Automatic Recognition and Generation of Indian Sign Language
Overview
Automatic Recognition and Generation of Indian Sign Language (ARGISL) is a joint collaborative project between the Department of Computer Science, RKMVERI (Belur), and the Faculty of Disability Management and Special Education (FDMSE), RKMVERI (Coimbatore). The main goal of this project is to develop an Indian Sign Language (ISL) recognition and generation system that can support and empower the Indian deaf community. We started our work with a collection of 2002 of the most commonly used daily words in ISL. The data was collected at the Coimbatore campus of RKMVERI, and with this we built a recognition model that currently achieves around 95% accuracy. Our long-term aim is to extend this work into more advanced systems that can bridge the communication gap between deaf and hearing individuals. If you like our work and wish to support us, you can do so here.
Indian Sign Language Word Dictionary
Here is the digital dictionary of ISL that covers 2002 commonly used words. Each entry includes a video demonstration of the sign. This dictionary will serve as a resource for ISL learners, educators, and researchers.
Publications
Our research findings have been shared through publications and conference papers so that others can learn from and build upon our work.
Recognition
📃STARK: Spatio-Temporal Attention for Representation of Keypoints for Continuous Sign Language Recognition
Suvajit Patra,
Soumitra Samanta
arXiv, 2026
📃Hierarchical Windowed Graph Attention Transformer Encoder and a Large Scale Dataset for Indian Sign Language Recognition
Suvajit Patra,
Arkadip Maitra,
Megha Tiwari,
K. Kumaran,
Swathy Prabhu,
Swami Punyeshwarananda,
Soumitra Samanta
Springer: Pattern Analysis and Applications, 2025
Generation
📃A Gloss-driven Indian Sign Language Production System using Learned Pose Representations
Suvajit Patra,
Arkadip Maitra,
Swami Punyeshwarananda,
Soumitra Samanta
International Joint Conference on Artificial Intelligence (IJCAI) (Accepted), 2026
📃A Continuous Indian Sign Language Production System from Isolated Signs
Suvajit Patra,
Arkadip Maitra,
Soumitra Samanta
IEEE: International Conference on Smart Computing & Communications (ICSCC), 2025
Acknowledgement
We would like to thank all the perticipants who have perticipated in our Indian sign language data collection process and the audio-visual unit of RKMVERI, Coimbatore for their technical support. We also like to thank VECC, Kolkata for parly supporting this Indian sign langauge project.
Contact
Suvajit Patra (he/him), PhD student, Computer Science Department, RKMVERI, Belur, suvajit.patra.cs20@gm.rkmvu.ac.in.
© Copyright 2024-26 by CS, RKMVERI, Belur.
Developed by Suvajit Patra, under the guidance of Soumitra Samanta.
All Rights Reserved.