Nisha Vellarayil Mohandas
Research Subjects
Nisha Vellarayil Mohandas has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Sajidha Syed Azimuddin
Vellore Institute of Technology, Chennai, India
5 shared publications - A. Vijayalakshmi
Vellore Institute of Technology, School of Computer Science and Engineering, Vandalur - Kelambakkam Road, Chennai, Tamil Nadu, 600127, India
1 shared publication - Aakif Mairaj
Indiana University, Kokomo, USA
1 shared publication Aditya Kshatriya
VIT Chennai
1 shared publication- Amit Kumar Tyagi
National Forensic Sciences University, Gandhinagar, India
1 shared publication B. Mahalakshmi
VIT University
1 shared publication
Latest Publications
- Open Access
- 23 November 2023
- Published by Springer Nature in Research Square
Meta Learning for Enhanced Web Security Against Malicious URLs
The identification of malicious Uniform Resource Locators (URLs) plays a pivotal role in strengthening network and cyber security. Over the years, the Internet has increasingly become a breeding ground for a multitude of cybercrimes. Malicious URLs can disrupt network performance, compromise data integrity, and introduce vulnerabilities into the system, impacting the network’s overall security and reliability. Taking precautions and implementing preventative techniques is crucial for shielding users from cyberattacks. This study utilizes meta-learning, which is a machine learning technique that learns from the predictions of other machine learning algorithms to make better decisions. For a comprehensive evaluation, the ISCX-URL-2016 dataset was utilized, a resource curated by the Canadian Institute for Cyber Security. The proposed approach focuses on utilizing URL strings as the primary source of information for detecting malicious content.The performance of different machine learning algorithms, including Histogram-based Gradient Boosting Classification Tree, XGBoost Classifier, CatBoost Classifier, Random Forest Classifier and more, were compared for the detection of malicious URLs. The two best performing base classifiers were identified and evaluated as potential meta-classifiers. Various combinations of base classifiers were then tested in conjunction with these two selected meta-classifiers. A significant improvement in the accuracy was observed and the best accuracy of 98.25% was achieved using All Base Classifiers + Meta Classifier XGBoost. Closely followed by the Three Best Base Classifiers + Meta Classifier XGBoost, which achieved an accuracy of 98.2%.
- 23 November 2023
- Published by Springer Nature in Research Square
Open AccessMeta Learning for Enhanced Web Security Against Malicious URLs
The identification of malicious Uniform Resource Locators (URLs) plays a pivotal role in strengthening network and cyber security. Over the years, the Internet has increasingly become a breeding ground for a multitude of cybercrimes. Malicious URLs can disrupt network performance, compromise data integrity, and introduce vulnerabilities into the system, impacting the network’s overall security and reliability. Taking precautions and implementing preventative techniques is crucial for shielding users from cyberattacks. This study utilizes meta-learning, which is a machine learning technique that learns from the predictions of other machine learning algorithms to make better decisions. For a comprehensive evaluation, the ISCX-URL-2016 dataset was utilized, a resource curated by the Canadian Institute for Cyber Security. The proposed approach focuses on utilizing URL strings as the primary source of information for detecting malicious content.The performance of different machine learning algorithms, including Histogram-based Gradient Boosting Classification Tree, XGBoost Classifier, CatBoost Classifier, Random Forest Classifier and more, were compared for the detection of malicious URLs. The two best performing base classifiers were identified and evaluated as potential meta-classifiers. Various combinations of base classifiers were then tested in conjunction with these two selected meta-classifiers. A significant improvement in the accuracy was observed and the best accuracy of 98.25% was achieved using All Base Classifiers + Meta Classifier XGBoost. Closely followed by the Three Best Base Classifiers + Meta Classifier XGBoost, which achieved an accuracy of 98.2%.
- 7 June 2023
- Published by Springer Nature in Journal of Ambient Intelligence and Humanized Computing
- Vol. 14 (8) , 11195-11205
Autonomous face mask detection using single shot multibox detector, and ResNet-50 with identity retrieval through face matching using deep siamese neural network
The COVID-19 pandemic poses a global health challenge. The World Health Organization states that face masks are proven to be effective, especially in public areas. Real-time monitoring of face masks is challenging and exhaustive for humans. To reduce human effort and to provide an enforcement mechanism, an autonomous system has been proposed to detect non-masked people and retrieve their identity using computer vision. The proposed method introduces a novel and efficient method that involves fine-tuning the pre-trained ResNet-50 model with a new head layer for classification between masked and non-masked people. The classifier is trained using adaptive momentum optimization algorithm with decaying learning rate and binary cross-entropy loss. Data augmentation and dropout regularization are employed to achieve best convergence. During real-time application of our classifier on videos, a Caffe face detector model based on Single Shot MultiBox Detector is used to extract the face regions of interest from each frame, on which the trained classifier is applied for detecting the non-masked people. The faces of these people are then captured, which is passed on to a deep siamese neural network, based on VGG-Face model for face matching. The captured faces are compared with the reference images from the database, by extracting the features and calculating cosine distance. If the faces match, the details of that person are retrieved from the database and displayed on the web application. The proposed method has secured best results where the trained classifier has achieved 99.74% accuracy, and the identity retrieval model achieved 98.24% accuracy.
