17 tears of teaching Experirnce in engineering college
EDUCATION
M.Tech Information Technology
RESEARCH, TEACHING, or OTHER INTERESTS
Computer Science, Artificial Intelligence, Psychiatry and Mental health, Multidisciplinary
FUTURE PROJECTS
Analyzing the impact of mental health in Quality of life
Analyzing the impact of mental health in Quality of life
Applications Invited related works ,data set
14
Scopus Publications
Scopus Publications
Artificial Intelligence to Enhance the Security of Internet of Things: A Comprehensive Review Dheekshaa V M, Harsha V, Sindhu M, S. Pavithra 3rd International Conference on Intelligent Data Communication Technologies and Internet of Things Idciot 2025, 2025 The Internet of Things is the term we call when devices are connected to the network and work together to provide a better experience for users, facilitate better decision-making, and improve operations. As with any invention, this one also has its disadvantages. IoT devices have complex structures and low processing capacities. There are serious security risks in designing IoT like poor authentication, data leakage, and DDoS attacks. In this review paper, how AI can secure IoT is discussed. A few AI-driven strategies that help mask all these weaknesses are also discussed regarding intrusion detection systems (IDS), behavioural analytics, anomaly detection, and real-time threat monitoring. It presents a four-layered IoT architecture consisting of perception, network, support, and application layers. The use of the IoT systems will become more flexible and efficient in data security and collection, secure and encrypted communication, and predictive security analytics. AI will play the same role in improving the threat perception, real-time identification, and response, as this will make it indispensable to ensure the IoT environment security. This work emphasizes the urgent need for solid AI-powered, multilayered adaptive security frameworks to match the evolving demands of cybersecurity.
Enhancing Power Grid Resilience against Cyber Threats in the Smart Grid Era R. Padmavathy, Sanjeev Kumar Singh, M. Sindhu, L. Hussien Jasim, Archana Saxena, Sukhvinder Singh Dari E3s Web of Conferences, 2024 In the age of smart grids, fortifying power grids against cyber threats has become of utmost importance. This paper reviews endeavours in the realms of defining, measuring, and scrutinising the resilience of smart grids. An exhaustive overview of both qualitative frameworks and quantitative metrics for resilience study is provided, underscoring the ideal characteristics of a resilience metric. The complexities in formulating and crafting such metrics in practical settings are also broached. Another focal point is the hierarchical outage management scheme, crafted to enhance the robustness of smart distribution systems, particularly those encompassing multi-microgrids. This scheme introduces a two-tiered approach: the preliminary stage revolves around resource scheduling via a model predictive control-based algorithm, whilst the subsequent stage centres on coordinating power transfers amongst microgrids. Moreover, the paper probes into the vulnerabilities of smart grids owing to the incorporation of information technology. A thorough exploration of security prerequisites, accounts of severe cyber-attacks, and a strategic methodology to detect and counteract these threats are elucidated. The paper wraps up by spotlighting future research trajectories, especially in forging a comprehensive framework for resilience and addressing challenges tied to multi-modal cyber/physical attacks and big data concerns.
Integrating Multi-Resolution U-Net for Morphological Segmentation with Multi-Scale Processing in Shape Recognition Dharmesh Dhabliya, M. Vigenesh, B Reddy, M. Sindhu, Prateek Garg, Meena Y R 2024 IEEE 4th International Conference on ICT in Business Industry and Government Ictbig 2024, 2024 Biological research, medical imaging, and satellite image processing all rely heavily on morphological segmentation. When dealing with complicated geometries, different intensities, and noise, traditional approaches fail miserably. The U-Net architecture of convolutional neural networks (CNNs) has demonstrated promising results in capturing both local and global picture characteristics. Nevertheless, when dealing with intricate morphological structures, their performance can be enhanced. This research presents a new method for morphological segmentation using a multi-resolution U-Net, which improves upon the classic U-Net design by adding a multi-resolution framework. Because of this adjustment, the network can now interpret picture characteristics at many scales concurrently, picking up both detailed information and larger contextual cues. If the borders between regions are not clearly defined or if morphological structures show a lot of variability, this method will be very helpful. Multiple resolution branches flow into the network’s encoder and decoder paths in the proposed multi-resolution U-Net design. The network receives feature maps of varied resolutions from each branch, which runs on a separate image scale. Through the use of multi-scale feature fusion, the network is able to get a deeper understanding of the image’s object hierarchy, leading to more accurate segmentation of smaller, more complex structures while still accurately representing larger regions. In benchmark datasets, the model outperforms other state-of-art segmentation techniques such as the classic U-Net in pixel accuracy along with other evaluation criteria such as Dice coefficient and Intersection of Union (IoU). This is particularly true in the medical images applications.
