Browsing by Author "Venkatesan, M."

Now showing 1 - 20 of 41

A Bag-of-Phonetic-Codes Modelfor Cyber-Bullying Detection in Twitter
(Institute of Electrical and Electronics Engineers Inc., 2018) Shekhar, A.; Venkatesan, M.
Social networking sites such as Twitter, Facebook, MySpace, Instagram are emerging as a strong medium of communication these days. These have become a part and parcel of daily life. People can express their thoughts and activities among their social circle with brings them closer to their community. However this freedom of expression has its drawbacks. Sometimes people show their aggression on Social Media which in turn hurts the sentiments of the targeted victims. Certain forms of cyber-bullying are sexual, racial and physical disability based. Hence a proper surveillance is necessary to tackle such situations. Twitter as a micro-blogging site sees cyber abuse on a daily basis. However, tweets are raw texts; containing a lot of misspelled words and censored words. This paper proposes a novel method to detect cyber-bullying, a Bag-of-Phonetic-Codes model. Using pronunciation of words as features can rectify misspelled words and can identify censored words. Correctly identifying duplicate words can lead to smaller vocabulary of words, thereby reducing the feature space. The inspiration for this proposed work is drawn from the famous Bag-of-Words model for extracting textual features. Phonetic code generation has been done using the Soundex Algorithm. Besides the proposed model, experiments were carried out with both supervised and unsupervised machine learning approaches on multiple datasets to understand the approaches and challenges in the domain of cyber-bullying detection. Â© 2018 IEEE.
An efficient image retrieval system for remote sensing images using deep hashing network
(Springer, 2020) Valaboju, S.; Venkatesan, M.
Due to the huge increase in volumes of remote sensing images, there is a requirement for retrieval systems which maintain the retrieval accuracy and efficiency which requires better learning of features and the binary hash codes which better discriminate the images of different classes of images. The existing retrieval systems for remote sensing images use CNNs for feature learning which fails to preserve the spatial properties of an image which in turn affect the quality of binary hash code and the retrieval performance. This Paper tries to address the above goals by using (1) Extracting Hierarchical features of convolutional neural network and using them to sequential learning to better learn the features preserving spatial and semantic properties. (2) Use lossless triplet loss with two more loss functions to generate the binary hash codes which better discriminate the images of different classes. The proposed architecture consists of three phases: (1) Fine-tuning a pre-trained model. (2) Extracting the hierarchical features of convolutional neural network. (3) Using those features to train the deep learning-based hashing network. Experiments are conducted on a publicly available dataset UCMD and show that when hierarchial convolutional features are considered there is a significant improvement in performance. Â© Springer Nature Singapore Pte Ltd 2020.
An Efficient Technique for Three-Dimensional Image Visualization through Two-Dimensional Images for Medical Data
(De Gruyter peter.golla@degruyter.com, 2020) Gunasekaran, G.; Venkatesan, M.
The main idea behind this work is to present three-dimensional (3D) image visualization through two-dimensional (2D) images that comprise various images. 3D image visualization is one of the essential methods for excerpting data from given pieces. The main goal of this work is to figure out the outlines of the given 3D geometric primitives in each part, and then integrate these outlines or frames to reconstruct 3D geometric primitives. The proposed technique is very useful and can be applied to many kinds of images. The experimental results showed a very good determination of the reconstructing process of 2D images. © 2020 Walter de Gruyter GmbH, Berlin/Boston 2020.
An enlarged map-reduce using 2logmean-PSO optimization for unstructured data
(SAGE Publications Inc., 2024) Kanimozhi, K.V.; Rajakumar, R.; Venkatesan, M.