- 7 June 2023
- Published by Springer Nature in Journal of Ambient Intelligence and Humanized Computing
- Vol. 14 (8) , 11195-11205
Autonomous face mask detection using single shot multibox detector, and ResNet-50 with identity retrieval through face matching using deep siamese neural network
The COVID-19 pandemic poses a global health challenge. The World Health Organization states that face masks are proven to be effective, especially in public areas. Real-time monitoring of face masks is challenging and exhaustive for humans. To reduce human effort and to provide an enforcement mechanism, an autonomous system has been proposed to detect non-masked people and retrieve their identity using computer vision. The proposed method introduces a novel and efficient method that involves fine-tuning the pre-trained ResNet-50 model with a new head layer for classification between masked and non-masked people. The classifier is trained using adaptive momentum optimization algorithm with decaying learning rate and binary cross-entropy loss. Data augmentation and dropout regularization are employed to achieve best convergence. During real-time application of our classifier on videos, a Caffe face detector model based on Single Shot MultiBox Detector is used to extract the face regions of interest from each frame, on which the trained classifier is applied for detecting the non-masked people. The faces of these people are then captured, which is passed on to a deep siamese neural network, based on VGG-Face model for face matching. The captured faces are compared with the reference images from the database, by extracting the features and calculating cosine distance. If the faces match, the details of that person are retrieved from the database and displayed on the web application. The proposed method has secured best results where the trained classifier has achieved 99.74% accuracy, and the identity retrieval model achieved 98.24% accuracy.
- 7 June 2023
- Published by Springer Nature
- p. 295-306
Using Machine Learning to Detect Botnets in Network Traffic
- 7 June 2023
- Published by Springer Nature
- p. 295-306
Using Machine Learning to Detect Botnets in Network Traffic
- Open Access
- 1 January 2023
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 2946 (1)
A comparison of speech-based grocery item classification for visually impaired people in Indian native languages
The ubiquitous usage of online shopping over time has increased its prominence. However, the lack of user-friendly UI for visually impaired people hinders the e-commerce shopping experience for them. Although, assistive technologies have become more available for the visually impaired, some features like a smooth User Interface and descriptions of products are fully not available for a complete e-shopping experience. Hence, to provide a wholesome support for the visually impaired to utilize the online e-shopping websites, a total voice-based implementation of online shopping using Indian languages (Hindi and Tamil) using CNN (convolutional neural network model) has been proposed. This proposed model is a speech-based classification model of grocery items using CNN. A comparison of the proposed CNN models (Hindi and Tamil) are made using three different activation functions, namely ReLu, sigmoid, and Softplus functions. It has been found that the proposed CNN model using ReLu activation function gives the highest training accuracy of 99.2 % and 97.4 for Hindi and Tamil languages. Also, it gives the highest testing accuracy of 94.5% for Hindi and 90.5% for Tamil languages. A comparison is done between the proposed CNN model and the open-source speech-to-text Google API merged with TF-IDF using the generated dataset. It has been found that, that proposed CNN model outperforms the Google API approach while translating the voice-based requests of the grocery items.
- 1 January 2023
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 2946 (1)
Open AccessA comparison of speech-based grocery item classification for visually impaired people in Indian native languages
The ubiquitous usage of online shopping over time has increased its prominence. However, the lack of user-friendly UI for visually impaired people hinders the e-commerce shopping experience for them. Although, assistive technologies have become more available for the visually impaired, some features like a smooth User Interface and descriptions of products are fully not available for a complete e-shopping experience. Hence, to provide a wholesome support for the visually impaired to utilize the online e-shopping websites, a total voice-based implementation of online shopping using Indian languages (Hindi and Tamil) using CNN (convolutional neural network model) has been proposed. This proposed model is a speech-based classification model of grocery items using CNN. A comparison of the proposed CNN models (Hindi and Tamil) are made using three different activation functions, namely ReLu, sigmoid, and Softplus functions. It has been found that the proposed CNN model using ReLu activation function gives the highest training accuracy of 99.2 % and 97.4 for Hindi and Tamil languages. Also, it gives the highest testing accuracy of 94.5% for Hindi and 90.5% for Tamil languages. A comparison is done between the proposed CNN model and the open-source speech-to-text Google API merged with TF-IDF using the generated dataset. It has been found that, that proposed CNN model outperforms the Google API approach while translating the voice-based requests of the grocery items.
