Peer-to-peer (P2P) systems are becoming a popular means of streaming audio and video content but ... more Peer-to-peer (P2P) systems are becoming a popular means of streaming audio and video content but they are prone to bandwidth starvation if selfish peers do not contribute bandwidth to other peers. We prove that an incentive mechanism can be created for a live streaming P2P protocol while preserving the asymptotic properties of randomized gossip-based streaming. In order to show the utility of our result, we adapt a distributed incentive scheme from P2P file storage literature to the live streaming scenario. We provide simulation results that confirm the ability to achieve a constant download rate (in time, per peer) that is needed for streaming applications on peers. The incentive scheme fairly differentiates peers' download rates according to the amount of useful bandwidth they contribute back to the P2P system, thus creating a powerful quality-of-service incentive for peers to contribute bandwidth to other peers. We propose a functional architecture and protocol format for a g...
Surveillance footage can catch a wide range of realistic anomalies. This research suggests using ... more Surveillance footage can catch a wide range of realistic anomalies. This research suggests using a weakly supervised strategy to avoid annotating anomalous segments in training videos, which is time consuming. In this approach only video level labels are used to obtain frame level anomaly scores. Weakly supervised video anomaly detection (WSVAD) suffers from the wrong identification of abnormal and normal instances during the training process. Therefore it is important to extract better quality features from the available videos. With this motivation, the present paper uses better quality transformer-based features named Videoswin Features followed by the attention layer based on dilated convolution and self attention to capture long and short range dependencies in temporal domain. This gives us a better understanding of available videos. The proposed framework is validated on real-world dataset i.e. ShanghaiTech Campus dataset which results in competitive performance than current stateof-the-art methods.
2019 IEEE Students Conference on Engineering and Systems (SCES)
In a dynamic environment, detection of changes in concept is very important for prediction and de... more In a dynamic environment, detection of changes in concept is very important for prediction and decision-based applications. Concept drift detection helps decision makers to perform smarter maintenance and operations at an appropriate time. In the context of Emails, concept drift is defined as how the concepts in Emails are changing with time. In this paper, a novel method is presented based on intention based segmentation to calculate concept drift in Email dataset. In intention based segmentation instead of comparing the contents of email as whole, it is divided into segments focusing on same intention and compared. Division of sentences is done on the basis of voices and tenses with the help of POS Tagging. After that, Hierarchical Clustering is used for clustering of segments and comparison is done using Vector Space Model. Importance of concept drift detection in Email domain is to find the emails which will be of less interest to user. It also helps to find the current interests of user by analyzing the changes in concepts over time.
2021 IEEE International Conference on Big Data (Big Data)
Adversarial attacks against deep learning models have gained significant attention and recent wor... more Adversarial attacks against deep learning models have gained significant attention and recent works have proposed explanations for the existence of adversarial examples and techniques to defend the models against these attacks. Attention in computer vision has been used to incorporate focused learning of important features and has led to improved accuracy. Recently, models with attention mechanisms have been proposed to enhance adversarial robustness. Following this context, this work aims at a general understanding of the impact of attention on adversarial robustness. This work presents a comparative study of adversarial robustness of non-attention and attention based image classification models trained on CIFAR-10, CIFAR-100 and Fashion MNIST datasets under the popular white box and black box attacks. The experimental results show that the robustness of attention based models may be dependent on the datasets used i.e. the number of classes involved in the classification. In contrast to the datasets with less number of classes, attention based models are observed to show better robustness towards classification.