Wavelet Transform Based Feature Extraction for Fault Diagnosis in Machinery A Vyshnavi, Manisha P. Mali, S. Hemelatha, Mohit Gupta, M. Sindhu, Abhishek Singla 2024 15th International Conference on Computing Communication and Networking Technologies Icccnt 2024, 2024 The wavelet remodel is a popular sign-processing technique used for feature extraction in fault prognosis in equipment. In this paper, we present an outline of the wavelet transform primarily based on function extraction techniques for fault prognosis. We discuss the standards and advantages of the wavelet remodel and its applicability to machinery fault diagnosis. We gift an in-depth evaluation of the diverse stages of feature extraction along with de-noising, time-frequency distribution, characteristic selection, sample popularity, and category. We additionally talk about the application of the facts-driven methods along with clustering techniques and neural networks for special equipment fault diagnosis troubles. Finally, we provide an outline of numerous case research that has used the wavelet rework based totally feature extraction for fault diagnosis in diverse styles of machinery. This evaluation highlights the usefulness of the wavelet remodel for extracting fault capabilities from equipment with the purpose of, in the long run, helping to enhance the accuracy of fault prognosis and to lessen protection prices.
Intelligent Time Management Recommendations Using Bayesian Optimization G. Chandramowleeswaran, M. Sindhu, Haider Alabdeli, Anshu Mishra, Priya Vinod, Harmeet Kaur 2024 IEEE International Conference on Communication Computing and Signal Processing Iicccs 2024, 2024 This paper focuses on the improvement of the intelligent time management system which employ Bayesian optimization for suggesting time management plans for each particular person. In this sense, through historical data of input-output patterns and users’ preferences, the system aims at increasing productivity and user satisfaction. In the study, Gaussian Processes are used as the surrogate model in the Bayesian optimization so that the required evaluations by the algorithm to realize optimal scheduling methodologies are kept to a minimum. Implementation is done as a web application where users submit their tasks and get the recommended schedule instantly. Indicators like, the degree of task accomplishment, time, and scheduling compliance, and probably the users’ satisfaction suggest that system helped enhance time management results. Lack of feedback from the users is removed through questionnaire that reveals the simplicity of the system and the quality of its recommended times, thereby supporting the idea of Bayesian optimization as a game changer in the management of time. This research significance points to the need for maintaining efficient and individualized approaches to time management strategies and agrees with others’ findings, which suggest that this is an area ample fiction research needs to acknowledge and pursue.
Detecting Fraudulent Marketing in Online Social Networks and Mitigating Cyber Threats with Advanced Security Approaches Mohan Garg, M. Sindhu, Paramjit Baxi’, N Raja Praveen K, Piyush Mathurkar, C. Karthikeyan 2024 IEEE 4th International Conference on ICT in Business Industry and Government Ictbig 2024, 2024 Using their fake marketing postings on social networking platforms, the unscrupulous people attempt to grab the attention of the customers. When it comes to Social Networking Services (SNS), it might be difficult to spot fraudulent postings that are posted by fictional individuals. It is for this reason that there is a need to identify fraudulent postings on social networking sites. Specifically, a technique known as SNS Fraudulent Detection (SFD) is being suggested in order to identify these harmful marketing postings. The DFA-T and WC based WC-NFS are both components of the SFD scheme that has been presented. A Penalty Score (PS) is computed by DFA-T based on the hazardous words that are included in the extracted URL. This score is determined by the properties of the URL. WC-NFS receives from PS the URL characteristics that were retrieved from DFA. After then, the WC accesses the URLs that are added to the WC-NFS in order to get the numeric value of the WC-Index (WCI). Both the current data set from social networking sites and appropriate machine learning methods are used in order to detect harmful URLs and posts. The malicious posts are identified using the data set. The suggested SFD distinguishes between benign and malicious URLs with a high mark of accurateness, as shown by the results of the experiments.
Enhanced Prediction of Early-Stage Autism Using Supervised Learning with Optimized Feature Engineering Amritpal Sidhu, Kiran K S, P.D. More, K.S. Bhuvaneshwari, Prateek Aggarwal, M. Sindhu 2024 IEEE 4th International Conference on ICT in Business Industry and Government Ictbig 2024, 2024 An autism spectrum disease (ASD) affects a person's brain development. Autism spectrum disorder is detected in children and young adults. For an early diagnosis of the disease, parental observation is crucial. That is why it is possible to provide children with effective therapy or treatment when they are young. Misdiagnosis is common due to the wide range of symptoms and levels of autism range syndrome. The datasets on children and toddlers with ASD that were identified at an early age were subjected to a limited number of Feature Transformation methods. The next step was to determine which classification methods performed best by applying them to various ASD datasets. Log, Power-Box Cox, and Yeo-Johnson Transformations for toddler datasets and Log transformations for children’s datasets were the FTs that followed the best prediction. The next step in using these altered datasets for early autism prediction is to use characteristic Selection Techniques in order to identify the most important ASD risk characteristic. In this comparison, the proposed method is the top machine learning technique with an incredible accuracy. Decision Tree, Random Forest (RF), K-Nearest Neighbours(KNN), Support Vector Machine(SVM), and Gradient Descent(GD) are some of the other models that perform a decent job, but they can't compare to the outstanding results that proposed method obtained. With an accuracy rate of 98.99%, proposed method stands out as the most reliable and efficient alternative among these methods.
