LATE G. N. SAPKAL COLLEGE OF ENGINEERING

Barriers and Limitations in Artificial Intelligence for Decoding Non-Human Signals

Advances in artificial intelligence, data analysis, and bioacoustics help bridge the gap between human and non-human species. By decoding animal vocalizations, gestures, and behavioral patterns, AI technologies offer new ways to understand animal emotions, intentions, and social interactions. This emerging field helps improve animal welfare, conservation efforts, and creates a broader understanding of cognition across species. Harnessing AI for communication challenges organizations to rethink the boundaries of language, intelligence, and empathy between humans and the natural world.

Exploring Deep Neural Architectures for Skin Lesion Detection and Classification

Skin cancer, such as melanoma, is an important global public health issue, in which the early detection and appropriate classification contributed to establish an early diagnosis that resulted in a more effective and efficient treatment. Recent advances in deep learning have sparked a new revolution in medical image analysis that has led to powerful automatic methods for skin lesion detection and classification. In this paper, we discuss the design and application of deep neural architectures, i.e., Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), Recurrent Neural Networks (RNNs), and hybrid models, for the purpose of dermatology diagnostics.

Inversion Based Face Swapping With Diffusion Model

"Inversion Based Face Swapping with Diffusion Model" presents an advanced face-swapping framework that leverages diffusion models and image inversion techniques to generate realistic and identity-preserving face replacements. The approach reconstructs images in the latent space, enabling seamless facial attribute transfer while maintaining the target image's pose, lighting, expressions, and background. The paper highlights the effectiveness of diffusion-based generative models in producing high-quality, photorealistic face swaps with minimal artifacts. It also discusses potential applications in digital content creation, entertainment, virtual reality, and privacy-preserving media editing, along with the ethical considerations of AI-generated facial manipulation.

"Crowdsense AI: Real Time Crowd Behavior Monitoring & Alert System"

"CrowdSense AI: Real-Time Crowd Behavior Monitoring & Alert System" presents an intelligent surveillance solution that leverages Artificial Intelligence (AI), Computer Vision, and Deep Learning to monitor crowd density and behavior in real time. The system analyzes live video feeds to detect overcrowding, unusual activities, and potential safety risks, generating instant alerts for authorities to enable timely intervention. It enhances public safety, improves crowd management, and supports smart city initiatives in locations such as railway stations, airports, stadiums, shopping malls, and public events. The paper also discusses the system architecture, key technologies, challenges, and future enhancements for AI-driven crowd monitoring systems.

"Implementation On Post Quantum Based Secure Email System"

"Implementation of a Post-Quantum Based Secure Email System " presents the design and implementation of an email security framework that employs post-quantum cryptographic algorithms to protect email communication against attacks from both classical and quantum computers. The system integrates quantum-resistant encryption and digital signature schemes to ensure data confidentiality, integrity, authentication, and non-repudiation. The paper discusses the architecture, implementation, performance, and security analysis of the proposed system, demonstrating its effectiveness in providing future-ready secure email communication for individuals and organizations in the quantum computing era.

Exploring Deep Neural Architectures for Skin Lesion Detection and Classification

Skin cancer, such as melanoma, is an important global public health issue, in which the early detection and appropriate classification contributed to establish an early diagnosis that resulted in a more effective and efficient treatment. Recent advances in deep learning have sparked a new revolution in medical image analysis that has led to powerful automatic methods for skin lesion detection and classification. In this paper, we discuss the design and application of deep neural architectures, i.e., Convolutional Neural Networks (CNNs), Vision Transformers (ViTs), Recurrent Neural Networks (RNNs), and hybrid models, for the purpose of dermatology diagnostics. We also provide a comparative study of benchmark datasets (ISIC, HAM10000, and PH2), efficacy of the preprocessing methods, and data augmentation and transfer learning strategies as a means of inflating performance metrics. The critical issues such as class imbalance, interpretability and cross-dataset generalization are also comprehensively addressed. This work also suggests some future research directions, including explainable AI and multimodal learning for dermatology. The objective of the review is to provide the researchers and clinicians in general, the overall perception of pros and cons of the available methods and the inspiration for creativity towards smart diagnostic systems of detecting and classifying skin cancer.

"Temporal Topic Modeling & Multi Topic Inference - A Bibliometric Analysis of Methods & Application "

"Temporal Topic Modeling & Multi-Topic Inference – A Bibliometric Analysis of Methods & Applications " presents a comprehensive bibliometric review of research on temporal topic modeling and multi-topic inference techniques. The paper analyzes publication trends, influential authors, research collaborations, and emerging themes while examining the evolution of methods used for dynamic topic discovery and analysis over time. It highlights key applications in text mining, social media analytics, healthcare, scientific literature, and business intelligence, providing insights into current research directions, challenges, and future opportunities in topic modeling and knowledge discovery.

Chapter - Solar Thermal Technologies and Nano-Enhanced Phase Change Materials for High-Efficiency Electric and Solar Mobility. 1st Edition - 2025

Chapter: Solar Thermal Technologies and Nano-Enhanced Phase Change Materials for High-Efficiency Electric and Solar Mobility presents an overview of advanced solar thermal technologies and nano-enhanced phase change materials (NEPCMs) for improving the thermal management and energy efficiency of electric and solar-powered vehicles. The chapter discusses the properties, heat storage capabilities, and applications of NEPCMs in enhancing battery performance, thermal regulation, and overall vehicle efficiency. It also explores recent advancements, challenges, and future research directions in sustainable mobility, highlighting the role of innovative thermal energy storage technologies in supporting clean and energy-efficient transportation systems.