The advancements in deep learning technologies have produced immense contribution to biomedical i... more The advancements in deep learning technologies have produced immense contribution to biomedical image analysis applications. With breast cancer being the common deadliest disease among women, early detection is the key means to improve survivability. Medical imaging like ultrasound presents an excellent visual representation of the functioning of the organs; however, for any radiologist analysing such scans is challenging and time consuming which delays the diagnosis process. Although various deep learning based approaches are proposed that achieved promising results, the present article introduces an efficient residual crossspatial attention guided inception U-Net (RCA-IUnet) model with minimal training parameters for tumor segmentation using breast ultrasound imaging to further improve the segmentation performance of varying tumor sizes. The RCA-IUnet model follows U-Net topology with residual inception depth-wise separable convolution and hybrid pooling (max pooling and spectral pooling) layers. In addition, cross-spatial attention filters are added to suppress the irrelevant features and focus on the target structure. The segmentation performance of the proposed model is validated on two publicly available datasets using standard segmentation evaluation metrics, where it outperformed the other state-of-the-art segmentation models.
It is evident that the infection with this severe acute respiratory syndrome coronavirus 2 (SARS-... more It is evident that the infection with this severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) starts with the upper respiratory tract and as the virus grows, the infection can progress to lungs and develop pneumonia. According to the statistics, approximately 14\% of the infected people with COVID-19 have severe cough and shortness of breath due to pneumonia, because as the viral infection increases, it damages the alveoli (small air sacs) and surrounding tissues. The conventional way of COVID-19 diagnosis is reverse transcription polymerase chain reaction (RT-PCR), which is less sensitive during early stages specially, if the patient is asymptomatic that may further lead to more severe pneumonia. To overcome this problem an early diagnosis method is proposed in this paper via one-class classification approach using a novel pinball loss function based one-class support vector machine (PB-OCSVM) considering posteroanterior chest X-ray images. Recently, several automated COV...
Machine Learning (ML) on massive scale datasets, called Big Data, has become a challenge for trad... more Machine Learning (ML) on massive scale datasets, called Big Data, has become a challenge for traditional computing and storage technologies. Henceforth, massive scale ML is an emerging domain of research. Least Square Twin Support Vector Machine (LSTSVM) is a faster variant of Support Vector Machine (SVM). However, it suffers from scalability issues and shows computational and/or storage bottlenecks on massive datasets. Proposed work designs a scalable solution to LSTSVM called Distributed LSTSVM (DLSTSVM). DLSTSVM is designed using distributed parallel computing on top of cluster of multiple machines. After applying horizontal partitioning on massive datasets, DLSTSVM trains it in distributed parallel fashion and finds two non-parallel hyper-planes as decision boundaries for two different classes. MapReduce paradigm is utilized to execute parallel computation on partitioned data in a way that averts memory constraints. Proposed technique achieves computational and storage scalabili...
2019 IEEE Students Conference on Engineering and Systems (SCES), 2019
Predictive Analytics means forecasting future demand using historical data. A robust, accurate an... more Predictive Analytics means forecasting future demand using historical data. A robust, accurate and fast prediction model is the need of the present constantly changing competitive environment. More specifically, for taxi demand prediction several statistical and regression-based models exist, but still, the accuracy is a challenging issue. This paper proposes a three-tier ensemble model for more accurate and faster taxi demand prediction. In the first stage, data cleaning is done followed by Mini Batch K-Means Clustering in the second stage whereas the Random Forest (a bagging technique) is applied at the last stage for demand prediction. The NYC TLC trip dataset is used for experiments, and the results obtained highlight that the proposed model outperforms the statistical models.
