Papers by Dr. Mantosh Biswas
arXiv (Cornell University), Jan 20, 2021
SAR (VV and VH polarization) and optical data are widely used in image fusion to use the complime... more SAR (VV and VH polarization) and optical data are widely used in image fusion to use the complimentary information of each other and to obtain the better-quality image (in terms of spatial and spectral features) for the improved classification results. This paper uses anisotropic diffusion with PCA for the fusion of SAR and optical data and patchbased SVM Classification with LBP (LBP-PSVM). Fusion results with VV polarization performed better than VH polarization using considered fusion method. For classification, the performance of LBP-PSVM using S1 (VV) with S2, S1 (VH) with S2 is compared with SVM classifier (without patch) and PSVM classifier (with patch), respectively. Classification results suggests that the LBP-PSVM classifier is more effective in comparison to SVM and PSVM classifiers for considered data.
Remote Sensing, Jan 3, 2023
This article is an open access article distributed under the terms and conditions of the Creative... more This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY
International Journal of Software Innovation
Impulse noise generally occurs because of bit errors in progression of image acquisition and tran... more Impulse noise generally occurs because of bit errors in progression of image acquisition and transmission. It is well known that median filtering method is an impulse noise removal method. Lots of modified median filters have been proposed in the last decades to improve the methods for noise suppression and detail preservation, which have their own deficiencies while identifying and restoring noise pixels. In this article, after deeply analyzing the reasons, such as decreased noise detection and noise removal accuracy that forms the basis of the deficiencies, this article proposes a modified weighted median filter method for color images corrupted by salt-and-pepper noise. In this method, a pixel is classified into either “noise free pixel” or “noise pixel” by checking the center pixel in the current filtering window with the extreme values (0 or 255) for an 8-bit image using noise detection step. Directional differences and the number of “good” pixels in the current filtering windo...
Wireless Personal Communications
In the modern era of Internet of Things (IoT), deep-learning-based systems are exhibited encourag... more In the modern era of Internet of Things (IoT), deep-learning-based systems are exhibited encouraging performance in hyperspectral image (HSI) classification, because of their capability in extracting key deep features from available images. However, this deep learning centered method typically needs a great amount of training samples in IoT and the main issue with HSI is that labeled samples are not adequate and it can force to overfitting/underfitting issue. Motivated by this, the authors proposed a novel deep learning-based intelligent decision support system, which attains promising accuracies with limited training samples using Manifold Batch Structure (MFBS). Three novel approaches have been proposed to design MFBS. Firstly, a manifold batch scanning approach utilized to conclude the spatial association among the neighboring pixels and, the spectral associations between distinct bands. The proposed manifold batch scanning approach integrates the spatial and spectral association within different batches as well as extracts the collective spatial-spectral information. Secondly, since hyperspectral images have ample of unlabeled pixels, hence we refer such samples in the semi-supervised way while construction of convolution kernels. Finally, the MFBS has developed on a network infrastructure that does not include any hyper parameters for alteration. The experiments results on such standard datasets have revealed that MFBS outperforms various related existing HSI classification framework that too in case related to small training datasets.
2019 International Conference on Communication and Electronics Systems (ICCES)
Super resolution algorithms always used as a tradeoff between the cost of the high definition (HD... more Super resolution algorithms always used as a tradeoff between the cost of the high definition (HD) cameras and the quality and/or clarity of the image obtained. There are various predefined algorithms that obtain Super Resolved images from Low Resolution (LR) images, some (such as, Convolutional Neural Network (CNN), Deep learning, Sparse Representation based algorithms) gives better results for e.g. deburring of zoomed part, removal of noise, color enhancement and so on but are computationally complex or hard to implement in real-time environment whereas some are very simple to use (such as, interpolation based, wavelet based algorithms) but lack quality for e.g. ringing artifacts, edge blurs, poor image quality etc. In this paper, we proposed a method that combines advantages of some of the above mentioned methods. Our proposed method obtains High Resolution (HR) image using saliency model for detection of visually dominant regions, Discrete Wavelet Transform (DWT) for extraction of high frequency details, finally Multi-layer Perceptron (MLP) and Particle Swarm Optimization (PSO) for interpolation. Experimental results visually and quantitatively show that for considered test images our proposed super resolution method appears to be most promising compared to bi-cubic, Chopade et al., Yu et al. and Man et al. methods.
2020 IEEE India Geoscience and Remote Sensing Symposium (InGARSS)
SAR (VV and VH polarization) and optical data are widely used in image fusion to use the complime... more SAR (VV and VH polarization) and optical data are widely used in image fusion to use the complimentary information of each other and to obtain the better-quality image (in terms of spatial and spectral features) for the improved classification results. This paper uses anisotropic diffusion with PCA for the fusion of SAR and optical data and patchbased SVM Classification with LBP (LBP-PSVM). Fusion results with VV polarization performed better than VH polarization using considered fusion method. For classification, the performance of LBP-PSVM using S1 (VV) with S2, S1 (VH) with S2 is compared with SVM classifier (without patch) and PSVM classifier (with patch), respectively. Classification results suggests that the LBP-PSVM classifier is more effective in comparison to SVM and PSVM classifiers for considered data.
2021 IEEE International India Geoscience and Remote Sensing Symposium (InGARSS)
2021 International Conference on Sustainable Energy and Future Electric Transportation (SEFET)
This paper presents a comprehensive study of three phase SEIG (Self-Excited Induction Generator) ... more This paper presents a comprehensive study of three phase SEIG (Self-Excited Induction Generator) feeding power to single phase loads exploiting renewable energy sources (RESs) and such systems are of great importance for the developing countries with remote location without grid connectivity. This paper includes the various excitation circuit topologies, voltage and frequency control schemes, steady-state and transient analysis techniques for three-phase SEIG connected to single-phase loads. The paper also presents the steady-state mathematical model of three-phase SEIG with single-phase loads, which is very useful for deciding the size of excitation capacitor in the different configurations.
Recent Advances in Computer Science and Communications
: Refining the quality of a noisy image is essential for many image applications. Various median ... more : Refining the quality of a noisy image is essential for many image applications. Various median filter variants have been introduced to suppress various noises, but they have their own limitations when detecting noise and restoring noise-free images. Denoising convolutional neural networks (DnCNNs), primarily developed for Gaussian noise removal, are influential nonlinear mapping models in image processing. After alterations in training data, they can be used to suppress other noise with outstanding results. This article suggests a frequency median filter method combined with deep learning for color images polluted by Salt and Pepper (SnP) noise. The analysis presented in this paper has primarily used a frequency median filter to suppress impulse noise wherein the restored value for the center pixel is evaluated by the frequency median rather than the traditional median. After which, the pretrained denoising convolutional neural network is hired to suppress the remaining noise and attain the output image finally. With a visual comparative study, simulation results on the taken test images show that the proposed method surpasses de-noising methods in terms of PSNR, SSIM, NMSE, Entropy, IEF, NCC, PCC and Running Time.
Lecture Notes in Networks and Systems, 2017
Edge in an image depends on the viewer’s perspective i.e., some viewers may feel it as edge and s... more Edge in an image depends on the viewer’s perspective i.e., some viewers may feel it as edge and some may not. Fuzzy logic can be used to solve this partial truth value concept. Many fuzzy-based edge detection methods have been proposed till now, but most of them used the static fuzzy inference system for edge detection, in which we have to change the membership functions for each image in order to get better results. Therefore, to overcome this drawback, we proposed fuzzy logic-based edge detection algorithm with dynamic generation of fuzzy interface system (FIS). The performance of the proposed method is demonstrated through computer simulation results over Sobel, Canny, EFLEDG, and EDFLM edge detection methods in terms of both subjective and fidelity criteria and our proposed method gave good results in terms of F-Measure and visual quality of resultant edge images.
