Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{187735,
        author = {Veeresha Kadlibala Mathada and Varun M and Tanay N M and Vignesh V and Dr. T N Anitha},
        title = {Real Time Hand Sign To Speech Translator},
        journal = {International Journal of Innovative Research in Technology},
        year = {2025},
        volume = {12},
        number = {6},
        pages = {6516-6523},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=187735},
        abstract = {Real-time communication between Deaf users and the hearing population remains limited due to the lack of efficient, deployable Sign Language Recognition (SLR) systems. This work presents a lightweight Real-Time Hand Sign to Speech Translator designed for signer-independent performance and low-latency operation. Using a Bias-Controlled Few-Shot Learning framework on a refined WLASL subset, the system extracts 1662-dimensional skeletal features through MediaPipe Holistic and applies a custom normalization strategy to reduce variations in signer anatomy and camera distance. A simplified LSTM model performs sequence classification, and recognized signs are converted to speech through a TTS module. Results demonstrate a compact, practical, and accessible solution suitable for real-world communication support.},
        keywords = {Few-Shot Learning (FSL), Key point Normalization, LSTM, Media Pipe Holistic, Real-Time Inference.},
        month = {November},
        }