Proceeding of the Electrical Engineering Computer Science and Informatics, 2019
Today big data has become the basis of discussion for the organizations. The big task associated ... more Today big data has become the basis of discussion for the organizations. The big task associated with big data stream is coping with its various challenges and performing the appropriate testing for the optimal analysis of the data which may benefit the processing of various activities, especially from a business perspective. Big data term follows the massive volume of data, (might be in units of petabytes or exabytes) exceeding the processing and analytical capacity of the conventional systems and thereby raising the need for analyzing and testing the big data before applications can be put into use. Testing such huge data coming from the various number of sources like the internet, smartphones, audios, videos, media, etc. is a challenge itself. The most favourable solution to test big data follows the automated/programmed approach. This paper outlines the big data characteristics, and various challenges associated with it followed by the approach, strategy, and proposed framework for testing big data applications.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
2016 International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2016
Feature selection is one of the most significant steps in machine learning that reduces the featu... more Feature selection is one of the most significant steps in machine learning that reduces the features space in order to achieve faster learning and yielding simpler models with high accuracy and interpretability. With rapid development in technologies, large scale high dimensional datasets are common today which degrades the performance of traditional feature selection techniques as they suffer with the scalability issues. Parallel feature selection is an obvious solution to deal with this problem. Due to advent of many distributed computing frameworks scalable computing has become a viable strategy in reference to feature selection. Present work proposes a distributed parallel feature selection technique that employs vertical distribution strategy for dataset to exploit parallel computation. It uses information gain filter based ranking method which evaluates multiple disjoint feature subsets of dataset in parallel. The key idea is the distribution of evaluation and rank generation of features over several computing nodes in parallel. Experiments are performed on multiple large scale and high dimension datasets and significant reduction in overall computation time is achieved.
International Journal of Database Theory and Application, 2015
This paper proposes a Multiclass Least Squares Twin Support Vector Machine (MLSTSVM) classifier f... more This paper proposes a Multiclass Least Squares Twin Support Vector Machine (MLSTSVM) classifier for multi-class classification problems. The formulation of MLSTSVM is obtained by extending the formulation of recently proposed binary Least Squares Twin Support Vector Machine (LSTSVM) classifier. For M-class classification problem, the proposed classifier seeks M-non parallel hyper-planes, one for each class, by solving M-linear equations. A regularization term is also added to improve the generalization ability. MLSTSVM works well for both linear and non-linear type of datasets. It is relatively simple and fast algorithm as compared to the other existing approaches. The performance of proposed approach has been evaluated on twelve benchmark datasets. The experimental result demonstrates the validity of proposed MLSTSVM classifier as compared to the typical multi-classifiers based on 'Support Vector Machine' and 'Twin Support Vector Machine'. Statistical analysis of the proposed classifier with existing classifiers is also performed by using Friedman's Test statistic and Nemenyi post hoc techniques.
Developers often collaborate to fix complex bugs, even in open source software systems (OSS) wher... more Developers often collaborate to fix complex bugs, even in open source software systems (OSS) where collaboration largely occurs through discussions in the bug tracker. The implicit Developer Social Networks (DSN) are created as a result of these discussions. Past research has investigated the usefulness of such DSNs in addressing many Software Engineering problems (e.g. Defect Prediction, Evolution of collaboration patterns, etc.). However, the multifaceted nature of DSNs constructed from bug reports data has been ignored in most of the past studies. That is, in most of the past studies, the link among developers exist only if they comment on the same bug report while in reality, the developers may be connected indirectly (e.g. pair of developers are connected even if they comment on two different bug reports which are associated with the same software component). Such unexplored relationships among developers can be used in defining new measures to identify important developers in ...
2019 6th International Conference on Electrical Engineering, Computer Science and Informatics (EECSI)
Today big data has become the basis of discussion for the organizations. The big task associated ... more Today big data has become the basis of discussion for the organizations. The big task associated with big data stream is coping with its various challenges and performing the appropriate testing for the optimal analysis of the data which may benefit the processing of various activities, especially from a business perspective. Big data term follows the massive volume of data, (might be in units of petabytes or exabytes) exceeding the processing and analytical capacity of the conventional systems and thereby raising the need for analyzing and testing the big data before applications can be put into use. Testing such huge data coming from the various number of sources like the internet, smartphones, audios, videos, media, etc. is a challenge itself. The most favourable solution to test big data follows the automated/programmed approach. This paper outlines the big data characteristics, and various challenges associated with it followed by the approach, strategy, and proposed framework for testing big data applications.
