http://dspace.dtu.ac.in:8080/jspui/handle/123456789/100 2026-06-12T12:49:15Z http://dspace.dtu.ac.in:8080/jspui/handle/repository/22744 Title: DESIGN AND DEVELOPMEN T OF SECURE FEDERATED LEARNING FRAMEWORK FOR DATA-SENSITIVE APPLICATIONS Authors: NARULA, MANU; Vishwakarma, Dinesh Kumar (SUPERVISOR); Meena, Jasraj (CO-SUPERVISOR) Abstract: In an era where security breaches and cyber fraud have become increasingly common, con- cerns about data privacy have also risen exponentially. This presents a substantial hur- dle for training Artificial Intelligence (AI) applications that handle sensitive data, such as healthcare and finance. Federated Learning (FL) provides a solution for Data-Sensitive Applications (DSA), enabling collaboration between competing parties while guaranteeing complete control over data privacy. Over time, the attacks on FL have evolved, and the native technology alone is insufficient to maintain the expected privacy levels when han- dling such data. Hence, efforts are made to integrate other technologies with FL to enhance certain aspects, such as resource management, security, privacy, and efficiency, with varying results and pros and cons. This thesis examines the integration of security mechanisms with FL, focusing on networks with resource scarcity, which include AdHoc networks, Internet of Things, and Internet of Medical Things (IoMT), commonly found in the field of healthcare and other DSA, addressing critical issues such as data privacy, security, access control, and scalability. It proposes solutions to enhance data security and privacy in healthcare systems by examining various existing studies and frameworks. FL is adopted for its ability to facilitate the integration of AI in domains where conven- tional centralized training approaches are infeasible due to data distribution constraints, privacy concerns, or competitive restrictions. Furthermore, FL enables remote or decen- tralized systems to achieve performance comparable to state-of-the-art centralized methods while incurring minimal resource overhead. By eliminating the need for centralized data aggregation, FL significantly mitigates the risks associated with large-scale data breaches and unauthorized access. This thesis makes a substantial contribution in summarizing the existing state of FL, its various types, supporting technologies, and challenges, designing hybrid security frameworks for FL aimed at DSA networks, and performing extensive evaluation of the state-of-the-art FL techniques. Our contributions are as follows: • An extensive systematic literature review has been conducted to investigate the cur- x rent state of FL implementations in DSA fields, such as healthcare and finance. This analysis highlights the complex challenges of implementing FL in mainstream health- care applications, including concerns about transmission costs, data security, privacy, and data/system heterogeneity. Additionally, we provide a detailed taxonomy for the existing literature, focusing on DSA. Alongside identifying the gaps, the review also highlights the fundamental challenges that FL may pose in real-world scenarios. • As organizations and developers explore FL solutions for various applications, the multitude of FL tools and frameworks can feel overwhelming for a beginner. We studied various tools and frameworks available in both open and proprietary forms to provide a concise view of their advantages, limitations, and utility, thereby filling a gap in the proposed literature. • It contributes to improving the security in resource-constrained DSA networks such as IoMT by employing quantization. Unlike prior techniques, we incorporate a workload- aware client selection scheme to overcome the quantization loss, optimize bandwidth, and provide stability to the training network by minimizing the straggler nodes with negligible performance overhead. Thus, providing a reliable security solution to resource-constrained DSA. The experimental results show a 77% to 95% decrease in straggler nodes and an 8% to 0% decrease in accuracy compared to standard FL, depending on the dataset’s complexity. • To ensure security in small edge networks, IoMT networks, and AdHoc networks, a dynamic, lightweight cryptographic FL framework is proposed. Conventional cryp- tographic techniques are computationally expensive for small networks. Thus, we propose a block cipher with an ever-changing key that takes multiple client and net- work attributes into account and is unique to all participants. The overhead in- duced by the cipher and key generation is significantly lower than that of state-of- the-art solutions available. The experimental evaluation indicates that the proposed scheme’s relative performance is comparable to that of native FL. The results reveal an increase of approximately 4.5% to 8% in per-round computation time compared to standard FL. However, this marginal computational overhead is justified