying down, stand and sit; and classifies dynamic activities into Walking, Walking Downstairs, and walking upstairs. CNN-ELM approach exhibits its major advantages: CNN extracts the features from the dataset which confiscates expert knowledge in extracting features and ELM classifies the transitional results. This framework is evaluated on the UCI-HAR dataset and achieves an accuracy of 96.86%.

This paper presents the design and development of an efficient and intelligent chatbot. The proposed idea is to create a health companion which can be used for a daily well-being. An agile approach was used for its development and deployed on a platform such that all types of users can communicate with it. This virtual assistant makes our life easier and saves time via automated user query replies. Responding to user queries regarding home remedies as well as the recent pandemic COVID-19, the bot is well-equipped and reliable.

Leukemia which is commonly known as blood cancer is a fatal type of cancer that affects white blood cells. It usually originates from the bone marrow and causes the development of abnormal blood cells called blasts. The diagnosis is made by blood tests and bone marrow biopsy which involve manual work and are time consuming. There is a need for development of an automatic tool for the detection of white blood cell cancer. Therefore, in this work, a classification model using Convolutional Neural Network with Deep Learning techniques as a basis is proposed. This work was implemented using Keras library with TensorFlow as backend. This model was trained and evaluated on cancer cell dataset C_NMC_2019 which includes white blood cell regions segmented from the microscopic blood smear images. The model offers an accuracy of 91% for training and 87% for testing which is satisfactory.

Text Summarization is the process of creating a short and coherent version of a longer document that holds the same meaning as that of the original data. This article illustrates the technique to read the text in a printed document (such as newspaper, brochure, web document, etc.) and generate a summary of text. The method proposed is named Text Summarization using Optical Character Recognition and Attention Neural Networks (TORA). TORA can perform extractive summarization of a news article with the aid of Recurrent Neural Networks, Bidirectional Long Short-Term Memory, and Bahadanu Attention Network. The experimental results of the proposed method are promising. The experimental results have shown 80% accuracy in producing the summary from the large text document.

IoT is a next generation technology that provides advanced levels of services over internet. As predicted, in the future billions and billions of devices will be connected to the internet across the globe, the security is the main challenge to be addressed to deploy IoT applications. These devices need to be authenticated before accessing them. As IoT devices are constrained by resources, implementing complex cryptographic schemes are difficult. Physically Unclonable Functions (PUF) is a digital fingerprinting technique that is used to achieve authentication without using secret keys stored in a device. In this paper, we are proposing an effective, authentication and key sharing scheme using PUFs. Our proposed system has less complexity providing a better solution for device authentication in constrained IoT applications.

Internet of Things (IoT) paradigm joins physical objects in the real world to the internet and prompts the creation of smart environments and applications. A real world object is the building block of the IoT, known as a smart object that keeps an observation of the environment. These objects can communicate among themselves and also possess data processing capabilities. A novel framework is developed for the discovery of the object in the IoT ecosystem based on the data gathered by the data center about the objects in the IoT ecosystem. The search technique depends on the data gathered and after analyzing the data comprehension is made, the central data center maintains the type of data that the objects often send. Based on the gathered data, analytics is made and each object is categorized into a predefined category, thereby enhance the search mechanism based on the clue of the data that the objects transmit to the central processing entity.

One of the major disasters that occurred in different parts of the world is flooding, which causes huge loss and destruction to the human atmosphere. To minimize the effect of flood and alert public, current existing methods and systems will act after being affected by the disaster but detecting these conditions prior is very important. Therefore, to alert the public the proposed design of the system uses IoT and AI concepts to monitor the level of rainfall which alerts the public through SMS and Android application.

A cost effective way of implementing automatic gate control system is proposed in this paper. This will reduce the burden on security personnel. Vehicles are granted access inside the premises of the organization after identifying the number plate through image processing techniques. The algorithm to identify and extract the number plate will be deployed on a cloud where it will be used as a custom API service. An RFID reader is used for authentication in cases where the number plate is not recognized. Sensors are used which help in detecting the vehicle and help in closing and opening the gate at the right time. The pedestrian is granted access to the organization premises by reading the tag information embedded on the ID card using an RFID reader. A web page is developed to facilitate administrator control to add vehicle or pedestrian details. The automatic gate control system will be helpful to reduce the bottleneck created due to the large number of vehicles that enter the premises, especially at peak hours. The system is designed so that it should be cost effective without compromising on security.