The pandemic of novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also known as ... more The pandemic of novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also known as COVID-19 has been spreading worldwide, causing rampant loss of lives. Medical imaging such as computed tomography (CT), X-ray, etc., plays a significant role in diagnosing the patients by presenting the visual representation of the functioning of the organs. However, for any radiologist analyzing such scans is a tedious and time-consuming task. The emerging deep learning technologies have displayed its strength in analyzing such scans to aid in the faster diagnosis of the diseases and viruses such as COVID-19. In the present article, an automated deep learning based model, COVID-19 hierarchical segmentation network (CHS-Net) is proposed that functions as a semantic hierarchical segmenter to identify the COVID-19 infected regions from lungs contour via CT medical imaging using two cascaded residual attention inception U-Net (RAIU-Net) models. RAIU-Net comprises of a residual inception U-Net model with spectral spatial and depth attention network (SSD) that is developed with the contraction and expansion phases of depthwise separable convolutions and hybrid pooling (max and spectral pooling) to efficiently encode and decode the semantic and varying resolution information. The CHS-Net is trained with the segmentation loss function that is the defined as the average of binary cross entropy loss and dice loss to penalize false negative and false positive predictions. The approach is compared with the recently proposed approaches and evaluated using the standard metrics like accuracy, precision, specificity, recall, dice coefficient and Jaccard similarity along with the visualized interpretation of the model prediction with GradCam++ and uncertainty maps. With extensive trials, it is observed that the proposed approach outperformed the recently proposed approaches and effectively segments the COVID-19 infected regions in the lungs.
With the advent of advancements in deep learning approaches, such as deep convolution neural netw... more With the advent of advancements in deep learning approaches, such as deep convolution neural network, residual neural network, adversarial network; U-Net architectures are most widely utilized in biomedical image segmentation to address the automation in identification and detection of the target regions or sub-regions. In recent studies, U-Net based approaches have illustrated state-of-the-art performance in different applications for the development of computer-aided diagnosis systems for early diagnosis and treatment of diseases such as brain tumor, lung cancer, alzheimer, breast cancer, etc., using various modalities. This article contributes in presenting the success of these approaches by describing the U-Net framework, followed by the comprehensive analysis of the U-Net variants by performing (1) inter-modality, and (2) intra-modality categorization to establish better insights into the associated challenges and solutions. Besides, this article also highlights the contribution of U-Net based frameworks in the ongoing pandemic, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also known as COVID-19. Finally, the strengths and similarities of these U-Net variants are analysed along with the challenges involved in biomedical image segmentation to uncover promising future research directions in this area.
Automatic image annotation is a technique by which computer systems automatically assigns appropr... more Automatic image annotation is a technique by which computer systems automatically assigns appropriate Keywords to input digital image. Smart cities are characterized by large volume of data, one of the prominent data types are images. In present research, images of smart cities are collected and then using automatic image annotation, several relevant indexing terms are proposed for every image. Using these terms, images may be categorized and further can be used in appropriate applications. Image annotation is also used in keyword based search applications. Here, the goal is to obtain images according to the input keyword given by the user. To accomplish this, Multi-label image annotation using K-nearest Neighbor (ML-KNN) is used. In this approach, images in the database need to be indexed with the terms in annotation vocabulary. Automatically annotated images of the image database prior to the search may thus be helpful in many situations such as distribution of land, water and residential areas which can aid in effective urban planning and governance.
2017 IEEE International Conference on Industrial and Information Systems (ICIIS), 2017
Transfer learning is an emerging research area which extracts knowledge from one or more than one... more Transfer learning is an emerging research area which extracts knowledge from one or more than one source domains and utilizes this gained knowledge to perform some task in a target domain. It has emerged as a popular topic in recent years, because this technique is considered to be helpful in reducing the cost of labeling. It has many applications on different domains such as Natural Language Processing, Image and Video Processing, etc. The aim is to study transfer learning and implement it for Sentiment Analysis of Tweets by using the knowledge of Yelp reviews. We find that transfer Learning approach is faster than the conventional machine learning approach and give comparable accuracy at much smaller dataset.
