A Posture–Depth–Motion Decomposition Framework for Hand Landmark–Based Sign Language Recognition

DOI

10.2991/978-94-6239-707-1_8How to use a DOI?

Keywords

ISl; Sign language; Mediapipe; gesture recognition; joint angle; flex bend

Abstract

For deaf and hard-of-hearing people, communication barriers remain a major obstacle, particularly in assistive and emergency communication situations. The majority of methods have mainly concentrated on gesture recognition and do not adequately address robustness under real-world variations like motion instability, posture inconsistency, and camera distance changes, despite the fact that recent vision-based Indian Sign Language recognition systems report high classification accuracy. The robustness-focused, vision-based ISL recognition system presented in this paper uses a commodity webcam to extract hand landmark-based features. The suggested framework combines motion stability assessment, finger posture consistency evaluation, and depth-aware feature extraction in an effort to increase signing reliability under different circumstances. To improve practical usability, recognized gestures are mapped to intent-level assistive phrases without attempting full sentence-level translation. Highlights of the paper: The following is a summary of this work’s primary contributions: Motion stability, posture consistency, and depth variation are all addressed by this robustness-focused, vision-based Indian Sign Language (ISL) recognition framework. MediaPipe hand landmarks can be used to extract interpretable biomechanical and kinematic features using a novel posture-depth-motion decomposition technique. In order to capture the intensity of expressive gestures, higher-order motion dynamics like velocity, acceleration, and jerk are integrated. An LSTM-based temporal modeling method to address changes in motion smoothness and signing speed. A mapping of intent-level assistive communication that improves practical usability without claiming complete linguistic translation.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - M. Neela Harish
AU  - G. Babu
PY  - 2026
DA  - 2026/06/18
TI  - A Posture–Depth–Motion Decomposition Framework for Hand Landmark–Based Sign Language Recognition
BT  - Proceedings of the International Conference on Recent Advances in Intelligent and Sustainable Technologies (RAIST 2026)
PB  - Atlantis Press
SP  - 87
EP  - 99
SN  - 2589-4919
UR  - https://doi.org/10.2991/978-94-6239-707-1_8
DO  - 10.2991/978-94-6239-707-1_8
ID  - NeelaHarish2026
ER  -