Features of an ordinary camera security system are enhanced with the help of computer vision and Machine Learning algorithms that treat images as input data. It is an inexpensive and unique way of providing all those functionalities that a conventional camera security system cannot provide. The system includes face recognition, object detection, and other computer vision related operations. This camera security system is developed as an IOT solution.

Aquatic debris detection and monitoring is of very essential to ensure safety of aquatic habitats, human health and water transport. In this paper, we introduce the prototype of an aquatic debris detection system equipped with multiple aquatic sensors which are used to measure various parameters of the water body to determine quality of water. Based on this sensing platform, we propose a quick and accurate aquatic debris detection algorithm for remote monitoring which detects debris in the image captured by the camera.

Fleets play a vital role in organizations where the main form of revenue is through the manufacturing and selling of products. Businesses that depend on fleets need to manage their fleets efficiently which gives rise to the need for fleet management systems. Several management systems exist in the market that provides various functions to the user, but since the technology is itself quite recent, a lot of flaws and drawbacks are seen within them. The system that is mentioned here focuses mainly on the simplicity factor, that is, to make a system that is simplistic in design and simplifies fleet management operations more than what is obtainable in the current market. Speed and efficiency are also areas which the product aims to optimize or improve. The product focuses on contributing some new features while providing the primary functions that are required of a fleet management system rather than turning all efforts and resources in trying to provide as many functions as possible.

Many research fields and real-time applications have the necessity to extract and process handwritten text from documents or images. Handwritten character recognition (HCR)/intelligent character recognition (ICR) is an extended technology of optical character recognition (OCR) which is utilized to convert scanned document/image into editable text form. Offline handwritten character recognition emerged as challenging research topic as it is more favorable to recognize the data years after the document is created. The primary objective of the work presented here is comparison of various offline HCR tools in terms of character and word recognition accuracy. Experimental evaluation of offline tools and Android apps like Google Docs, ABBYY FineReader, CamScanner, Pen to Print, etc., is carried out and recognition accuracy is deduced. This analysis assists researchers to determine which tool is efficient for HCR in terms of recognition rate and also helps in using the tool/app in various real-time scenarios like recognizing bank check amount (written in words), recognizing postal address, captcha reading, healthcare sector, digitize handwritten notes, evaluation of answer scripts, etc.

Agriculture is one of the biggest economic sectors in India. Around 70% of the Indian economy depends on Agriculture. We know farmers as the lifeblood of the nation. But the condition of farmers is becoming worse day by day. They cannot get a generous share of profit for their hard work. Even the rate of farmers has decreased hazardously in recent years. In this paper, we are trying to resolve some of the main issues faced by farmers. Which includes debt rate, price crash, failure of crops, and so on. This paper presents a generic framework for e-agricultural portal comprising detailed knowledge about soil fertility to predict the correct fertilizers, seeds, and pesticides to be used for a particular area. We maintain this portal to help farmers for buying fertilizers, seeds, pesticides, and selling crops at a better rate using a decentralized application.

Software Defined Networking (SDN) is one of the emerging paradigms in the area of networking with its ability to scale and accommodate new features and addressing existing drawbacks. This paper works in this direction by exploring the avenues in Routing in SDN implementing QoS. As there is lot of different kinds of data moving across the network there is a requirement of providing QoS along with assuring that data is transferred from source to destination. This paper provides an insight into the requirement of QoS in routing with existing approaches and proposes a new scheme of accommodating QoS to improve the service.

The availability of service through cloud computing upon the emergence of IoT and wireless networking has exposed the security of information to an even greater risk of abuse. These security threats are caused by cyber criminals through Malware, Cracker, Insider and Zombie. Among the strategies attacking the network DoS or DDoS is disastrous as its objective is to paralyse the complete network system. The bursty or volatile nature of traffic is indicative of a surge in traffic flow. Such an event also implies the incidence of a DDoS attack. If left unchecked this attack will cause a disruption in service. The study focuses on deriving a statistical model for discriminating malign traffic by devising a predictive model using error optimization technique and Hurst Correlogram analysis for effectively detecting and controlling the spread of bot-induced DDoS attack in a cloud environment. This work proposes ARIMA model that detects DDoS with an accuracy of 92% which is further assessed through MLR with an accuracy of 84%. The Hurst estimation of the univariate lies very close to the Hurst of raw data needs mention and proves the maturity of the prediction strategy.