Peer-to-peer (P2P) systems are becoming a popular means of streaming audio and video content but ... more Peer-to-peer (P2P) systems are becoming a popular means of streaming audio and video content but they are prone to bandwidth starvation if selfish peers do not contribute bandwidth to other peers. We prove that an incentive mechanism can be created for a live streaming P2P protocol while preserving the asymptotic properties of randomized gossip-based streaming. In order to show the utility of our result, we adapt a distributed incentive scheme from P2P file storage literature to the live streaming scenario. We provide simulation results that confirm the ability to achieve a constant download rate (in time, per peer) that is needed for streaming applications on peers. The incentive scheme fairly differentiates peers' download rates according to the amount of useful bandwidth they contribute back to the P2P system, thus creating a powerful quality-of-service incentive for peers to contribute bandwidth to other peers. We propose a functional architecture and protocol format for a g...
Surveillance footage can catch a wide range of realistic anomalies. This research suggests using ... more Surveillance footage can catch a wide range of realistic anomalies. This research suggests using a weakly supervised strategy to avoid annotating anomalous segments in training videos, which is time consuming. In this approach only video level labels are used to obtain frame level anomaly scores. Weakly supervised video anomaly detection (WSVAD) suffers from the wrong identification of abnormal and normal instances during the training process. Therefore it is important to extract better quality features from the available videos. With this motivation, the present paper uses better quality transformer-based features named Videoswin Features followed by the attention layer based on dilated convolution and self attention to capture long and short range dependencies in temporal domain. This gives us a better understanding of available videos. The proposed framework is validated on real-world dataset i.e. ShanghaiTech Campus dataset which results in competitive performance than current stateof-the-art methods.
2019 IEEE Students Conference on Engineering and Systems (SCES)
In a dynamic environment, detection of changes in concept is very important for prediction and de... more In a dynamic environment, detection of changes in concept is very important for prediction and decision-based applications. Concept drift detection helps decision makers to perform smarter maintenance and operations at an appropriate time. In the context of Emails, concept drift is defined as how the concepts in Emails are changing with time. In this paper, a novel method is presented based on intention based segmentation to calculate concept drift in Email dataset. In intention based segmentation instead of comparing the contents of email as whole, it is divided into segments focusing on same intention and compared. Division of sentences is done on the basis of voices and tenses with the help of POS Tagging. After that, Hierarchical Clustering is used for clustering of segments and comparison is done using Vector Space Model. Importance of concept drift detection in Email domain is to find the emails which will be of less interest to user. It also helps to find the current interests of user by analyzing the changes in concepts over time.
2021 IEEE International Conference on Big Data (Big Data)
Adversarial attacks against deep learning models have gained significant attention and recent wor... more Adversarial attacks against deep learning models have gained significant attention and recent works have proposed explanations for the existence of adversarial examples and techniques to defend the models against these attacks. Attention in computer vision has been used to incorporate focused learning of important features and has led to improved accuracy. Recently, models with attention mechanisms have been proposed to enhance adversarial robustness. Following this context, this work aims at a general understanding of the impact of attention on adversarial robustness. This work presents a comparative study of adversarial robustness of non-attention and attention based image classification models trained on CIFAR-10, CIFAR-100 and Fashion MNIST datasets under the popular white box and black box attacks. The experimental results show that the robustness of attention based models may be dependent on the datasets used i.e. the number of classes involved in the classification. In contrast to the datasets with less number of classes, attention based models are observed to show better robustness towards classification.
The advancements in deep learning technologies have produced immense contribution to biomedical i... more The advancements in deep learning technologies have produced immense contribution to biomedical image analysis applications. With breast cancer being the common deadliest disease among women, early detection is the key means to improve survivability. Medical imaging like ultrasound presents an excellent visual representation of the functioning of the organs; however, for any radiologist analysing such scans is challenging and time consuming which delays the diagnosis process. Although various deep learning based approaches are proposed that achieved promising results, the present article introduces an efficient residual crossspatial attention guided inception U-Net (RCA-IUnet) model with minimal training parameters for tumor segmentation using breast ultrasound imaging to further improve the segmentation performance of varying tumor sizes. The RCA-IUnet model follows U-Net topology with residual inception depth-wise separable convolution and hybrid pooling (max pooling and spectral pooling) layers. In addition, cross-spatial attention filters are added to suppress the irrelevant features and focus on the target structure. The segmentation performance of the proposed model is validated on two publicly available datasets using standard segmentation evaluation metrics, where it outperformed the other state-of-the-art segmentation models.
