A Machine Learning-Based Indian Sign Language Recognition System with Integrated Audio-Visual Aid for Children

View at publisher Summarise with VIRAT
Authors: Varun Fatepuria, Nivedita Agarwal, A. Anitha, Rishitaa Bohra
Cited by: 0
References: 3

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

This paper presents a preliminary machine learning based Indian Sign Language (ISL) alphabet recognition system aimed at assisting young children and beginner learners with hearing impairments. Unlike many existing systems that primarily focus on American or British Sign Languages, this work addresses the relative lack of educational tools for ISL. A custom alphabet-level dataset (A-Z) was created and augmented, and hand landmarks were extracted using MediaPipe. A lightweight Random Forest classifier enables real-time recognition on standard laptop hardware. To support early-stage learning, the system integrates audio output and visual reinforcement for each recognized alphabet. Experimental results demonstrate high accuracy on the custom dataset. While the current system focuses on static alphabet gestures, it serves as a foundational platform, with future extensions planned for word-level and sentence-level ISL recognition.

Fields of Study & Subjects

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

Publication details

Journal / source
International Conference on Advances in Intelligent Computing and Applications (AICAPS)
Pages
1–5
Publication date
11 February 2026
Language
EN