Shah & Anchor Kutchhi Engineering College

This research aims to develop a comprehensive mental health prediction system based on Internet of Things (IoT) and deep learning methods. Specifically, this prediction system is especially sensitive to changing stress levels, to allow prompt intervention and personalized support, in response to the critical demand related to precise, convenient, and stigma-free mental health observation. The proposed system utilizes Interval type-2 Fuzzy Gradient recurrent mixed multimodal Convolutional Neural Network (IFG-CNN) that has three major modules. (i) Stress detector component performs a question–answer based evaluation with the help of Robustly Optimized BERT Approach (RoBERTa). (ii) and A facial emotion recognition module takes visual data and uses Scale-Invariant Feature Transform (SIFT) to extract features and then uses emotion classification using a Rotation-Invariant Surface Attention Radial basis function neural Network (RISAR-Net) trained by the white-faced capuchin optimizer. (iii) The classification module is a mental health status module that classifies them into highly stressed, moderately stressed and normal. The proposed model has a better prediction accuracy with accuracy of 97.85

Ballistic aggregates (BAs) have long been used in astrophysics to model interstellar dust, particularly for analyzing their light scattering and absorption characteristics. In this work, we extend the application of BA structures into the plasmonic domain by investigating their optical response at the nanoscale. Using the discrete dipole approximation (DDA), we numerically simulate the extinction and absorption behavior of 50 nm radius metal nanoparticles (Al, Ag, Au) arranged in BA, BAM1, and BAM2 configurations. Our results confirm that the optical properties of these clusters are strongly influenced by porosity, number of monomers, and the refractive index of the embedding medium. Notably, we identify a transition wavelength, below which extinction increases with porosity, while above it, the trend reverses. We also observe a redshift in surface plasmon resonance (SPR) with decreasing porosity and increasing cluster size, offering broad tunability of the extinction spectrum. The near-field analysis further reveals significant field enhancement due to interparticle coupling. These findings suggest that nanoscale BA-type clusters could serve as viable candidates for plasmon-enhanced light absorption in photovoltaic applications.

Maligancy of the brain is a disorder that affects the health of populations worldwide and for diagnosis, calls for complex skills and high-resolution techniques. This work proposes the use of a new strategy for the detection and segmentation of brain tumors with the help of MRI images. Concerning MRI, although the tumors are high in water content and possess hydrogen atoms, they give variations in signal intensity when compared to normal tissues. Thus, the utilization of CNNs for classifying and detecting brain tumors is quite efficient, provided the combined application of the CLAHE algorithm to boost the image contrast, where the outcomes of the classification attained a high accuracy. This chapter reflects on the Sustainable Development Goals in general by contributing to the development of more effective diagnostic tools and, consequently, enhancing the health of individuals. It is in line with the need to have sustainable healthcare systems for the constant and fair diagnosis and treatment of brain tumors regardless of the climate situation.

This work analyzes four novel NOMA-based optimization frameworks: (1) relay-assisted NOMA with DF and AF techniques, showing significant outage improvement over OMA (DF superior); (2) Sensing-NOMA (SNOMA) integrating spectrum sensing with user clustering and power allocation to maximize secondary throughput while protecting primary users; (3) Hybrid NOMA (HNOMA) game-theoretic scheme enabling autonomous coalition formation and dynamic switching between NOMA and OMA modes via preference relations and sequential games, further enhanced by an energy-efficient power allocation game; and (4) deep learning-based NOMA performing signal detection and fair power allocation without explicit channel state information, outperforming conventional SIC in detection accuracy and fairness. Extensive simulations validate all four schemes over existing OMA/NOMA baselines in outage probability, throughput, SE, EE, and fairness, leading to the conclusion that integrating NOMA with relaying, sensing, game theory, and provides a robust, scalable solution for future CRNs, while open challenges and research directions are also highlighted.

Skin cancer is one of the most common types of cancer in the world, and finding it early can greatly improve a patient’s chances of recovery. This paper presents a deep learning method for automatic classification of dermoscopic skin lesions into seven diagnostic categories using the HAM10000 dataset. Transfer learning is employed with Vision Transformer ViT-Base/16 pre-trained on ImageNet-21K and self-supervised DINOv2 Meta AI pre-trained on 142 million images (LVD-142M), with performance compared against EfficientNet-B4 and ResNet-50. The best performing model is ViT-Base/16, which achieves a classification accuracy of 96.95%, AUC-ROC of 0.987, and a macro-averaged F1-score of 0.952 in the held-out test set, with 94% recall in melanoma, the most clinically critical class. The results demonstrate that Vision Transformers consistently outperform traditional CNN architectures on this task. A fully deployed Flask-based web application is presented featuring real-time inference, model selection, probability visualisation, and Grad-CAM explainability.