Sinhala Sign Language Learning Aid Using LSTM-Based Spatio-Temporal Modeling

View at publisher Summarise with VIRAT
Authors: Amila Kehelwatta, Narmada Karunaratne, Tilani Gunawardena
Cited by: 0
References: 16

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 the design and implementation of an interactive Sinhala Sign Language (SSL) Learning Aid using deep learning and computer vision. Leveraging Long Short-Term Memory (LSTM) and Bidirectional LSTM (BLSTM) networks, the system performs word-level SSL gesture recognition, unlike prior character-level approaches that lack linguistic and practical relevance. The proposed model extracts spatio-temporal features from sequential gesture data to achieve accurate and stable real-time classification. Experimental results show that the LSTM model outperformed the BLSTM, reaching an accuracy of 94.64% and demonstrating superior temporal learning efficiency. This research addresses key limitations identified in earlier studies like restricted temporal modeling, reliance on isolated gestures, and absence of adaptive learning mechanisms by introducing a robust and accessible framework that enhances the practicality of SSL education and supports greater inclusion within Sri Lanka’s Deaf community.

Fields of Study & Subjects

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

Publication details

Journal / source
IEEE International Conference on Industrial and Information Systems (ICIIS)
Pages
551–556
Publication date
16 January 2026
Language
EN