Jasper John Sahayam

Purpose: Accurate brain tumor grade classification from magnetic resonance imaging is essential for diagnosis, treatment planning, and prognosis assessment. Although deep learning models have demonstrated strong performance, many lack clinical validation, robustness across datasets, and interpretability for decision support. This study proposes an explainable hybrid deep learning framework that integrates tumor segmentation, feature optimization, grade classification, and model attribution for high-grade and low-grade gliomas, with evaluation across multiple MRI datasets under controlled experimental conditions. Methods: The proposed framework, Towards Interpretable Multiclass Semantic Intelligence: Explainable Brain Tumor Segmentation and Automated Grade Stratifi cation, integrates brain tumor segmentation, optimized feature selection, and discriminative classifi cation to support automated tumor analysis and grade stratifi cation. Tumor regions are segmented using a Path Aggregation Network (PANet) integrated with Graph Convolutional Neural Networks (GCNNs) and ResNet-50 to preserve relevant structural characteristics. From the segmented tumor regions, shape, texture, and Histogram of Oriented Gradients (HOG) features are extracted to characterize morphological and visual patterns. A modified Reptile Search Algorithm (RSA) is then employed to select discriminative features, aiming to reduce feature redundancy and mitigate over fitting. The selected features are subsequently classified using DenseNet-201 within a Siamese architecture to enhance robustness to inter-patient variability. Model interpretability is incorporated through Gradient-weighted Class Activation Mapping (Grad-CAM), which provides visual explanations of the regions contributing to model predictions. The framework is evaluated using the BraTS 2019, BraTS 2020, and BraTS 2021 datasets, together with additional MRI datasets. Results: The proposed model achieved classification accuracies of 99.12%, 99.23%, and 98.86% on the benchmark datasets. The average precision, sensitivity, and specificity exceeded 98% across datasets, demonstrating strong discrimination between tumor grades and consistent generalization across imaging protocols. Visualization analysis highlighted tumor regions contributing to classification decisions. Conclusions: The proposed model achieved classification accuracies of 99.12%, 99.23%, and 98.86% on the benchmark datasets. The average precision, sensitivity, and specificity exceeded 98% across datasets, demonstrating strong discrimination between tumor grades and consistent generalization across imaging protocols. Visualization analysis highlighted tumor regions contributing to classification decisions. Conclusions: The proposed framework demonstrates promising performance and provides interpretable model-attribution information under the evaluated experimental conditions. It may serve as a potential computer-aided decision-support approach, but prospective multicenter validation, calibration, clinician-in-the-loop assessment, and workflow integration are required before routine clinical deployment.

In the modeling and designing of PhotoVoltaic (PV) systems, parameter characterization in PV cell/module models remains a crucial field of research. Diode‐based models, such as single‐diode model (SDM), double‐diode model (DDM), and the three‐diode model, are frequently employed, and SDM and DDM are the most significant models. As a result, the difference between the estimated and experimental current can be minimized by using an objective function to solve the parameter characterization of such models. Metaheuristic optimization algorithms have recently been employed to get around the difficulty of finding accurate and highly reliable outcomes quickly. As a result, this research modifies the fundamental SDM and DDM and considers an objective function based on the modified models. Additionally, an improved version of a novel metaheuristic algorithm called White Shark Optimizer (WSO) is proposed by modifying the force control parameters of the WSO, and a chaotic generator is infused to improve the exploitation ability of WSO. The modified algorithm is named IWSO and applied to extract the PV parameters. This paper uses the new objective function to compare the conventional and the modified PV models. The outcomes of the experiment demonstrated IWSO's dominance over competing algorithms. With an average Freidman's ranking test value of 1.171, the proposed IWSO is superior to all selected algorithms. The average accuracy of modified SDM and DDM is 12% better than the traditional PV models. According to the findings, IWSO's estimated parameter values are the best, with the smallest difference between estimated and experimental current.

This article briefs about a smart multifunctional single-phase inverter control for a domestic solar photo voltaic (PV)-based distributed generation that can work in both a grid-connected mode and an islanded mode by making the inverter mimic the operation of a synchronous generator. The control objectives were threefold: to provide the required active and reactive power for normal operating conditions and under varying operating conditions, to maintain the rated voltage and the rated frequency for the islanded mode, and to switch between the two modes of operation with the least amount of disturbance for the system while behaving as a virtual synchronous generator (VSG). The control structure is divided into three major loops: the outermost loop responsible for power control, the middle loop responsible for voltage control, and the innermost loop responsible for current control. The proposed control methodology incorporates the functionalities of the grid-connected and the islanded-mode control into a single complex structure and thus provides support to the grid under abnormal conditions while providing good-quality power to consumers under grid failure. The efficacy of the system is good. The operation under various modes were simulated in MATLAB Simulink, and the proportional integral (PI) controllers used for current controllers were tuned using particle swarm optimization (PSO). It can be concluded that the control structure becoming complex is benefitted by the added advantages of the smart PV system. The smart domestic PV system helps the prosumer to actively provide frequency support and voltage support, adding frequency support to the existing multifunctional PV systems.

Summary: The recent development in technologies is the driving force for development in sensors, especially in military applications. Due to the openness nature of the ad hoc sensor network, the system gets easily affected, which may lead to some packet drop, transmission delay, high network overhead, and more energy consumption. In this paper, the security during data transmission is provided to the network by using the proposed Secure Atom Search Routing (SASR) algorithm, which is adopted from the behavior of molecular dynamics. For global optimization problems, this algorithm provides an effective solution based on the constraint and interaction force of atoms. Moreover, the performance of SASR is improved by providing a proper balance between exploitation and exploration. Since the knowledge base processes all the data, the computational complexity is reduced and the lifetime of the network is increased. The simulation and performance are carried out for the proposed Knowledge and Intrusion Detection based Secure Atom Search Routing (KID‐SASR) protocol and is compared with the existing methods based on the metrics trust, delay, throughput, energy, packet delivery ratio, network lifetime, trust detection rate, and communication cost. The results obtained show improvement in the overall performance of the system.