by the significantly enhanced security guarantees provided by the proposed approach, par- ticularly in resource-constrained network environments, where conventional security mechanisms with similar computational budgets offer comparatively lower protection • To ensure privacy in networks with sufficient resources, we propose a hybrid FL xi framework that utilizes differential privacy in conjunction with homomorphic encryp- tion. To counter the loss incurred by the differential privacy scheme, we employ the workload-aware client selection. The simulation results demonstrate minimal compu- tational overhead and comparable performance to other state-of-the-art techniques. The proposed Fed-HDVE imposes a minimal time overhead of 2.8 to 3.5 seconds, depending on the model size. The performance evaluation, analysis, and experimental results indicate that the proposed solutions offer a viable and effective security solution. Moreover, the comparative study demonstrates that the suggested approaches perform on par with existing solutions, with lower computational overhead and resource consumption. 2026-02-01T00:00:00Z http://dspace.dtu.ac.in:8080/jspui/handle/repository/22696 Title: FACE DETECTION AND TRACKING Authors: MOOL, AKSHAY; Panda, Jeebananda (SUPERVISOR); Sharma, Kapil (CO-SUPERVISOR) Abstract: With the use of various technological advancements and devices in today’s routine life, detection and tracking of human faces and facial features become very essential areas of focus. They become important so that more techniques could be developed for increasing their working efficiency. The field of Computer Vision uses these intermediary processes of face detection and tracking to track and analyze the input of visual information about humans, their faces and/or body movements, and correspondingly proceed to the desired application. The present thesis work has been taken up for the development of (i) an optimiz- able face detection and tracking model based on facial landmark localisation and feature tracking, for better and efficient processing of faces in high quality video streams, and (ii) a Non-Neighbourhood Background Elimination component us- ing the built model with mathematical and statistical modelling, for reducing the processing time and computations required for finding the target face in a frame when it has already been detected. Many algorithms have been developed to facilitate Face Detection and Track- ing applications. Viola and Jones were able to develop an algorithm in 2004, that achieves real-time performance with decent accuracy in detecting faces. It was one of the first algorithms to achieve such efficient performance, that is why it is still used as a standard algorithm to compare against other upcoming algorithms, and therefore has been specifically discussed in this thesis. vii There is a lot of visual information that is generated in various fields, rang- ing from daily routine to specialized applications. Processing all these types of information efficiently is a valid concern and need focused research. Most face detection algorithms have to deal with low quality data in videos, since they’re mainly focused on surveillance applications, whose information capturing devices capture less information per frame. Consequently, this thesis reviews some state- of-the-art face detection algorithms and compares their processing efficiency on low and high quality videos. The comparative analysis reveals that these recent and modern algorithms do not work as effectively on high quality videos as they do on lower quality videos. Therefore, there is an increasing need to focus re- search on analysis of high quality information in videos in an efficient manner, so as to keep up the pace of their analysis with the information that is generated. High quality videos (data generated by current applications like social media, Multimedia content, etc) mostly exist in offline mode, that could be used for post processing by the Computer Vision applications. To address this need, an effort has been made to focus on developing such an algorithm that gives faster results on high quality videos, at par with the algorithms working on live low quality video feeds. The proposed algorithm uses Convolutional-MTCNN as base algo- rithm, and speeds it up for high definition videos. The proposed model speeds up the face detection process really fast, up to 19+ FPS, while still maintaining above 90% accuracy. This paper also presents a novel solution to the problem of occlusion and detecting partial or fully hidden faces in the videos. This is achieved by using statistical and probabilistic approaches, given that the face has been identified in first few frames, to give the algorithm an estimate of where the face should be in the occluded region. Since the focus of our research is to efficiently process high quality data, some viii commercially used face detection algorithms