Kanta Prasad Sharma – Page 5

Improving Offline Gurmukhi Character Recognition: A Comparative Study of Feature Selection Techniques

Illustration of offline Gurmukhi character recognition using machine learning, featuring Gurmukhi script samples, a computer display, and feature selection icons.

This study investigates how different feature selection techniques enhance offline Gurmukhi character recognition. By analyzing various feature extraction methods and comparing their performance across machine learning models, the research identifies approaches that improve accuracy while reducing computational load. The findings support the development of more efficient and reliable OCR systems for Gurmukhi script, enabling better […]

Performance Analysis for Predicting Primary School Dropouts Using Machine Learning

Illustration showing machine learning models analyzing student data to predict primary school dropout risks using visual indicators like charts, children icons, and decision pathways.

This study evaluates the performance of various machine learning models in predicting primary school dropout risks by analyzing student-related data such as attendance, academic performance, and socio-economic factors. By comparing multiple algorithms, the research identifies the most effective approaches for early detection of at-risk students. The model’s predictive insights help educators implement targeted interventions, improve […]

Quantum Behaved Binary Gravitational Search Algorithm with Random Forest for Twitter Spammer Detection

Illustration showing AI algorithms detecting Twitter spammers, featuring gravitational search elements, decision-tree symbols, and digital user icons.

This hybrid model detects Twitter spammers using quantum-inspired optimization and random forest classification. The gravitational search algorithm improves feature selection and search efficiency. Random Forest enhances decision accuracy for identifying malicious accounts. Together, they deliver a robust system for social media spam detection. For more information, click on the link below. Results in Engineering (SCI)2025https://www.sciencedirect.com/science/article/pii/S2590123025000817

Quantum-Inspired Firefly Algorithm with Ant Miner Plus for Fake News Detection

Illustration showing AI algorithms detecting fake news, featuring digital firefly-inspired nodes, ant trail patterns, and a news verification symbol.

This study integrates firefly swarm intelligence with ant-based rule learning for detecting fake news. The hybrid algorithm improves classification accuracy and reduces misinformation spread. Quantum-inspired optimization enhances search efficiency within large datasets. Designed to support reliable and automated news verification systems. For more information, click on the link below. International Journal of Modern Physics C […]

Enhancing Human Activity Recognition in Smart Surveillance Using Transfer Learning

Illustration of a smart surveillance system analyzing human movement patterns on a monitor, supported by transfer learning and AI-powered activity recognition.

This research uses transfer learning to improve activity recognition in surveillance systems. The model accurately identifies human movements from video streams in real time. It enhances monitoring efficiency for safety and smart city applications. The approach delivers higher accuracy with reduced training effort. For more information, click on the link below. National Academy Science Letters […]

Blockchain-Assisted Trust Framework for Increased Survivability in Internet of Vehicles

Illustration of connected vehicles secured through blockchain nodes, showing a network of cars, digital blocks, and encrypted links representing trust and communication in IoV systems.

A secure blockchain framework designed to enhance trust in Internet-connected vehicles. It ensures data integrity by validating communication through decentralized nodes. The system improves survivability against cyber threats in vehicular networks. Built to support safer, more reliable intelligent transportation environments. For more information, click on the link below. Internet Technology Letters (SCIE)2025 https://onlinelibrary.wiley.com/doi/abs/10.1002/itl2.70029

Evaluating MobileNetV2 Architecture for Resource-Efficient Facial Emotion Recognition

Illustration of a facial emotion recognition system using a lightweight neural network architecture, showing a face input, neural network diagram, and emotion output icons.

This journal explores a lightweight deep-learning approach for recognizing human emotions from facial images.Using the MobileNetV2 architecture, it delivers high accuracy while maintaining low computational requirements.The model processes facial features efficiently, making it suitable for real-time and mobile applications.Its resource-optimized design ensures fast performance without compromising recognition quality. For more information, click on the link […]

Transformative Impact of Explainable Artificial Intelligence: Bridging Complexity and Trust

A glowing digital brain icon in the center surrounded by symbols representing analysis, trust, performance, and human reasoning, illustrating the concept of Explainable Artificial Intelligence.

A comprehensive study exploring how Explainable AI (XAI) enhances transparency, interpretability, and trust in complex AI systems. The research highlights how XAI helps users understand decision-making processes, reduces ambiguity, and strengthens human–machine collaboration across real-world applications. For more information, click on the link below. Discover Artificial Intelligence (Scopus)2025https://link.springer.com/article/10.1007/s44163-025-00281-1

Author: Kanta Prasad Sharma