Text clustering system is a proper technique which mainly segments large measure of textual documents into clusters. The size of the material influences the clustering of text by reducing its performance. In this manner, the textual document comprises sparse and uninformative features, and thus raises the computational time and decreases the execution of primary clustering process. Feature selection is a crucial system to choose another subset of instructive text feature to enhance text clustering execution and diminish computational time. The implemented model proposes a 2logmean-particle swarm optimization algorithm for the unstructured text clustering. In this newly proposed technique, all the texts are initially converted into ASCII value, and then by using the particle swarm optimization, the document text is clustered. The outcomes display that clustering accuracy of the implemented method is high compared to the existing K-means algorithm. Furthermore, performances of newly implemented techniques are evaluated concerning scalability, less computation speed with colossal dimensionality reduction. © The Author(s) 2019.
Applying Multi-Modal Quantum Deep Learning Algorithms for Enhanced Fake News Detection
(Jagiellonian University, 2025) Aishwarya, C.; Venkatesan, M.; Prabhavathy; Akanksha, D.
The pervasive spread of fake news across digital platforms has prompted the development of advanced detection systems. This review surveys and compares state-of-the-art multimodal deep learning models, including SpotFake, BDANN, MVAE, EANN, and the attention-based model by Guo et al., across benchmark datasets such as Twitter and Weibo. We present detailed performance comparisons, with SpotFake achieving an accuracy of 86.1% on the Twitter dataset. Key contributions of this review include the introduction of taxonomy tables based on fusion strategy and model architecture, a critical comparison of early, late, and hybrid fusion mechanisms, and a comprehensive evaluation of cross-modal generalization capabilities. In addition, we explore recent efforts in Quantum Machine Learning (QML), highlighting variational quantum circuits and hybrid quantum-classical models as promising approaches for enhancing scalability and efficiency. This work serves as a roadmap for building robust, interpretable, and scalable fake news detection systems that integrate both classical and quantum techniques. Povzetek: Pregled primerja multimodalne modele za zaznavanje lažnih novic (SpotFake, BDANN, MVAE, EANN, Guo) na Twitterju in Weibou ter predstavi taksonomije fuzije in arhitektur. Obravnava tudi obetavne kvantne pristope, ki lahko izboljšajo skalabilnost in u?inkovitost prihodnjih sistemov. © (2026). All right reserved.
Attention-Based Bitemporal Image Deep Feature-Level Change Detection for High Resolution Imagery
(Springer Science and Business Media Deutschland GmbH, 2023) Naik, N.; Chandrasekaran, K.; Venkatesan, M.; Prabhavathy, P.
To understand the intricacy of changes on the surface of the land, change detection is an important field in the area of remote sensing. Bitemporal remote sensing images are resourceful information to perform the analysis related to classification and change detection. Most of the architectures proposed for improving the performance of change detection in high resolution images pose a challenge due to composite texture features and finer image details. In this paper, we propose a change detection approach for bitemporal images using supervised learning. Firstly, extraction of the features is performed using a pretrained neural network. Then, the extracted features are provided to a (DSDEN) deep supervised-based difference evaluation network. Then, channel and spatial-based attention components are incorporated for fusing the difference image features with the deep features of raw images for the reconstruction of the final change map. The experimental evaluation on public â€œLEVIR-CDâ€ dataset demonstrates the effectiveness and superiority over traditional methods. Â© 2023, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
A Bag-of-Phonetic-Codes Modelfor Cyber-Bullying Detection in Twitter
(2018) Shekhar, A.; Venkatesan, M.
Social networking sites such as Twitter, Facebook, MySpace, Instagram are emerging as a strong medium of communication these days. These have become a part and parcel of daily life. People can express their thoughts and activities among their social circle with brings them closer to their community. However this freedom of expression has its drawbacks. Sometimes people show their aggression on Social Media which in turn hurts the sentiments of the targeted victims. Certain forms of cyber-bullying are sexual, racial and physical disability based. Hence a proper surveillance is necessary to tackle such situations. Twitter as a micro-blogging site sees cyber abuse on a daily basis. However, tweets are raw texts; containing a lot of misspelled words and censored words. This paper proposes a novel method to detect cyber-bullying, a Bag-of-Phonetic-Codes model. Using pronunciation of words as features can rectify misspelled words and can identify censored words. Correctly identifying duplicate words can lead to smaller vocabulary of words, thereby reducing the feature space. The inspiration for this proposed work is drawn from the famous Bag-of-Words model for extracting textual features. Phonetic code generation has been done using the Soundex Algorithm. Besides the proposed model, experiments were carried out with both supervised and unsupervised machine learning approaches on multiple datasets to understand the approaches and challenges in the domain of cyber-bullying detection. � 2018 IEEE.