in open literature have also been considered in our research. Models like FaceNet, HOG, YuNet, alongwith Viola- Jones algorithm and MTCNN, have been discussed and analysed in our thesis. The research done is compared against these models, in an effort to improve their performance in commercial settings. Further analysis lead to the conclusion that modern face detection algorithms fail to provide optimal results when they have to deal with larger amounts of data per frame while processing higher quality videos. This thesis discusses an- other proposed work that tackles discussed problem and offers a solution to deploy commercially used state-of-the-art face detection algorithms to process only the regions of interest in a frame, and discard the rest to decrease the data to be processed. The model maintains the accuracy of the base algorithm while de- creasing the processing time per frame, thereby increasing the overall efficiency. The selection of region of interest is dependent on the detection of facial window in the previous frame. Therefore, the choice of base algorithm plays an important role in determining the speed of the model. The model achieves increased pro- cessing speeds of about 69–76% more than the standalone usage of the detection algorithms for analyzed frame rates. 2024-10-01T00:00:00Z http://dspace.dtu.ac.in:8080/jspui/handle/repository/22673 Title: ANALYSIS OF EEG SIGNALS IN SPATIAL, SPECTRAL AND TEMPORAL DOMAINS FOR CLASSIFICATION Authors: HOLKER, RUCHI Abstract: EEG signals serve as a non-invasive, real-time biomarker of brain function, offering sensitive metrics for diagnosing and monitoring various neurological and psychiatric disorders. Their ability to capture subtle changes in electrical brain activity makes EEG an invaluable tool for detecting patterns and dysfunctions underlying conditions like Alcohol Use Disorder (AUD) and Attention-Deficit/Hyperactivity Disorder (ADHD). Unlike subjective behavioral assessments, EEG provides objective, quantifiable metrics that reflect the dynamic interplay of neural networks across temporal, spectral, and spatial domains. This thesis introduces a comprehensive set of twenty seven Quantitative EEG (QEEG) features to create a detailed and multifaceted representation of brain activity. These neuro- biomarkers are grouped into three main categories. Power features quantify the signal's strength and statistical properties, including total amplitude power, standard deviation, skewness, kurtosis, and both the mean and standard deviation of the signal envelope, reflecting the strength, variability, and asymmetry of neural oscillations. In a similar category, Range EEG features (rEEG) further probe peak-to-peak dynamics with statistics like mean, median, lower and upper percentile margins, width, coefficient of variation, asymmetry, and standard deviation, offering a richly detailed view of voltage fluctuations across windows. Spectral features analyze the frequency components and complexity by measuring spectral absolute power and relative power, Shannon entropy, spectral flatness, spectral difference, spectral edge frequency, permutation entropy and fractal dimension, revealing abnormalities in brain rhythms. The third category is Inter-Hemispherical Connectivity features that measure the interaction between the brain's left and right hemispheres using metrics like the Brain Symmetry Index (BSI), correlation, mean and maximum coherence, and the frequency of maximum coherence, which are crucial for understanding network-level dysfunction. Another significant contribution is the development of a robust, generalized end-to-end signal processing and feature selection pipeline that converts raw EEG recordings into QEEG biomarkers for accurate diagnosis of behavioral and neurological disorders. The process begins with artifact removal and referencing to prepare high-quality, clean EEG data. This is followed by sub-time segmentation using overlapping temporal windows to preserve the continuity of neural dynamics over time. The resulting EEG signal is subjected to a vii broad-band spectral filter bank, dividing the signal into ten non-overlapping frequency bands covering the full range from 0–100Hz, ensuring that both slower and faster oscillations (including high-gamma activity) are analyzed. Within each spectral band, common spatial pattern (CSP) filtering pinpoints the most class-informative spatial components, maximizing discriminability between healthy and patient subject groups and reducing the influence of irrelevant channels. From these spectrally and spatially filtered signals, the twenty seven QEEG features are extracted, and subsequently averaged over different temporal windows, producing a high-dimensional feature space that enables comprehensive modelling of neural function. To address redundancy and highlight only the most predictive features, advanced filter-wrapper feature selection is employed to