Progress in Advanced Computing and Intelligent Engineering

Table of Contents

1. Frontmatter

2. Advanced Image and Video Processing Applications

   1. Frontmatter

   2. Crack Detection on Inner Tunnel Surface Using Image Processing

      Debanshu Biswas, Ipsit Nayak, Shaibal Choudhury, Trishaani Acharjee, Sidhant, Mayank Mishra

      Abstract

      Cracks in the concrete structures such as cracks in the inner surface of tunnels are minor fault, however, can cause major damage or loss of lives if not checked frequently. The current method of detecting cracked surface is manual inspection by hand measuring tools and drawing sheets which may not be feasible as the tunnel needs to be blocked for a limited time period, till the inspection is in progress. By image processing, we can analyze the digital images captured from inside the tunnel for localization of cracks. The algorithm proposed in this paper can be applied to an image of the cracked surface of a tunnel for detecting the crack. Moreover, the length of the crack can also be measured in pixels.

   3. Novel Approach for Resolution Enhancement of Satellite Images Using Wavelet Techniques

      Mansing Rathod, Jayashree Khanapuri, Dilendra Hiran

      Abstract

      Today, many researchers are working on satellite images, to solve resolution problems. Some of the techniques are used to enhance the resolutions that are Stationary Wavelet (SWT), Discrete Wavelet (DWT) and Integer Wavelet Transforms (IWT). These wavelet transforms are considered for improving the quality in terms of resolution enhancement. Low resolution (LR) image is used for processing and decomposed by mentioned transforms. Interpolation techniques are applied to manipulate the output of transform images. Some estimated images determined through the interpolation factor 2 to make equal sizes of images. All the images are integrated with inverse wavelet transform to generate high resolution (HR). In this work, we have compared wavelet transform. LANDSAT5 satellite images are considered for verification of implemented techniques. In this paper, three types of results are presented that are enhanced the image with size 128 × 128 to 512 × 512, 256 × 256 to 512 × 512 and iterative resolution. Some of the quality parameters are applied, i.e. PSNR, RMSE, MAE and MSE for verification for the performance of implemented techniques.

   4. Vehicle Number Plate Detection: An Edge Image Based Approach

      Kalyan Kumar Jena, Soumya Ranjan Nayak, Sasmita Mishra, Sarojananda Mishra

      Abstract

      Transportation system (vehicle communication) plays a major role in today’s scenario. Detection of vehicle number plate exactly in blurry conditions was the most challenging issue found in the last three decades. Although many intensive studies were undertaken, none addressed this problem exhaustively. Various methods are introduced by several researchers for detecting the vehicle number from the vehicle number plate images. The purpose of this study was to investigate this current issue by implementing an edge-based approach on the basis of quantitative combination of Canny, Morphological and Sobel methods for the accurate detection of vehicle number in blurry conditions. The experimental results demonstrated that the proposed scheme outperforms its counterparts in terms of Sobel, Prewitt, Roberts, Laplacian of Gaussian (LoG), Morphological and Canny methods in all aspects with higher peak signal to noise ratio (PSNR) and signal to noise ratio (SNR) values. Hence, the proposed hybrid scheme is better and robust and results in accurate estimation of vehicle number from the blurry vehicle number plate (BVNP) images for the given datasets.

   5. A High-Precision Pixel Mapping Method for Image-Sensitive Areas Based on SVR

      Huang Jing, Amit Yadav, Asif Khan, Dakshina Yadav

      Abstract

      It is necessary to monitor the grain size characteristics of particles at production site to control the production equipment, for the assurance of product quality. In this respect, prior research finds that it is critical to evaluate the accuracy of tiny particles since the current practice indicates that the existing methods illustrate multiple shortcomings including large measure error, low accuracy and poor repeatability. Therefore, to improve the accuracy of particles, monitoring this study proposed a calibration method based on SVR algorithm to predict the accurate pixel size of the particles. Results revealed that high-precision pixel mapping of the sensitive area transforms the pixel mapping of the particle image closer to the actual size and improves the measurement precision of the whole system.

   6. Nucleus Segmentation from Microscopic Bone Marrow Image

      Shilpa, Rajesh Gopakumar, Vasundhara Acharya

      Abstract

      Acute myeloid leukemia (AML) is a type of cancer that originates in the bone marrow and moves into the blood quickly. The disease progresses rapidly if not treated. Therefore identification of the disease at an early stage is very important. In the proposed work, different image segmentation techniques like K-medoids and watershed, region growing and active contour have been applied for the segmentation of the nucleus from the blast cells and the results are analyzed.