"High Fidelity Blood Cell Detection in Microscopy: Comparative Evaluation of YOLOv9 & Faster R - CNN Architecture"

"High Fidelity Blood Cell Detection in Microscopy: Comparative Evaluation of YOLOv9 & Faster R-CNN Architecture" presents a comparative study of two advanced deep learning models, YOLOv9 and Faster R-CNN, for accurate blood cell detection in microscopic images. The paper evaluates both architectures based on detection accuracy, precision, recall, processing speed, and computational efficiency. It highlights their effectiveness in identifying and classifying different blood cell types to support automated medical diagnosis and laboratory analysis. The study provides insights into the strengths and limitations of each model, helping researchers and healthcare professionals select suitable object detection techniques for high-precision hematological image analysis.

" Implementation on Machine learning , blockchain & decision process for Securing Smart Grid"

"Implementation of Machine Learning, Blockchain, and Decision Process for Securing Smart Grid" presents a secure framework that integrates Machine Learning, Blockchain technology, and intelligent decision-making processes to enhance the security and reliability of smart grid systems. The proposed approach enables real-time threat detection, secure data sharing, and decentralized energy management while protecting the grid from cyberattacks and unauthorized access. The paper discusses the system architecture, implementation, security analysis, and performance evaluation, demonstrating how the combined technologies improve the resilience, efficiency, and trustworthiness of modern smart grid infrastructure.

"Machine Learning & Real Time Telemetry Predictive Autoscaling Framework of Cloud Infrastructure "

"Machine Learning & Real-Time Telemetry Predictive Autoscaling Framework of Cloud Infrastructure" presents an intelligent cloud resource management framework that combines Machine Learning and real-time telemetry data to enable predictive autoscaling of cloud infrastructure. The proposed system analyzes workload patterns, resource utilization, and performance metrics to forecast future demand and automatically scale computing resources. This approach improves application performance, reduces resource wastage, minimizes operational costs, and ensures high availability. The paper discusses the framework architecture, predictive models, implementation, and performance evaluation, demonstrating its effectiveness in optimizing cloud infrastructure management in dynamic computing environments.

"Secure Digital Medical Images Using Henon Map, Dynamic S-Box, and Elliptic Curve Cryptography (ECC)"

Secure Digital Medical Images Using Henon Map, Dynamic S-Box, and Elliptic Curve Cryptography (ECC) presents a robust encryption framework for protecting digital medical images during storage and transmission. The proposed approach integrates the Henon chaotic map for pixel permutation, a Dynamic S-Box for enhanced confusion, and Elliptic Curve Cryptography (ECC) for secure key generation and exchange. This multi-layered security mechanism ensures data confidentiality, integrity, and resistance against cryptographic attacks while maintaining image quality. The paper discusses the implementation, security analysis, and performance evaluation of the proposed method, demonstrating its effectiveness in safeguarding sensitive medical image data in modern healthcare systems.

"Secure Digital Medical Images Using Henon Map, Dynamic S-Box, and Elliptic Curve Cryptography (ECC)"

"Secure Digital Medical Images Using Henon Map, Dynamic S-Box, and Elliptic Curve Cryptography (ECC)" presents a robust encryption framework for protecting digital medical images during storage and transmission. The proposed approach integrates the Henon chaotic map for pixel permutation, a Dynamic S-Box for enhanced confusion, and Elliptic Curve Cryptography (ECC) for secure key generation and exchange. This multi-layered security mechanism ensures data confidentiality, integrity, and resistance against cryptographic attacks while maintaining image quality. The paper discusses the implementation, security analysis, and performance evaluation of the proposed method, demonstrating its effectiveness in safeguarding sensitive medical image data in modern healthcare systems.

"High Fidelity Blood Cell Detection in Microscopy: Comparative Evaluation of YOLOv9 & Faster R - CNN Architecture"

High Fidelity Blood Cell Detection in Microscopy: Comparative Evaluation of YOLOv9 & Faster R-CNN Architecture presents a comparative analysis of two advanced deep learning models, YOLOv9 and Faster R-CNN, for accurate blood cell detection in microscopic images. The study evaluates the models based on detection accuracy, precision, recall, inference speed, and computational efficiency to identify the most effective approach for automated hematological analysis. The paper discusses the implementation, performance comparison, and practical applications of both architectures, demonstrating their potential to enhance disease diagnosis, laboratory automation, and clinical decision-making through reliable blood cell detection.

"Implementation on Machine learning , blockchain & decision process for Securing Smart Grid"

Implementation of Machine Learning, Blockchain, and Decision Process for Securing Smart Grid presents a secure framework that integrates Machine Learning, Blockchain technology, and intelligent decision-making processes to enhance the security and reliability of smart grid systems. The proposed approach enables real-time threat detection, secure data sharing, and decentralized energy management while protecting the grid from cyberattacks and unauthorized access. The paper discusses the system architecture, implementation, security analysis, and performance evaluation, demonstrating how the combined technologies improve the resilience, efficiency, and trustworthiness of modern smart grid infrastructure.