It is evident that the infection with this severe acute respiratory syndrome coronavirus 2 (SARS-... more It is evident that the infection with this severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) starts with the upper respiratory tract and as the virus grows, the infection can progress to lungs and develop pneumonia. According to the statistics, approximately 14\% of the infected people with COVID-19 have severe cough and shortness of breath due to pneumonia, because as the viral infection increases, it damages the alveoli (small air sacs) and surrounding tissues. The conventional way of COVID-19 diagnosis is reverse transcription polymerase chain reaction (RT-PCR), which is less sensitive during early stages specially, if the patient is asymptomatic that may further lead to more severe pneumonia. To overcome this problem an early diagnosis method is proposed in this paper via one-class classification approach using a novel pinball loss function based one-class support vector machine (PB-OCSVM) considering posteroanterior chest X-ray images. Recently, several automated COV...
Machine Learning (ML) on massive scale datasets, called Big Data, has become a challenge for trad... more Machine Learning (ML) on massive scale datasets, called Big Data, has become a challenge for traditional computing and storage technologies. Henceforth, massive scale ML is an emerging domain of research. Least Square Twin Support Vector Machine (LSTSVM) is a faster variant of Support Vector Machine (SVM). However, it suffers from scalability issues and shows computational and/or storage bottlenecks on massive datasets. Proposed work designs a scalable solution to LSTSVM called Distributed LSTSVM (DLSTSVM). DLSTSVM is designed using distributed parallel computing on top of cluster of multiple machines. After applying horizontal partitioning on massive datasets, DLSTSVM trains it in distributed parallel fashion and finds two non-parallel hyper-planes as decision boundaries for two different classes. MapReduce paradigm is utilized to execute parallel computation on partitioned data in a way that averts memory constraints. Proposed technique achieves computational and storage scalabili...
2019 IEEE Students Conference on Engineering and Systems (SCES), 2019
Predictive Analytics means forecasting future demand using historical data. A robust, accurate an... more Predictive Analytics means forecasting future demand using historical data. A robust, accurate and fast prediction model is the need of the present constantly changing competitive environment. More specifically, for taxi demand prediction several statistical and regression-based models exist, but still, the accuracy is a challenging issue. This paper proposes a three-tier ensemble model for more accurate and faster taxi demand prediction. In the first stage, data cleaning is done followed by Mini Batch K-Means Clustering in the second stage whereas the Random Forest (a bagging technique) is applied at the last stage for demand prediction. The NYC TLC trip dataset is used for experiments, and the results obtained highlight that the proposed model outperforms the statistical models.
Proceeding of the Electrical Engineering Computer Science and Informatics, 2019
Today big data has become the basis of discussion for the organizations. The big task associated ... more Today big data has become the basis of discussion for the organizations. The big task associated with big data stream is coping with its various challenges and performing the appropriate testing for the optimal analysis of the data which may benefit the processing of various activities, especially from a business perspective. Big data term follows the massive volume of data, (might be in units of petabytes or exabytes) exceeding the processing and analytical capacity of the conventional systems and thereby raising the need for analyzing and testing the big data before applications can be put into use. Testing such huge data coming from the various number of sources like the internet, smartphones, audios, videos, media, etc. is a challenge itself. The most favourable solution to test big data follows the automated/programmed approach. This paper outlines the big data characteristics, and various challenges associated with it followed by the approach, strategy, and proposed framework for testing big data applications.