A picture provides the necessary information to the user. However, not all photos will do so since the backdrop is blurred. Blurry within the image could be a condition due to various factors like camera shifting, inappropriate aiming, aperture use and different external influences. Such photos are called deteriorated photographs which have been degraded. Owing to the rapid body activity, haze is sometimes created and can be categorized into two separate types, i.e., uniform and non-uniform. Distorted pictures of poor quality have, however, received fewer attentions. There are several forms a corrupted file can be measured and reconstituted. These issues can be overcome by adding a profound learning strategy, like a convolutional neural network, which simultaneously allows deblurring and super resolution. Photo super resolution is an efficient means of taking high resolution pictures with low resolution photographs. Using the advantages of super resolution, neural networks are superior experimentally to current deep learning algorithms. By this suggested model, we show, scientifically, that the consistency and quantity of our system is obtained. We may achieve good quality resolution in documented and unknown amounts of blur.

The wide acceptance of the IoT technology and deployment of large scale IoT networks has led to it being used for diverse applications. This highly distributed sensor infrastructure is able to acquire vast amount of data, but the weight, power and size (SWAP) constraints limit the amount of processing that can be done by the edge devices. Naturally, the next step in the evolution of the system was to integrate the IoT systems with the Cloud infrastructure and harness its computing and storage capability to realize highly intelligent systems which could be used for purposes of enhancing Homeland security, monetization of data, Smart Agriculture, Traffic Monitoring and Control systems among others. It was seen that the large amount of raw data acquired by these sensors if piped to the cloud directly imposes a large load on the communication infrastructure. This led to the evolution of the Fog architecture in which higher computing power is deployed between the Edge devices and the Cloud at a location where the SWAP constraints are far less stringent. Consequently, processed information is sent to the Cloud. Quite a few of these applications require security of data to be assured while in transit as well as while at rest. The Tangle technology is adopted for ensuring integrity of data in the Foggy section of the network while the Cloud has its own mechanisms. Very little work has been done to address the end to end integrity requirements in Fog computing networks by researchers as of date.

As the effect of COVID-19 is increasing rapidly every day, it becomes very difficult for the survival of many people, and there is a high effect on the economic situations of every country which is affected by it. In the same way, it is affecting all states of India and causing an economic crisis. This paper deals with the analysis of the impact caused by COVID-19 on each state in India and also gives an estimated date on which the effect will reduce and may even become zero along with the analysis report of overall India. For the predicting of the effect, we have used generalized autoregressive conditional heteroskedasticity (GARCH) algorithm which has produced root mean squared error (RMSE) of around 5.89 for some of the states and other with 20.05 due to the data abnormality. The predicted data for each state is projected in the figures using the line plots. And the resulted graphs are explained clearly. The accuracy of the proposed model is around 94.6%–96.8% for the states with good data and less RMSE and 80% for the states with abnormal data and high RMSE value. From the produced results of the proposed methodology, the dates of which the effect of COVID-19 will decrease are calculated for the states having a high number of cases.

There have been several forays into developing a collision avoidance and navigation assistance system for the visually impaired. We aim to transform the simple white cane into a tool which will detect obstacles, determine approximate distance from obstacles and navigate the user through this maze of obstacles safely. The system fuses data from diverse sensors namely, a FMCW Radar and Ultrasonic sensors and creates a map of the obstructions lying ahead. In addition, the signals acquired from a downward looking LIDAR sensor are also processed independently to detect ditches, pot-holes and stairs. This information is also super-imposed onto the Obstruction map to caution the user of obstacles and guide him safely. A smart-phone based application (Kannu–Eye) has also been developed which supports the smart cane with a wide range of additional functions like “Find My Cane”, Emergency services, and navigation assistance. The HMI for Kannu has been developed such that all its operations are completely voice-controlled thus providing a simple and effective user interfaces. Kannu has been implemented on an Android platform.

Health care has been a major concern for everyone right from the inception of mankind. There is no doubt that medical science has done remarkable headway in this arena too. Here, the detection of disease a person is suffering from is a key aspect and hence, in this project we proposed a model which primarily focuses on having easy diagnosis and prediction of the disease. Moreover, the primary objective of this project is to provide remote diagnosis to the patients. In this suggested system the user can provide the input either by speech or entering directly into the UI. The proposed model once detecting the disease also displays the description of the same and a more info link for further elucidation. Along with it the user is also provided with important health tips, a balanced diet plan and required exercises which will help the user to make the required changes in their diet and daily routine which would lead to a potential healthy lifestyle. Apart from this, as per the current situation of the ongoing pandemic we also added few other features other than the above mentioned like a COVID-19 tracker which helps the user to stay updated with the state wise count which gives us the total confirmed,active, recovered and deceased cases according to the Ministry of Health and Family Welfare department and it is purely dynamic in nature. Also, we are providing the potential causes of the disease and the preventive measures to be taken by the user as per the government guidelines. By this, we are trying to achieve better health care in a technological aspect. Using the latest technologies will make it much efficient to know about various symptoms and predict the cause at an early stage which will help in taking necessary steps to minimize the damage.

