Indian Sign Language Recognition using YoloV5 on Custom Dataset

View at publisher Summarise with VIRAT
Authors: Aayush Gautam, Aarya Aishwaryam, Satwinder Kaur, Garima Joshi
Cited by: 0
References: 13

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

Hand gesture recognition systems serve as a vital role for lessening the communication gap between people who can hear and the deaf communities. The YOLO-V5 object detection model, coupled with a custom dataset of Indian alphabets, has been employed in this paper to recommend an improved method for the Indian hand gesture alphabet detection. By generating a wide range of sign language alphabet signs with enhanced variability and reliability, our approach tackles the limitations of earlier datasets. The significance of this unique dataset curation in computer vision techniques for user convenience is illustrated by this method, which yields considerable accuracy gains over baseline methods. Based on our experimental results, our YOLO-V5 achieves overall accuracy of 92%. This research enhances real-time sign language recognition systems and highlights the significant impact of superior datasets on accurate performance.

Fields of Study & Subjects

Hand Gesture Recognition Systems Advanced Neural Network Applications Human Pose and Action Recognition

Publication details

Journal / source
International Conference on Signal Processing and Electronics Design (ICSPED)
Pages
212–217
Publication date
23 January 2026
Language
EN