Sign Language Recognition Using Computer Vision and Machine Learning

View at publisher Summarise with VIRAT
Authors: Gokulnath Naidu, Ritikesh Alhate, Aryan Bhatjoshi, Akash Chaudhari, Abhishek Shiwalkar
Cited by: 0
References: 5

Summary

No summary yet. VIRAT can read this paper and write a plain-language summary of its aim, methods, key findings and limitations.

Summarise with VIRAT Free to try with your Science Catalogs account.

Abstract

Sign Language Recognition (SLR) continues to be an important tool for supporting communication between deaf or hard-of-hearing individuals and people who do not understand sign language. In this work, we develop a vision-based system that combines conventional image processing with deep learning to recognize both static and moving gestures. Instead of relying only on static hand images or wearable sensors, the proposed approach uses hand segmentation, landmark extraction, and a joint CNN-RNN model to capture both the shape of the hand and its motion over time. Testing was carried out on a custom dataset containing ASL alphabets and dynamic gestures. With the expanded dataset, the system achieved an overall accuracy of about 98%, showing a clear improvement over initial CNN-only trials, which struggled with similar hand shapes when trained on limited data. The model runs at 15.48 FPS, allowing smooth real-time gesture interpretation. These results suggest that integrating temporal modeling with carefully designed preprocessing steps offers a noticeable performance gain compared to typical static-framebased methods. The system shows strong potential for use in assistive communication tools, educational platforms, and realtime translation applications.

Fields of Study & Subjects

Hand Gesture Recognition Systems Human Pose and Action Recognition Hearing Impairment and Communication

Publication details

Journal / source
Biennial International Conference on Nascent Technologies in Engineering (ICNTE)
Pages
1–6
Publication date
16 January 2026
Language
EN