• Sandeep Bharti
• Sandeep Bharti
• Ayaan Rajput

Sign language is a rich, multidimensional form of communication that uses coordinated hand movements, facial expressions, spatial organization, and timing to convey meaning. Unlike spoken languages, it depends entirely on visual and physical cues. Indian Sign Language (ISL), widely used by the Deaf community in India, reflects substantial variation across regions and is shaped by diverse gestural expressions. While ISL is deeply embedded in the lives of Deaf individuals, communication with those unfamiliar with sign language often proves challenging, due to the stark contrast between gestural and verbal modes of interaction. Bridging this communication gap requires intelligent systems capable of interpreting ISL gestures in real-time. Despite the growth of computer vision technologies, most studies have focused on American Sign Language and isolated signs, leaving dynamic ISL sequences less examined. To address this, our research presents a ViViT-based framework, optimized with deeper transformer layers and fine-tuned attention mechanisms to better capture complex spatiotemporal features. The model was trained on VISL-PICT, a purpose-built dataset of 508 gesture videos. Our approach attained 96.69% overall accuracy and 99.55% top-5 accuracy, offering a robust solution for automated ISL recognition.

Recognition of Indian Sign Language Gestures, Analysis of Temporal and Spatial Hand Movements, Deep Learning Using Transformer Architectures, Automated Interpretation of Sign Language in Real-Time

"An Intelligent Machine Learning Framework for Automated Yawning and Fatigue Monitoring", International Journal of Emerging Technologies and Innovative Research (www.jetir.org), ISSN:2349-5162, Vol.13, Issue 1, page no.a7-a14, January-2026, Available :http://www.jetir.org/papers/JETIR2601002.pdf

"An Intelligent Machine Learning Framework for Automated Yawning and Fatigue Monitoring", International Journal of Emerging Technologies and Innovative Research (www.jetir.org | UGC and issn Approved), ISSN:2349-5162, Vol.13, Issue 1, page no. ppa7-a14, January-2026, Available at : http://www.jetir.org/papers/JETIR2601002.pdf