2018 Second International Conference on Intelligent Computing and Control Systems (ICICCS), 2018
Edge in image processing is considered as those pixels whose intensity value changes drastically ... more Edge in image processing is considered as those pixels whose intensity value changes drastically and finding the object boundary is the main task of any edge detection technique. There have been various Ant Colony Optimization (ACO), Particle Swarm Optimization (PSO) based techniques that have been applied to solve edge detection problem, but most of them have not considered the noisy environment which in itself makes edge detection further more difficult task and the user-defined threshold approach doesn't always give desired results. The paper proposes a Whale Optimization Algorithm (WOA) based edge detection technique with weighted fitness function including homogeneity, uniformity and average gradient magnitude as main factors for detecting the edges of additive gaussian noise images. The experiment results have shown that the proposed technique has performed better under noisy environment for conventional edge detectors: Sobel, Canny and ACO based technique for both objective criteria i.e. restored edge images and subjective criteria i.e. PSNR, Precision, Recall and F -measure.
Lecture Notes in Electrical Engineering, 2020
Maximum power point tracker (MPPT) controls the DC/DC converter for extracting maximum power from... more Maximum power point tracker (MPPT) controls the DC/DC converter for extracting maximum power from solar photovoltaic (SPV) array connected with power generation system. MPPT operates at its maximum power point (MPP) (Vmp, Imp) irrespective to load conditions and input weather conditions. Use of by-pass diodes in series-connected SPV modules under non-uniform insolation is a key cause for many power peaks in the power–voltage characteristics of SPV array. Henceforth the problem of MPPT under partial shading becomes a nonlinear optimization problem. A new quick and reliable MPPT technique is proposed in this paper to identify the global MPP under partial shadow conditions. The computation time and correctness in tracking global MPP are compared with standard soft computing techniques: modified binary (MB) search, differential evolution (DE) techniques, and particle swarm optimization (PSO). The results show correctness of the presented random binary search technique in tracking the global MPP in very less time than the conventional soft computing techniques. The technique is quick, simple, and oscillation-free for tracking global MPP in least iterations; hence, the computation (hardware) requirements are less than that using PSO and DE MPPT techniques.
IEEE Transactions on Industry Applications, 2022
This article presents the design, development, and performance analysis of a three-phase synchron... more This article presents the design, development, and performance analysis of a three-phase synchronized ac chopper (SynACC)-based controller for small hydrogeneration systems. The three-phase SynACC-based controller is used to control the voltage and frequency of a three-phase self-excited induction generator (SEIG)-based small hydropower generation system. The SynACC-controlled dump load is connected at the point of common coupling across the consumer load that controls the system frequency and voltage using instantaneous power balancing method. The SynACC-based controller has synchronized the system frequency feedback-based pulsewidth modulated current delivered to the dump load with the SEIG output voltage in order to make the equal instantaneous currents and similar frequency spectrum of current in all the three phases of the dump load circuit. The equal instantaneous currents and similar frequency spectrum of currents in all the three phases of the dump load circuit significantly improve the system stability. That also improves the system efficiency by removing the multiple and fast fluctuating torque generation problem in three-phase SEIG-based small hydrogeneration systems. The SynACC controller improves the system power factor, since it reduces the reactive power demanded by the dump load circuit. The laboratory prototype of the scheme has been developed and detailed mathematical modeling along with the simulation and experimental results are presented in the article to validate the claims. The technique presented in this article may also be used to develop an improved power quality heating furnace system for chemical and process industry applications.