A new Identity-Based Broadcast Encryption (IBBE) scheme is presented that uses integer matrices in finite field as keys. The plain text is encrypted such that only the unrevoked users from a set of registered users can decrypt the ciphertext. The sizes of the ciphertext, public key, and the private keys vary linearly with the total number of registered users. In the proposed scheme, the revocation process is direct. That is a separate access controller does not exist between the users and the storage to filter out the revoked users, because the revoked users anyhow cannot decrypt the ciphertext.

This paper reviews the different methods that are used in sentiment analysis for detecting depression in social media users. Sentiment analysis of the text is generally executed sequentially: every set of text is broken down progressively into components such as sentences, phrases, and words. The tokenized text is then processed such that each component of speech is identified for its sentiment and then assigned a score in the range of ‒1–1, with ‒1 indicating the most negative emotion and + 1 indicating the most positive emotion. The aggregate score of the entire text is then used to determine the overall sentiment of the text. The paper outlines a proposal of a cumulative analysis method to detect the depression level of a person by extracting data from social media posts in conjunction with other data pertaining to the user such as sleep, food intake, and exercise patterns. The proposed method performs a multi-varied analysis on the type of emotion exhibited by the data using machine learning techniques on extracted data for a given user.

This paper proposes a new technique named “Process Logo” for visualizing the causal relationship between the activities of a process (Control flow). Traditional process mining algorithms rely on representing the activity as a sequence of operations modeled using nodes and edges, as the number of activities increases, the representation of the entire control flow becomes quite tedious. Process logo is a compact yet highly informative method for visually representing the process model. It visually summarizes the number of activities, sequence of execution, relative significance, and dependency between activities. It uses a dynamic programming method—sequence alignment and clustering approach with Levenshtein measure as a distance measure. The proposed method is evaluated on the synthetic event log, the experimental result is promising.

The N-degree of a vertex $${v}_{i}$$ v i is the sum of the degree of the vertices in the open neighborhood of $${v}_{i}$$ v i . The maximum degree energy of graphs has been reported recently. Motivated by this work, we have introduced the maximum N-degree energy of graphs denoted by $$E_{N} \left( G \right)$$ E N G and computed the maximum N-degree energy of certain standard families of graphs, and also properties of the maximum N-degree eigenvalues are discussed.

Power-aware dynamic virtual machines (VMs) allocation in Cloud Data Centers (DCs) taking advantage of cloud computing paradigm. Each VM request is characterized by four parameters: CPU, RAM, disk and bandwidth. Allocators are designed in order to accept as many VM requests as possible, taking into account the power consumption of the network devices. In the proposed system, three different allocation strategies introduced are Analytic IT Resource allocation and Fuzzy IT Resource allocation and genetic algorithm-based allocation. The experimental results are conducted to show two different allocation strategies of VM under data centers.

The Paper proposes an idea for enhancing the usefulness of the technology for the betterment of the visually impaired. The project includes an Android app which captures the image of the surrounding to the blind person and send it to an Image captioning algorithm. The image captioning algorithm processes the image and converts it to text. It uses both Computer Vision and Natural Language Processing to generate the captions. The textual description of the image is then converted to speech which is then sent to the user. Technologies such as Deep Learning, Cloud Computing and tools like Flask and Tensorflow are been used at the backend. Cloud is been used as a serving for the Image Captioning API. The Android app is made using Android Studio. The textual description which is converted to speech is sent back to the user so that it gives them an idea of what is present in their surrounding.

Water quality has deteriorated over the years, largely due to man-made disasters. Water is a fundamental entity of life for humans. Water plays a vital role in supporting growth, not only in human life but also in the life of planet earth. Each organism on earth is solely dependent on water. Water is life and life is water. The fundamental driver for water quality issues is over-misuse of regular assets. The paper proposes two major segments one is the Internet of Things (IoT) based Water Quality Analysis (WQA) based framework execution and a consolidated IoT grounded three-phase water sanitization framework so as to get a joined framework meeting the goal of ongoing water quality estimation and it’s cleaning. The Measurement Kit comprises pH, Temperature, Dissolved Oxygen (D.O.) and Turbidity sensors with Bluetooth module for estimation, information obtained and logging reason. Purification framework contains three phases be specific primary, secondary and tertiary stages. The measurement kit has been verified with standard methods of assessment and the results are more than 93% accurate.