An Experimental Evaluation of an Effective QR Code Based Duplicate Product Detection Using Blockchain Technology R. Geetha, Jayakumar D, Prabakaran P, S. Nivedha, M. Sindhu, R.C. Thivyarathi 2024 International Conference on Intelligent Systems for Cybersecurity Iscs 2024, 2024 Over the span of the last few years, innovations in blockchain technology have received a great deal of attention. Currency exchange, which has a variety of applications in addition to digital money, is the subject that receives the most attention. Because of this, it may have an effect on multiple industries. Large-scale financial transactions have become more transparent and straightforward as a result of the implementation of blockchain technology. Through the utilization of blockchain technology, we are able to recognize counterfeit merchandise. When it comes to making a purchase of any kind in this day and age, the question of authenticity frequently makes an appearance. In addition, studies have shown that these aspects are essential for the expansion of the economy. In light of this, it is of the utmost importance to educate consumers about the transparency of the items in order to decrease the number of counterfeit goods. With blockchain technology, we have taken a step closer to eliminating the ever-increasing presence of false and hazardous items from the global market. This is a cause for concern. Each and every person ought to be made aware of the fact that the utilization of technology will result in a reduction in the production of merchandise that is counterfeit. Every item must be given a one-of-a-kind digital code in order to guarantee that the production and packaging processes are carried out correctly. In this research, a novel method for detecting counterfeit goods is presented. This method is referred to as the Blockchain and QR based Product Analyzer (BQRPA). It undergoes cross-validation with the traditional Classical Fake Product Identifier (CFPI) in order to determine whether or not it is effective. Within the context of the software installation process, this programme performs a scan of the product code in order to ascertain whether or not the product in question is a counterfeit.
Experimental Evaluation of a Secured Phrase Searching Model Using Cipher Principles Over Cloud Computing Platform R. Geetha, M. Kanniga Parameshwari, Praveena. V, T. Venkatesan, M. Sindhu, S. Bhuvana 2024 International Conference on Intelligent Systems for Cybersecurity Iscs 2024, 2024 For many cloud-based IoT device learning applications, including sensible clinical data analytics, phrase search—which allows retrieval of files containing a given term—plays a crucial role. In order to prevent sensitive information from being exposed by third-party providers, data owners typically encrypt files (such as medical records) before sending them to the cloud. Having said that, this does make the search operation quite challenging and current searchable encryption systems for multiple keyword searches can't do word searches because they can't determine the location connection of multiple key phrases in a query word using encrypted data stored on the cloud server side. This study presents a model called MCLPS, which stands for “Modified Cyber Law for Phrase Searching.” Its purpose is to identify specific phrases in a cloud environment. To test how well the model works, it is cross-validated with the traditional BET. The safety of information in the cloud has long been an important concern for cloud service providers and their clients, despite the fact that there are several elements that could affect data security during searches. The reason behind this is the presence of dangers posed by both external data sources and internal partners, namely, their employees. Therefore, due to its massive structure, maintaining this security perfect is an ongoing headache and difficult effort in the cloud. In most cases, this is achieved by encrypting data at all times and taking extra precautions to ensure that data is not compromised during the audition or any other internal exercises. For safe phrase search in cloud IoT, this study suggested the MCLPS model and the BET encryption technique. The data integration, recovery, indexing, sliding, encryption, decryption, and transfer processes are all encompassed in the suggested works. This proposed study introduces a new approach to encrypted searching that allows encrypted phrase searches, and the MCLPS ad BET procedures was developed to secure large data sets before storing them in the cloud.
A Machine Learning-Based Forecasting Model with Deep Learning Capabilities for Unbalanced Time Series Data A. Amudha, Sahil Suri, M. Sindhu, Shubhi Goyal, Mansi Kukreja, Kishor R. Pathak 2024 Global Conference on Communications and Information Technologies Gccit 2024, 2024 During the last several years, deep learning capabilities have surpassed those of more conventional models on a wide variety of machine learning-related tasks. Neural networks with deep layers have indeed been effectively used to time series forecasting issues, which are crucial to the field of data mining. Their ability to instantly understand the temporal connections contained in time series has made them a viable option. Yet, deciding on a deep neural network architecture and then optimizing its parameters calls for a high level of skill. Thus, it is essential to conduct comprehensive studies of the predictive capacities of all current architectural frameworks. The completion of this work is hampered by two primary factors: (1) a complete examination of the most recent research on depth learning's application to time series forecasting, and (2) an empirical inquiry analysing the effectiveness of the most commonly used architectures. Seven distinct deep learning models are evaluated and compared for their efficiency and accuracy. We present scores and distributions based on our analysis of the suggested models trained with various architectures and hyperparameter settings. The 50,000 time series in the databases are used to solve 12 unique forecasting issues. Using this data to train approximately 38,000 models, we present the largest deep learning research to date for forecasting time series. While LSTMs were shown to be the most significantly predictive, both LSTMs & CNNs showed promise in this research. Comparable performance is attained by CNNs, but with less variance in results as well as a more based on the best available overall.