   7. Developing a Framework for Acquisition and Analysis of Speeches

      Md. Billal Hossain, Mohammad Shamsul Arefin, Mohammad Ashfak Habib

      Abstract

      Speech plays a vital role for human communication. Proper delivery of speech can enable a person to connect with a large number of people. Nowadays, a lot of valuable speeches are being provided by many popular people throughout the world and it will be very helpful if important information can be extracted from those speeches by analyzing them. An automatic speech-to-text converter can facilitate the task of speech analysis. There have been carried out a lot of works for the conversion of speech to text in the last few decades. This paper presents a framework for the acquisition of speech along with the location of the speaker and then conversion of that speech into text. We have worked with speeches containing three different languages. To evaluate our framework, we collected speeches from several locations and the result shows that the framework can be used for efficient collection and analysis of the speeches.

   8. Gait-Based Person Identification, Gender Classification, and Age Estimation: A Review

      Rupali Patua, Tripti Muchhal, Saikat Basu

      Abstract

      In this era where both techniques and technology are going digital, there is a need of designing and developing such a system that automatically identifies the person and also the person’s attributes with the help of the biometric system. The biometric of every person is unique. Therefore, in this context, the use of the biometric system to identify the identity of an individual is a popular approach. Gait is a biometric approach that helps to verify the identity of a person by their walking patterns. Owing to the advantages of gait as a biometric, its popularity among the researchers is amplified in recent years. In this paper, a review is represented, where several papers with different methods, have been mentioned which recognize the person and its attributes such as gender and age based on the gait of the person.

   9. Efficient Watermarking in Color Video Using DWT-DFT and SVD Technique

      Mangal Patil, Preeti Bamane, Supriya Hasarmani

      Abstract

      With the advancement in Internet technology, a digital video can be easily modified, copied, and distributed among a large audience. Copyright protection and security become very essential aspects because of the extensive use of digital multimedia applications. Digital watermarking is being used for copyright protection, data authenticity for multimedia contents such as image, audio, and video. In this paper, a DWT-DFT-SVD-based method is opted to improve robustness and overall computational requirements. The computed PSNR between original video signal and watermarked signal is improved up to 60 db. The normalized correlation value of the original and the extracted watermark image have a high level of imperceptibility. The proposed scheme shows strong robustness against several geometric and non-geometric attacks.

   10. Motif Discovery and Anomaly Detection in an ECG Using Matrix Profile

       Rutuja Wankhedkar, Sanjay Kumar Jain

       Abstract

       Time Series Data mining is a popular field in data science to discover and extract useful information from the time series data. Time Series Motif discovery is one of the tasks in data mining to discover frequent patterns which are unknown previously. Motif discovery has gained a lot of attention since its advent in 2002. Many motif discovery techniques were introduced and applied in various domains like E-commerce, Weather Prediction, Seismology, etc. In this paper, we introduce a technique for anomaly detection and motif discovery in the ECG data using Matrix Profile which has been introduced recently in the literature. Anomaly detection in ECG helps to detect the abnormal heartbeats before the process of diagnosis and motif discovery helps to locate the highly similar beats in the ECG. Using Matrix Profile for the task of anomaly detection and motif discovery in our proposed technique provides our technique with properties that are inherited from Matrix Profile. Thus, the proposed technique in this paper has properties like exactness, simple and parameter-free, space-efficient, anytime, handle missing data, free from the curse of dimensionality.

   11. ISS: Intelligent Security System Using Facial Recognition

       Rajesh Kumar Verma, Praveen Singh, Chhabi Rani Panigrahi, Bibudhendu Pati

       Abstract

       Security is mandatory and of utmost importance for all organizations. While the legitimate people should be allowed inside enterprises, the illegal ones should be barred from entering and this can be achieved using face recognition techniques. In this work, we have come up with a robust architecture using artificial intelligence and Internet of things that can be used across different enterprises. We have also derived the methodology and solution for implementing a more advanced security system. For proper demonstration, we have also considered one of the business use cases along with proposed processing work flow.

   12. G.V Black Classification of Dental Caries Using CNN

       Prerna Singh, Priti Sehgal

       Abstract

       Dental caries is one of the most predominant pathologies in the world. Early detection of dental caries leads to the prevention of tooth decay. According to G.V Black classification model dental caries can be classified into six classes (Class I–Class VI) based on the location of caries. The proposed work classifies caries infected tooth based on G.V black classification using deep convolution network architecture (DCNN). In the proposed approach, feature extraction of the preprocessed images is done using Local Ternary Pattern (LTP) and feature reduction is done using Principal Component Analysis (PCA). The pretrained models used in the study are AlexNet architecture and GoogleNet architecture. AlexNet is 8 layer DCNN and classifies the tooth with an accuracy of 93%, 90% sensitivity and 92% specificity. GoogleNet is a 22 layer DCNN and classifies with an accuracy of 94%, 91% sensitivity and 93% specificity.