Big data computation model for landslide risk analysis using remote sensing data
(IGI Global, 2018) Venkatesan, M.; Prabhavathy, P.
Effective and efficient strategies to acquire, manage, and analyze data leads to better decision making and competitive advantage. The development of cloud computing and the big data era brings up challenges to traditional data mining algorithms. The processing capacity, architecture, and algorithms of traditional database systems are not coping with big data analysis. Big data are now rapidly growing in all science and engineering domains, including biological, biomedical sciences, and disaster management. The characteristics of complexity formulate an extreme challenge for discovering useful knowledge from the big data. Spatial data is complex big data. The aim of this chapter is to propose a multi-ranking decision tree big data approach to handle complex spatial landslide data. The proposed classifier performance is validated with massive real-time dataset. The results indicate that the classifier exhibits both time efficiency and scalability. © 2018, IGI Global. All rights reserved.
Capsule Network–based architectures for the segmentation of sub-retinal serous fluid in optical coherence tomography images of central serous chorioretinopathy
(Springer Science and Business Media Deutschland GmbH, 2021) Pawan, S.J.; Sankar, R.; Jain, A.; Jain, M.; Darshan, D.V.; Anoop, B.N.; Kothari, A.R.; Venkatesan, M.; Rajan, J.
Central serous chorioretinopathy (CSCR) is a chorioretinal disorder of the eye characterized by serous detachment of the neurosensory retina at the posterior pole of the eye. CSCR results from the accumulation of subretinal fluid (SRF) due to idiopathic defects at the level of the retinal pigment epithelial (RPE) that allows serous fluid from the choriocapillaris to diffuse into the subretinal space between RPE and neurosensory retinal layers. This condition is presently investigated by clinicians using invasive angiography or non-invasive optical coherence tomography (OCT) imaging. OCT images provide a representation of the fluid underlying the retina, and in the absence of automated segmentation tools, currently only a qualitative assessment of the same is used to follow the progression of the disease. Automated segmentation of the SRF can prove to be extremely useful for the assessment of progression and for the timely management of CSCR. In this paper, we adopt an existing architecture called SegCaps, which is based on the recently introduced Capsule Networks concept, for the segmentation of SRF from CSCR OCT images. Furthermore, we propose an enhancement to SegCaps, which we have termed as DRIP-Caps, that utilizes the concepts of Dilation, Residual Connections, Inception Blocks, and Capsule Pooling to address the defined problem. The proposed model outperforms the benchmark UNet architecture while reducing the number of trainable parameters by 54.21%. Moreover, it reduces the computation complexity of SegCaps by reducing the number of trainable parameters by 37.85%, with competitive performance. The experiments demonstrate the generalizability of the proposed model, as evidenced by its remarkable performance even with a limited number of training samples. [Figure not available: see fulltext.]. © 2021, International Federation for Medical and Biological Engineering.
Deep Learning-Based Prediction, Classification, Clustering Models for Time Series Analysis: A Systematic Review
(Springer Science and Business Media Deutschland GmbH, 2022) Naik, N.N.; Chandrasekaran, K.; Venkatesan, M.; Prabhavathy, P.
