isolated-sign-language-recognition

Code for INCLUDE paper with pre-trained models

[ICCV 2025] "Cross-View Isolated Sign Language Recognition via View Synthesis and Feature Disentanglement".

Vision-Based Isolated Dynamic Uzbek Sign Language (UzSL) Recognition with Manual and Non-Manual Features for Real-Time Applications

SignBart is an efficient Isolated Sign Language Recognition model that decouples x and y coordinates using a lightweight encoder-decoder architecture. It achieves high accuracy with fewer than 750K parameters, outperforming traditional models on datasets like LSA-64, WLASL, and ASL-Citizen.

Affordable and accessible solution that converts Sign language into readable text in real-time, empowering Deaf and Hard-of-hearing individuals to fully engage in a hearing world

Isolated Turkish Sign Language (TID) word recognition from webcam landmarks - MediaPipe + BiLSTM

In this work, we shared the source code which are utilized in the study named ODE Transformers for Isolated Sign Language Recognition: A Study on Robustness and Scalability. This code implements the multimodal experiments which reported in the article.

In this work, we shared the source code which are utilized in the study named ODE Transformers for Isolated Sign Language Recognition: A Study on Robustness and Scalability. This code implements skeleton based experiments which reported in the article

An end-to-end system for isolated sign language recognition, featuring data preprocessing, deep learning-based model training, real-time prediction, and a web interface for live sign recognition to enhance accessibility and communication.