Communication is a very important part of human life. But few individuals face challenges incommunicating with others due to various reasons such as physical or psychological issues.These issues can lead to inability to hear, speak or both. This limits their access to socialinteraction. This led to sign language creation. Sign language relies on hand gestures, facialexpressions, and body movements. It remains inaccessible to most people, which widens thiscommunication gap. Advances in Artificial Intelligence, Computer Vision, and NaturalLanguage Processing have opened new ways to bridge this gap using automated gesturerecognition systems. Early systems relied on data gloves, depth sensors, and CNN-basedmodels to detect hand gestures. Their major shortcomings were expensive hardware and limitedgesture vocabulary. Many approaches achieved recognition only under controlled conditionsand lacked robustness to varying lighting and real-world environments. However, most priorsystems did not provide communication support such as speech conversion or multilingualtranslation. The proposed system addresses these shortcomings using YOLOv8 for handdetection, MediaPipe for 21-point landmark extraction, and a Vision Transformer for gestureclassification, achieving 91% accuracy with CPU inference latency below 5 seconds.Additional features include speech-to-text, text-to-speech, and translation across six Indianregional languages. A Streamlit front-end allows a hardware-free and accessible userexperience.
Building similarity graph...
Analyzing shared references across papers
Loading...
Raghava Mattegunta
Building similarity graph...
Analyzing shared references across papers
Loading...
Raghava Mattegunta (Mon,) studied this question.
www.synapsesocial.com/papers/69e866616e0dea528ddeac76 — DOI: https://doi.org/10.5281/zenodo.19660324
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: