Sandeep Chand Kumain

Bio: Sandeep Chand Kumain is an academic researcher from Graphic Era Hill University. The author has contributed to research in topics: Digital image processing & Noise. The author has an hindex of 1, co-authored 4 publications receiving 4 citations. Previous affiliations of Sandeep Chand Kumain include National Institute of Technology, Srinagar.

An efficient Gaussian Noise Reduction Technique For Noisy Images using optimized filter approach

Abstract: In the Multimedia era, removal of the Noises from an image becomes a key challenge in the field of Digital Image Processing (DIP) and Computer Vision. Noise may be mixed with an image during capturing time, transmission time or due to dust particle on the screen of capturing device. Therefore, removal of these unwanted signals from the image is urgently required for the better analysis of the image and the de-noised image is more meaningful for Object detection, Edge detection and many more. There are various types of image noise, however, Gaussian Noise and Impulse Noise are commonly found in the image. This work focuses on the outliers and Mean Filter to improve the performance for Gaussian noise reduction from the image. In experimental assessments, artificial noise has been mixed using MATLAB to MSRA (10k images) dataset, this dataset is used to evaluate our proposed technique. The experiment results show that the proposed approach improves the performance in noise reduction over other filter approaches.

VBNC: Voting Based Noise Classification Framework Using Deep CNN

Abstract: In this revolutionary digital world, the role of images and videos for transmitting the information is much more than the text Now, people prefer short, concise and informative data over the large volume of texts like the information in the form of images is preferable Noise availability in an image is a major challenge in digital image processing and this presence of noise misleads the results during various computer vision operations The image denoise step is must require in this case Multiple noise reduction algorithms are developed and used till now, but mostly the work is done with the known information of noise In this research work, the author(s) focused on noise type identification in an image that will helpful for the images having unknown information about the type of noise is present Further, it will help to select the appropriate algorithm for noise reduction purposes For this purpose, the author(s) proposed a voting based noise classification framework that utilized deep learning technology

Smart Multipurpose Robotic Car

Abstract: In this modern digital world, everyone is moving towards automation. Robotics allows automation where machines perform a well-defined step safely and productively, in autonomous or partial autonomous manners. Many articles have been published in recent times to present automatic/robotic cars which can perform some process such as, smart surveillance, obstacle detection, and collision aversion, etc. The design of the smart robotic car is presented through this paper and demonstrates its operations. The proposed developed model can perform multiple operations such as smart irrigation, obstacle detection, line-following, grass cutting, and vacuum cleaning. All these operations are operated or managed through the smartphone. The proposed smart model work as the expert system which utilized the features of Artificial Intelligence. The smart robotic car can sense the environment and take the proper decision without human guidance.

SEF-VSOD: A Static Ensemble Framework For Video Saliency Using Modified Doubly U-Net

Abstract: In this digital age, video and images are commonly used to process information. Although an image or video contains a lot of information, not all of it is useful. One of the key areas of computer vision is salient object detection, whose major goal is to replicate the human visual system and recognize the scene's most prominent object. The salient object detection techniques have a significant impact on a wide range of applications, including video summarization, automated cropping, etc. Deep learning approaches have recently become more common in the saliency detection problem. However, it is highly challenging for a single model to perform well in all circumstances due to the different constraints in the model creation process (training, parameter adjustment). To address this issue, the author(s) present a static ensemble framework for estimating video saliency. This ensemble approach aided in model performance improvement. The proposed model's performance is evaluated using two well-known publicly available video datasets, namely ViSal and DAVSOD-Easy. Based on the various quantitative evaluation parameters the proposed model performance is compared with the state-of- the-art SOD models.

Adaptive Genetic Algorithm-Aided Neural Network With Channel State Information Tensor Decomposition for Indoor Localization

Abstract: Channel state information (CSI) can provide phase and amplitude of multichannel subcarrier to better describe signal propagation characteristics. Therefore, CSI has become one of the most commonly used features in indoor Wi-Fi localization. In addition, compared to the CSI geometric localization method, the CSI fingerprint localization method has the advantages of easy implementation and high accuracy. However, as the scale of the fingerprint database increases, the training cost and processing complexity of CSI fingerprints will also greatly increase. Based on this, this article proposes to combine backpropagation neural network (BPNN) and adaptive genetic algorithm (AGA) with CSI tensor decomposition for indoor Wi-Fi fingerprint localization. Specifically, the tensor decomposition algorithm based on the parallel factor (PARAFAC) analysis model and the alternate least squares (ALSs) iterative algorithm are combined to reduce the interference of the environment. Then, we use the tensor wavelet decomposition algorithm for feature extraction and obtain the CSI fingerprint. Finally, in order to find the optimal weights and thresholds and then obtain the estimated location coordinates, we introduce an AGA to optimize BPNN. The experimental results show that the proposed algorithm has high localization accuracy, while improving the data processing ability and fitting the nonlinear relationship between CSI location fingerprints and location coordinates.

Feature Extraction with Handcrafted Methods and Convolutional Neural Networks for Facial Emotion Recognition

Abstract: This research compares the facial expression recognition accuracy achieved using image features extracted (a) manually through handcrafted methods and (b) automatically through convolutional neural networks (CNNs) from different depths, with and without retraining. The Karolinska Directed Emotional Faces, Japanese Female Facial Expression, and Radboud Faces Database databases have been used, which differ in image number and characteristics. Local binary patterns and histogram of oriented gradients have been selected as handcrafted methods and the features extracted are examined in terms of image and cell size. Five CNNs have been used, including three from the residual architecture of increasing depth, Inception_v3, and EfficientNet-B0. The CNN-based features are extracted from the pre-trained networks from the 25%, 50%, 75%, and 100% of their depths and, after their retraining on the new databases. Each method is also evaluated in terms of calculation time. CNN-based feature extraction has proved to be more efficient since the classification results are superior and the computational time is shorter. The best performance is achieved when the features are extracted from shallower layers of pre-trained CNNs (50% or 75% of their depth), achieving high accuracy results with shorter computational time. CNN retraining is, in principle, beneficial in terms of classification accuracy, mainly for the larger databases by an average of 8%, also increasing the computational time by an average of 70%. Its contribution in terms of classification accuracy is minimal when applied in smaller databases. Finally, the effect of two types of noise on the models is examined, with ResNet50 appearing to be the most robust to noise.

Image lossless encoding and encryption method of EBCOT Tier1 based on 4D hyperchaos

Abstract: To satisfy requirements of high security and transmission efficiency in application scenarios with high image quality, an image lossless encryption and compression algorithm based on four-dimensional hyperchaos and embedded block coding with optimal truncation (EBCOT) is proposed in this paper. First, according to a character that the amplitude of the high-frequency part of the wavelet coefficient is less than the amplitude of the low-frequency part, an encryption algorithm for the wavelet coefficients is proposed to improve the security while reducing the impact on the compression performance. Second, the bit-plane coding and arithmetic coding in EBCOT Tier1 are embedded with encryption points, and the encryption process and compression process are combined to propose a secure EBCOT Tier1 code, which could further improve the security of the algorithm. Furthermore, this paper proposes a new four-dimensional hyperchaotic system, using Secure Hash Algorithm-256 (SHA-256) to generate initial values of the chaotic system, so that the algorithm could resist known plaintext and selected plaintext attacks. The results of the mean square error of all restored images are 0 and the information entropy of this algorithm is close to the theoretical value. The experimental results show that the algorithm has high security and lossless compression performance.