identify the most discriminant features. An ensemble feature selection approach is used: initially, filter-based methods such as ANOVA, Chi-square, Gini Index, and Information Gain Ratio statistically rank the features, the obtained ranks are averaged, and a wrapper technique—typically Sequential Forward Selection (SFS)—iteratively builds the optimal feature subset by maximizing classifier performance with cross-validation. This process reduces the feature set into a compact, highly informative set, supporting models that generalizes accurately across independent patient cohorts and multiple brain disorders. Finally, a novel and generalized framework is designed to extract Functional Connectivity features by capturing linear monotonic inter-channel associations, enabling robust identification of functional interactions between distinct brain regions. Functional connectivity in the time domain is quantified using the Pearson correlation coefficient, serving as a quantitative EEG (QEEG) feature that captures inter-electrode synchrony and underlying neural patterns to enhance the accuracy of disorder classification. The result is a fully integrated and computationally efficient pipeline that combines broad-spectrum feature capture, performs network-level analysis, and rigorous feature reduction to reach state-of- the-art accuracy in classifying behavioral and neurological disorders successfully. 2026-02-01T00:00:00Z http://dspace.dtu.ac.in:8080/jspui/handle/repository/22527 Title: AUTOMATIC MEDICAL TRANSCRIPTS SUMMARIZATION USING MACHINE LEARNING TECHNIQUES Authors: BEDI, PARMINDER PAL SINGH Abstract: The healthcare sector and biomedical domain are essential for public health and medical advancement, providing services from clinical care to research. Healthcare facilities offer crucial services like check-ups and disease management, while the biomedical domain drives medical innovation through research and experimentation. With the increasing volume of biomedical literature, automatic text summarization is vital for efficiently extracting insights. These algorithms, equipped with domain-specific knowledge, simplify complex information, facilitating knowledge dissemination and collaboration. Additionally, in the rapidly evolving field of biomedical research, automatic summarization systems ensure timely access to up-to-date information by monitoring and summarizing the latest literature and databases. There are two main approaches of Automatic Text Summarization: Extractive and Abstractive. Extractive summarization involves selecting and extracting specific sentences or phrases directly from the source text, prioritizing their frequency or relevance to compose the summary. In contrast, Abstractive summarization interprets and paraphrases the content to create new sentences conveying the essential meaning in a concise form. In this research work, extractive text summarization techniques in biomedical domain are explored, focusing on issues such as redundancy, coherence, and the risk of overlooking crucial information. Extractive summarization techniques in the biomedical domain utilize various algorithms and approaches, including Frequency-based Methods, Graph-based Algorithms, and Machine Learning Approaches, to identify and extract key sentences or phrases from biomedical documents. Hybrid approaches combine multiple techniques to improve accuracy and coverage, effectively summarizing complex biomedical texts while addressing challenges such as redundancy and information loss. To address the identified research gaps, numerous novel approaches have been proposed for biomedical text summarization. Firstly, a novel approach using the Methathesaurus from UMLS to extract named entity concepts is proposed which applies the BERT method to generate concise summaries from Pubmed and Mtsamples. Further, an unsupervised approach focusing on semantic similarity and keyword-phrase extraction for both single- document and multi-document summarization is proposed. Furthermore, to further improve vi upon the results, a distinctive framework utilizing deep neural networks for contextually aware summarization of biomedical literature is proposed which employs a binary classifier and bidirectional long-short term memory recurrent neural network. To validate the proposed approaches, comparisons are made with baseline methods in biomedical text summarization, including a recent graph-based approach with the FP- Growth method. The results indicate that the last proposed approach outperforms state-of- the-art methods, achieving the highest ROUGE score of 0.96, surpassing the scores of the first and second approach (0.74, 0.76). The research concludes that the proposed methods demonstrate superior results in the medical domain compared to existing state-of-the-art techniques, highlighting the efficacy of the developed summarization approaches for biomedical literature. 2024-05-01T00:00:00Z