Analysis of time series is a prominent issue in the field of data analysis. With large amount of existing data in time series, multiple algorithms for analyzing time series data are being proposed. A variety of deep learning models are being designed to enhance the diversity of datasets related to time series across different fields. In comparison with the existing methods, only few have incorporated deep neural networks to perform this task. In most of the cases, deep neural networks are being applied for image data but it can also be used for sequential data such as text and audio. Here, we throw light on the recent advancements in hybrid deep learning models which consist of combination of various frameworks of deep neural networks with statistical models that have led to an improvement in time series analysis. Deep learning models are categorized into discriminative, and generative models provide an insight into the data based on the perception of conditional or joint probability. In this paper, we have surveyed newly devised algorithms and limitations of prediction, classification and clustering for time series analysis which describes how the temporal information can be merged into the analysis of the time series data. Â© 2022, The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd.
Detecting Fake News: A Comparative Evaluation of Machine Learning Techniques
(Institute of Electrical and Electronics Engineers Inc., 2024) Aishwarya, C.; Venkatesan, M.; Prabhavathy, P.; Shetty, A.S.
Fake news is a significant and well-acknowledged problem in contemporary society due to its rapid spread via social media and various online networking platforms, thereby making it difficult to determine the validity of information. In this study, we examine literature for this issue, prevalent datasets like LIAR, Politifact, and COVID-19, as well as classical machine learning and deep learning models such as SVM, BiLSTM, and CNN- BiGRU for fake news detection, and analyze their effectiveness and scope of application for fake news detection. Â© 2024 IEEE.
An efficient image retrieval system for remote sensing images using deep hashing network
(2020) Valaboju, S.; Venkatesan, M.
Due to the huge increase in volumes of remote sensing images, there is a requirement for retrieval systems which maintain the retrieval accuracy and efficiency which requires better learning of features and the binary hash codes which better discriminate the images of different classes of images. The existing retrieval systems for remote sensing images use CNNs for feature learning which fails to preserve the spatial properties of an image which in turn affect the quality of binary hash code and the retrieval performance. This Paper tries to address the above goals by using (1) Extracting Hierarchical features of convolutional neural network and using them to sequential learning to better learn the features preserving spatial and semantic properties. (2) Use lossless triplet loss with two more loss functions to generate the binary hash codes which better discriminate the images of different classes. The proposed architecture consists of three phases: (1) Fine-tuning a pre-trained model. (2) Extracting the hierarchical features of convolutional neural network. (3) Using those features to train the deep learning-based hashing network. Experiments are conducted on a publicly available dataset UCMD and show that when hierarchial convolutional features are considered there is a significant improvement in performance. � Springer Nature Singapore Pte Ltd 2020.
Forecasting Land-Use and Land-Cover Change Using Hybrid CNN-LSTM Model
(Institute of Electrical and Electronics Engineers Inc., 2024) Varma, B.; Naik, N.; Chandrasekaran, K.; Venkatesan, M.; Rajan, J.
Land-use and land-cover (LULC) information helps analyze future trends and is essential for environmental management and sustainable planning. Time-series satellite images are employed in this study to forecast changes in LULC. Deep-learning (DL) frameworks have been widely used for modeling dynamic LULC changes at the regional level. However, improving the accuracy of the existing prediction models is necessary. This letter proposes an integrated convolutional neural network (CNN) and long short-term memory network (LSTM) known as a hybrid CNN-LSTM model to address the fine-scale LULC prediction requirement. The efficiency of the proposed approach was examined using LULC data for the Dakshina Kannada District of Karnataka State, India. The proposed model achieved an overall accuracy of 95.11% and a kappa coefficient of 0.92, based on the ground-truth data for 2014. The model's predictions for 2035, based on data from 2005 to 2014, revealed the following trends: Urbanization exhibited a pattern of rapid expansion and increased growth. The integrated CNN-LSTM model extracted spatial and temporal features for effectively predicting LULC changes. Infrastructure development, population density, and enhanced economic activities were the major driving factors of changes in LULC for the study region. Robust LULC change forecasting will strengthen LULC evaluations, aid in understanding complex land-use systems, and empower decision-makers to formulate effective land management strategies in the coming years. © 2004-2012 IEEE.