- 9 September 2022
- Published by Springer Nature
- p. 329-346
Sanskriti—A Distributed E-Commerce Site Implementation Using BlockChain
- 9 September 2022
- Published by Springer Nature
- p. 329-346
Sanskriti—A Distributed E-Commerce Site Implementation Using BlockChain
- 17 December 2021
- Published by Springer Nature
- p. 507-513
Person Re-identification Using Deep Learning with Mask-RCNN
- 17 December 2021
- Published by Springer Nature
- p. 507-513
Person Re-identification Using Deep Learning with Mask-RCNN
- Open Access
- 25 December 2018
- Published by Institute of Advanced Engineering and Science in Indonesian Journal of Electrical Engineering and Computer Science
- Vol. 14 (1) , 471-477
A symmetry based anomaly detection in brain using cellular automata for computer aided diagnosis
Computer aided diagnosis (CAD) is an advancing technology in medical imaging. CAD acts as an additional computing power for doctors to interpret the medical images which leads to a more accurate diagnosis of the disease.CAD system increases the chances of detection of brain lesions by assisting the physicians in decreasing the observational oversight in the early stage of diseases.This paper focuses on the development of a cellular automata based model to find the anomaly prone areas in human brains.Because of the bilateral symmetric nature of human brain, a symmetry based cellular automata model is proposed.An algorithm is designed based on the proposed model to detect the anomaly prone areas in brain images. The proposed model can be a standalone model or it can be incorporated to a sophisticated computer aided diagnosis system. By incorporating asymmetry information into a computer aided diagnosis system, enhances its performance in identifying the anomalies exists in bilaterally symmetrical brain images.
- 25 December 2018
- Published by Institute of Advanced Engineering and Science in Indonesian Journal of Electrical Engineering and Computer Science
- Vol. 14 (1) , 471-477
Open AccessA symmetry based anomaly detection in brain using cellular automata for computer aided diagnosis
Computer aided diagnosis (CAD) is an advancing technology in medical imaging. CAD acts as an additional computing power for doctors to interpret the medical images which leads to a more accurate diagnosis of the disease.CAD system increases the chances of detection of brain lesions by assisting the physicians in decreasing the observational oversight in the early stage of diseases.This paper focuses on the development of a cellular automata based model to find the anomaly prone areas in human brains.Because of the bilateral symmetric nature of human brain, a symmetry based cellular automata model is proposed.An algorithm is designed based on the proposed model to detect the anomaly prone areas in brain images. The proposed model can be a standalone model or it can be incorporated to a sophisticated computer aided diagnosis system. By incorporating asymmetry information into a computer aided diagnosis system, enhances its performance in identifying the anomalies exists in bilaterally symmetrical brain images.
- Open Access
Classification of Two Dimensional Cellular Automata Rules for Symmetric Pattern Generation
Cellular automata (CA) are parallel computational models that comprise of a grid of cells. CA is mainly used for modeling complex systems in various fields, where the geometric structure of the lattices is different. In the absence of a CA model to accommodate different types of lattices in CA, an angle-based CA model is proposed to accommodate various lattices. In the proposed model, the neighborhood structure in a two dimensional cellular automata (2D-CA) is viewed as a star graph. The vertices of the proposed graph are determined by a parameter, angle . Based on the angle , the neighborhood of the CA, which is treated as the vertices of the graph, varies. So this model is suitable for the representation of different types of two dimensional lattices such as square lattice, rectangular lattice, hexagonal lattice, etc. in CA. A mathematical model is formulated for representing CA rules which suit for different types of symmetric lattices. The star graph representation helps to find out the internal symmetries exists in CA rules. Classification of CA rules based on the symmetry exists in the rules, which generates symmetric patterns are discussed in this work.Open AccessClassification of Two Dimensional Cellular Automata Rules for Symmetric Pattern Generation
Cellular automata (CA) are parallel computational models that comprise of a grid of cells. CA is mainly used for modeling complex systems in various fields, where the geometric structure of the lattices is different. In the absence of a CA model to accommodate different types of lattices in CA, an angle-based CA model is proposed to accommodate various lattices. In the proposed model, the neighborhood structure in a two dimensional cellular automata (2D-CA) is viewed as a star graph. The vertices of the proposed graph are determined by a parameter, angle . Based on the angle , the neighborhood of the CA, which is treated as the vertices of the graph, varies. So this model is suitable for the representation of different types of two dimensional lattices such as square lattice, rectangular lattice, hexagonal lattice, etc. in CA. A mathematical model is formulated for representing CA rules which suit for different types of symmetric lattices. The star graph representation helps to find out the internal symmetries exists in CA rules. Classification of CA rules based on the symmetry exists in the rules, which generates symmetric patterns are discussed in this work.