This is a PDF file of an article that has undergone enhancements after acceptance, such as the ad... more This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
2016 International Conference on Advances in Computing, Communications and Informatics (ICACCI), 2016
Feature selection is one of the most significant steps in machine learning that reduces the featu... more Feature selection is one of the most significant steps in machine learning that reduces the features space in order to achieve faster learning and yielding simpler models with high accuracy and interpretability. With rapid development in technologies, large scale high dimensional datasets are common today which degrades the performance of traditional feature selection techniques as they suffer with the scalability issues. Parallel feature selection is an obvious solution to deal with this problem. Due to advent of many distributed computing frameworks scalable computing has become a viable strategy in reference to feature selection. Present work proposes a distributed parallel feature selection technique that employs vertical distribution strategy for dataset to exploit parallel computation. It uses information gain filter based ranking method which evaluates multiple disjoint feature subsets of dataset in parallel. The key idea is the distribution of evaluation and rank generation of features over several computing nodes in parallel. Experiments are performed on multiple large scale and high dimension datasets and significant reduction in overall computation time is achieved.
International Journal of Database Theory and Application, 2015
This paper proposes a Multiclass Least Squares Twin Support Vector Machine (MLSTSVM) classifier f... more This paper proposes a Multiclass Least Squares Twin Support Vector Machine (MLSTSVM) classifier for multi-class classification problems. The formulation of MLSTSVM is obtained by extending the formulation of recently proposed binary Least Squares Twin Support Vector Machine (LSTSVM) classifier. For M-class classification problem, the proposed classifier seeks M-non parallel hyper-planes, one for each class, by solving M-linear equations. A regularization term is also added to improve the generalization ability. MLSTSVM works well for both linear and non-linear type of datasets. It is relatively simple and fast algorithm as compared to the other existing approaches. The performance of proposed approach has been evaluated on twelve benchmark datasets. The experimental result demonstrates the validity of proposed MLSTSVM classifier as compared to the typical multi-classifiers based on 'Support Vector Machine' and 'Twin Support Vector Machine'. Statistical analysis of the proposed classifier with existing classifiers is also performed by using Friedman's Test statistic and Nemenyi post hoc techniques.
Developers often collaborate to fix complex bugs, even in open source software systems (OSS) wher... more Developers often collaborate to fix complex bugs, even in open source software systems (OSS) where collaboration largely occurs through discussions in the bug tracker. The implicit Developer Social Networks (DSN) are created as a result of these discussions. Past research has investigated the usefulness of such DSNs in addressing many Software Engineering problems (e.g. Defect Prediction, Evolution of collaboration patterns, etc.). However, the multifaceted nature of DSNs constructed from bug reports data has been ignored in most of the past studies. That is, in most of the past studies, the link among developers exist only if they comment on the same bug report while in reality, the developers may be connected indirectly (e.g. pair of developers are connected even if they comment on two different bug reports which are associated with the same software component). Such unexplored relationships among developers can be used in defining new measures to identify important developers in ...
2019 6th International Conference on Electrical Engineering, Computer Science and Informatics (EECSI)
Today big data has become the basis of discussion for the organizations. The big task associated ... more Today big data has become the basis of discussion for the organizations. The big task associated with big data stream is coping with its various challenges and performing the appropriate testing for the optimal analysis of the data which may benefit the processing of various activities, especially from a business perspective. Big data term follows the massive volume of data, (might be in units of petabytes or exabytes) exceeding the processing and analytical capacity of the conventional systems and thereby raising the need for analyzing and testing the big data before applications can be put into use. Testing such huge data coming from the various number of sources like the internet, smartphones, audios, videos, media, etc. is a challenge itself. The most favourable solution to test big data follows the automated/programmed approach. This paper outlines the big data characteristics, and various challenges associated with it followed by the approach, strategy, and proposed framework for testing big data applications.
