MediaPipe ‐ LSTM ‐Enhanced Framework for Real‐Time Dynamic Sign Language Recognition in Inclusive Communication Systems
Engineering Reports · 2025-07 · 7 (7)
DOI: 10.1002/eng2.70142 ↗Kurz gefasst
Fachartikel in der Zeitschrift Engineering Reports, online erschienen am 22.07.2025. Verfasst von N. Anithadevi und 3 weiteren Autor:innen. Die Arbeit wurde bisher 15-mal zitiert (Crossref) und nennt 33 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Sign language serves as a crucial mode of communication for individuals with hearing and speech impairments. However, the lack of accessible recognition technologies continues to hinder inclusivity. This study presents an advanced vision‐based dynamic sign language recognition system that addresses challenges related to gesture variability and contextual diversity. The proposed system leverages MediaPipe's landmark tracking alongside an enhanced stacked long short‐term memory (LSTM) network for real‐time identification of continuous gestures. Additionally, a transformer‐based paraphrase module translates recognized gestures into semantically precise search queries, enabling seamless interaction with virtual personal assistants. To enhance recognition accuracy, a dedicated dataset of Indian sign language (ISL) gestures, encompassing 30 dynamic sign categories, was developed. This dataset facilitates precise tracking of hand, facial, and posture attributes. The proposed approach achieves exceptional performance, with 97% accuracy, 98.17% precision, and an F1‐score of 94.88%, surpassing state‐of‐the‐art spatiotemporal feature learning techniques. The system's scalability enables diverse applications in education, healthcare, and customer service, fostering greater accessibility for the deaf and mute communities. By integrating robust real‐time gesture recognition with semantic understanding, this research establishes a new benchmark for dynamic sign language recognition, laying the groundwork for future advancements in attention mechanisms and optimized spatiotemporal learning.
Zitieren
APA
Anithadevi, N., Palanisamy, S., Rubini, S. S., & Shrestha, S. (2025). MediaPipe ‐ LSTM ‐Enhanced Framework for Real‐Time Dynamic Sign Language Recognition in Inclusive Communication Systems. Engineering Reports, 7(7). https://doi.org/10.1002/eng2.70142
Vancouver
Anithadevi N, Palanisamy S, Rubini SS, Shrestha S. MediaPipe ‐ LSTM ‐Enhanced Framework for Real‐Time Dynamic Sign Language Recognition in Inclusive Communication Systems. Engineering Reports. 2025;7(7). doi:10.1002/eng2.70142
BibTeX
@article{Anithadevi2025,
author = {Anithadevi, N. and Palanisamy, SatheeshKumar and Rubini, S. Saranya and Shrestha, Sachin},
title = {MediaPipe ‐ LSTM ‐Enhanced Framework for Real‐Time Dynamic Sign Language Recognition in Inclusive Communication Systems},
journal = {Engineering Reports},
year = {2025},
volume = {7},
number = {7},
doi = {10.1002/eng2.70142}
}
EUVIMED