Grapevine SDLC Model for Real-Time Fake News Classification
(Springer Science and Business Media Deutschland GmbH, 2025) Aishwarya, C.; Shekokar, T.P.; Naga Mukesh, K.; Venkatesan, M.; Prabhavathy, P.
In an era of rapid information distribution, the presence of fake news presents enormous difficulties to society, influencing public opinion and decision-making on a global scale. To address this issue, a reliable and efficient system capable of detecting and classifying fake news in real time must be developed. This project proposes the design and implementation of a specialized Software Development Life Cycle (SDLC) model, called the Grapevine SDLC, specifically designed for developing a real-time fake news classifier using Large Language Models (LLMs) and Apache Kafka. The Grapevine SDLC takes a methodical, iterative approach, starting with a thorough requirements analysis that identifies both system capabilities and limitations. During the design and development phase, the system architecture is crafted with a focus on scalability and real-time processing, integrating LLMs for highly accurate content analysis and categorization. Kafkaâ€™s distributed messaging platform ensures seamless and efficient data streaming, enabling the system to handle large volumes of data in real time. Further, the model includes continuous monitoring and feedback loops to improve detection accuracy and adapt to evolving fake news patterns. Â© The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.
Graph based Unsupervised Learning Methods for Edge and Node Anomaly Detection in Social Network
(2019) Venkatesan, M.; Prabhavathy, P.
In the last decade online social networks analysis has become an interesting area of research for researchers, to study and analyze the activities of users using which the user interaction pattern can be identified and capture any anomalies within an user community. Detecting such users can help in identifying malicious individuals such as automated bots, fake accounts, spammers, sexual predators, and fraudsters. An anomaly (outliers, deviant patterns, exceptions, abnormal data points, malicious user) is an important task in social network analysis. The major hurdle in social networks anomaly detection is to identify irregular patterns in data that behaves significantly different from regular patterns. The focus of this paper is to propose graph based unsupervised machine learning methods for edge anomaly and node anomaly detection in social network data. � 2019 IEEE.
Graph based Unsupervised Learning Methods for Edge and Node Anomaly Detection in Social Network
(Institute of Electrical and Electronics Engineers Inc., 2019) Venkatesan, M.; Prabhavathy, P.
In the last decade online social networks analysis has become an interesting area of research for researchers, to study and analyze the activities of users using which the user interaction pattern can be identified and capture any anomalies within an user community. Detecting such users can help in identifying malicious individuals such as automated bots, fake accounts, spammers, sexual predators, and fraudsters. An anomaly (outliers, deviant patterns, exceptions, abnormal data points, malicious user) is an important task in social network analysis. The major hurdle in social networks anomaly detection is to identify irregular patterns in data that behaves significantly different from regular patterns. The focus of this paper is to propose graph based unsupervised machine learning methods for edge anomaly and node anomaly detection in social network data. Â© 2019 IEEE.
Hybrid Approach for Intrusion Detection System
(2018) Singh, P.; Venkatesan, M.
In the recent research, Intrusion Detection sys- tem in Machine Learning has been giving good detection and high accuracy on novel attacks. The major purpose of this study is implementing a method that combines Random-Forest classification technique and K-Means clustering Algorithms. In misuse-detection, random-forest algorithm will build a patterns of intrusion over a training data. And in anomaly-detection, intrusions will be identified by the outlier-detection mechanism in the random-forest algorithm. This hybrid-detection system will combine the advantage of anomaly and mis-use detection and improves the performance of detection. This paper mainly focused on evaluating the performance of hybrid approaches namely Gaussian Mixture clustering with Random Forest Classifiers and K-Means clustering with Random Forest Classifiers in-order to detect intrusion. These algorithms were evaluated for the four categories of attacks based on accuracy, false-alarm-rate, and detection-rate. From our experiments conducted, K-Means clustering with Random Forest Classifiers outperformed over the Gaussian Mixture clustering with Random Forest Classifiers. � 2018 IEEE.