The pandemic of novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also known as ... more The pandemic of novel severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also known as COVID-19 has been spreading worldwide, causing rampant loss of lives. Medical imaging such as computed tomography (CT), X-ray, etc., plays a significant role in diagnosing the patients by presenting the visual representation of the functioning of the organs. However, for any radiologist analyzing such scans is a tedious and time-consuming task. The emerging deep learning technologies have displayed its strength in analyzing such scans to aid in the faster diagnosis of the diseases and viruses such as COVID-19. In the present article, an automated deep learning based model, COVID-19 hierarchical segmentation network (CHS-Net) is proposed that functions as a semantic hierarchical segmenter to identify the COVID-19 infected regions from lungs contour via CT medical imaging using two cascaded residual attention inception U-Net (RAIU-Net) models. RAIU-Net comprises of a residual inception U-Net model with spectral spatial and depth attention network (SSD) that is developed with the contraction and expansion phases of depthwise separable convolutions and hybrid pooling (max and spectral pooling) to efficiently encode and decode the semantic and varying resolution information. The CHS-Net is trained with the segmentation loss function that is the defined as the average of binary cross entropy loss and dice loss to penalize false negative and false positive predictions. The approach is compared with the recently proposed approaches and evaluated using the standard metrics like accuracy, precision, specificity, recall, dice coefficient and Jaccard similarity along with the visualized interpretation of the model prediction with GradCam++ and uncertainty maps. With extensive trials, it is observed that the proposed approach outperformed the recently proposed approaches and effectively segments the COVID-19 infected regions in the lungs.
With the advent of advancements in deep learning approaches, such as deep convolution neural netw... more With the advent of advancements in deep learning approaches, such as deep convolution neural network, residual neural network, adversarial network; U-Net architectures are most widely utilized in biomedical image segmentation to address the automation in identification and detection of the target regions or sub-regions. In recent studies, U-Net based approaches have illustrated state-of-the-art performance in different applications for the development of computer-aided diagnosis systems for early diagnosis and treatment of diseases such as brain tumor, lung cancer, alzheimer, breast cancer, etc., using various modalities. This article contributes in presenting the success of these approaches by describing the U-Net framework, followed by the comprehensive analysis of the U-Net variants by performing (1) inter-modality, and (2) intra-modality categorization to establish better insights into the associated challenges and solutions. Besides, this article also highlights the contribution of U-Net based frameworks in the ongoing pandemic, severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) also known as COVID-19. Finally, the strengths and similarities of these U-Net variants are analysed along with the challenges involved in biomedical image segmentation to uncover promising future research directions in this area.
Automatic image annotation is a technique by which computer systems automatically assigns appropr... more Automatic image annotation is a technique by which computer systems automatically assigns appropriate Keywords to input digital image. Smart cities are characterized by large volume of data, one of the prominent data types are images. In present research, images of smart cities are collected and then using automatic image annotation, several relevant indexing terms are proposed for every image. Using these terms, images may be categorized and further can be used in appropriate applications. Image annotation is also used in keyword based search applications. Here, the goal is to obtain images according to the input keyword given by the user. To accomplish this, Multi-label image annotation using K-nearest Neighbor (ML-KNN) is used. In this approach, images in the database need to be indexed with the terms in annotation vocabulary. Automatically annotated images of the image database prior to the search may thus be helpful in many situations such as distribution of land, water and residential areas which can aid in effective urban planning and governance.
2017 IEEE International Conference on Industrial and Information Systems (ICIIS), 2017
Transfer learning is an emerging research area which extracts knowledge from one or more than one... more Transfer learning is an emerging research area which extracts knowledge from one or more than one source domains and utilizes this gained knowledge to perform some task in a target domain. It has emerged as a popular topic in recent years, because this technique is considered to be helpful in reducing the cost of labeling. It has many applications on different domains such as Natural Language Processing, Image and Video Processing, etc. The aim is to study transfer learning and implement it for Sentiment Analysis of Tweets by using the knowledge of Yelp reviews. We find that transfer Learning approach is faster than the conventional machine learning approach and give comparable accuracy at much smaller dataset.
Uploads
Papers by Sonali Agarwal