Real-Time Sign Language Recognition Framework Utilizing 3D Pose Estimation and Recurrent Neural Networks
International Journal of Scientific Research in Computer Science, Engineering and Information Technology · 10.06.2026 · 12 (3) · 662-672
DOI: 10.32628/cseit26123364 ↗Kurz gefasst
Fachartikel in der Zeitschrift International Journal of Scientific Research in Computer Science, Engineering and Information Technology, online erschienen am 10.06.2026. Verfasst von I Pawan Kumar und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Sign language serves as the primary mode of communication for the hearing and speech-impaired community. However, a significant communication gap exists between sign language users and non-signers, necessitating the development of automated translation systems. This paper presents a robust Sign Language Recognition (SLR) system designed to interpret isolated gestures from video sequences. The proposed methodology leverages Google MediaPipe for efficient, lightweight 3D pose extraction, capturing spatial dependencies of hand and body keypoints. These temporal sequences are subsequently processed using Long Short-Term Memory (LSTM) networks to classify gestures. Experimental results utilizing a subset of the WLASL dataset demonstrate that the proposed skeleton-based approach achieves high recognition accuracy while maintaining low computational complexity, making it suitable for deployment in real-time accessibility applications.
Zitieren
APA
Kumar, I. P., Reddy, K. R., Kumar, G. S., Rajasekhar, N., Kumar, N. V., & Begum, F. (2026). Real-Time Sign Language Recognition Framework Utilizing 3D Pose Estimation and Recurrent Neural Networks. International Journal of Scientific Research in Computer Science, Engineering and Information Technology, 12(3), 662-672. https://doi.org/10.32628/cseit26123364
Vancouver
Kumar IP, Reddy KR, Kumar GS, Rajasekhar N, Kumar NV, Begum F. Real-Time Sign Language Recognition Framework Utilizing 3D Pose Estimation and Recurrent Neural Networks. International Journal of Scientific Research in Computer Science, Engineering and Information Technology. 2026;12(3):662-672. doi:10.32628/cseit26123364
BibTeX
@article{Kumar2026,
author = {Kumar, I Pawan and Reddy, K Rupesh and Kumar, G Sai and Rajasekhar, Nakka and Kumar, Nallabothula Vinod and Begum, Famida},
title = {Real-Time Sign Language Recognition Framework Utilizing 3D Pose Estimation and Recurrent Neural Networks},
journal = {International Journal of Scientific Research in Computer Science, Engineering and Information Technology},
year = {2026},
volume = {12},
number = {3},
pages = {662--672},
doi = {10.32628/cseit26123364}
}
EUVIMED