Advances in Intelligent Systems and Computing, 2018
Classification is the process of setting class labels to pixels based on some obtained properties... more Classification is the process of setting class labels to pixels based on some obtained properties. Hyperspectral images (HSI) have very high dimensionality, which results in higher cost and complexity for analyzing and classifying them as superfast processors and large storage devices are required. Moreover, due to limited training samples and labeled data, classification remains an arduous task. Many methods have been presented till now for classification of HSI based on traditional methods that use handcrafted features beforehand, principal component analysis and its variations, decision trees, random forests, SVM-based methods, and neural networks, but most of these consider only the spectral information for classification resulting in low classification accuracy. Nowadays, increasing spatial resolution of HSI demands obtaining spatial data for further improving classification performance. We, therefore, present a classification method which obtains spectral as well as spatial features using convolutional neural network (CNN) model and then a logistic regression (LR) classifier that uses the activation function softmax for predicting classification results. Our proposed method is compared with considered techniques and tests on HSIs, namely, Indian pines and Pavia University, which have shown better performance regarding parameters such as overall accuracy (OA), average accuracy (AA), and kappa coefficient (K).
Innovations in Computer Science and Engineering, 2018
Lecture Notes in Electrical Engineering, 2020
The agriculture sector is essential for every country because it provides a basic income to a lar... more The agriculture sector is essential for every country because it provides a basic income to a large number of people and food as well, which is a fundamental requirement to survive on this planet. We see as time passes, significant changes come in the present era, which begins with Green Revolution. Due to improper knowledge of plant diseases, farmers use fertilizers in excess, which ultimately degrade the quality of food. Earlier farmers use experts to determine the type of plant disease, which was expensive and time consuming. In today's time, image processing is used to recognize and catalog plant diseases using the lesion region of plant leaf, and there are different modus-operandi for plant disease scent from leaf using neural networks (NN), support vector machine (SVM), and others. In this paper, we improve the architecture of the neural networking by working on ten different types of training algorithms and the proper choice of neurons in the concealed layer. Our proposed approach gives 98.30% accuracy on general plant leaf disease and 100% accuracy on specific plant leaf disease based on Bayesian regularization, automation of cluster, and without overfitting on considered plant diseases over various other implemented methods.
Lecture Notes in Electrical Engineering, 2018
2020 IEEE International Conference on Power Electronics, Drives and Energy Systems (PEDES), 2020
The paper presents the design, analysis and control of a Canonical Switching Inverse Buck-Boost C... more The paper presents the design, analysis and control of a Canonical Switching Inverse Buck-Boost Converter (CSwI-BBC) based electric vehicle drive system. The presented CSwI-BBC converter topology is capable of performing buck and boost both operation upon the input voltage and thereby offers a wide operation range for the speed control of brushless DC motor (BLDCM) drive using DC link control algorithm. The presented system in closed- loop is capable to controlling the speed of drive by adjusting the reference DC link voltage. The scheme is mathematically modeled in this paper and Exhaustive performance analysis of CSwI-BBC based electrical vehicle drive system has been presented through the simulation results acquired in MATLAB environment. The developed Simulink model is subjected to different operating conditions and performance of the system is found quite satisfactory under steady-state and dynamic test conditions.
2018 Second International Conference on Intelligent Computing and Control Systems (ICICCS), 2018
Classification of hyperspectral images (HSI) signifies giving each and every pixel in the image a... more Classification of hyperspectral images (HSI) signifies giving each and every pixel in the image a class on the basis of some information obtained. Convolutional neural networks (CNNs), in recent years, have been widely explored and applied for HSI classification as they have shown better performance than other existing techniques in terms of accuracy of classification. Most of the existing techniques such as traditional techniques using manually crafted features, feature reduction techniques like independent component analysis (ICA), decision trees (DT) and support vector machines (SVMs) consider only the spectral information for classification. They are unable to utilize the spatial information adequately which hinders classification performance. CNN has shown great promise as it is able to conceive the spatial information along with the spectral information thus enhancing classification accuracy. However, as HSI has very high dimensionality and insufficient samples for training, e...
Uploads
Papers by Dr. Mantosh Biswas