Hybrid Approach for Intrusion Detection System
(Institute of Electrical and Electronics Engineers Inc., 2018) Singh, P.; Venkatesan, M.
In the recent research, Intrusion Detection sys- tem in Machine Learning has been giving good detection and high accuracy on novel attacks. The major purpose of this study is implementing a method that combines Random-Forest classification technique and K-Means clustering Algorithms. In misuse-detection, random-forest algorithm will build a patterns of intrusion over a training data. And in anomaly-detection, intrusions will be identified by the outlier-detection mechanism in the random-forest algorithm. This hybrid-detection system will combine the advantage of anomaly and mis-use detection and improves the performance of detection. This paper mainly focused on evaluating the performance of hybrid approaches namely Gaussian Mixture clustering with Random Forest Classifiers and K-Means clustering with Random Forest Classifiers in-order to detect intrusion. These algorithms were evaluated for the four categories of attacks based on accuracy, false-alarm-rate, and detection-rate. From our experiments conducted, K-Means clustering with Random Forest Classifiers outperformed over the Gaussian Mixture clustering with Random Forest Classifiers. Â© 2018 IEEE.
Hybrid dimensionality reduction technique for hyperspectral images using random projection and manifold learning
(Springer Science and Business Media Deutschland GmbH info@springer-sbm.com, 2020) Mohan, A.; Venkatesan, M.
Hyperspectral images (HSI) are contiguous band images having hundreds of bands. However, most of the bands are redundant and irrelevant. Curse of dimensionality is a significant problem in hyperspectral image analysis. The band extraction technique is one of the dimensionality reduction (DR) method applicable in HSI. Linear dimensionality reduction techniques fail for hyperspectral images due to its nonlinearity nature. Nonlinear reduction techniques are computationally complex. Therefore this paper introduces a hybrid dimensionality reduction technique for band extraction in hyperspectral images. It is a combination of linear random projection (RP) and nonlinear technique. The random projection method reduces the dimensionality of hyperspectral images linearly using either Gaussian or Sparse distribution matrix. Sparse random projection (SRP) is computationally less complex. This reduced image is fed into a nonlinear technique and performs band extraction in minimal computational time and maximum classification accuracy. For experimental analysis of the proposed method, the hybrid technique is compared with Kernel PCA (KPCA) using different random matrix and found a promising improvement in results for their hybrid models in minimum computation time than classic nonlinear technique. Â© Springer Nature Switzerland AG 2020.
Hybrid intelligent bayesian model for analyzing spatial data
(2018) Velmurugan, J.; Venkatesan, M.
Spatial data mining refers to the extraction of Geo Spatial Knowledge, maintaining their spatial relationships, along with other interesting patterns not explicitly stored in spatial datasets. The overall objective of this research work is to apply GIS based data mining classification modeling techniques to assess the spatial landslide risk analysis in Nilgris district, Tamilnadu, India. Landslide is one of the most important hazards that affect different parts of India in the every year. Landslides cover broad range impact on the people of the affected area in terms of the devastation caused to material and human resources. Landslide is generated by various factors such as rainfall, soil, slope, land use and land covers, geology, etc. Each landslide factor has a different level of values. The ranking of values and assignment of weight to the landslide factor gives good classification of landslide risk level. Data science and soft computing play major role in landslide risk analysis. The rank and weight are assigned to the landslide factor and its different levels using classification data science techniques. In this paper, we proposed a new model with integration of rough set and Bayesian classification called Hybrid Intelligent Bayesian Model (HIBM) to analyze the possibilities of various landslide risk level. The proposed model is compared with real-time data, and performance is validated with other data science models. � 2018, Springer Nature Singapore Pte Ltd.