Applying Computer Vision with Deep Learning for Real-Time Sign Language Recognition

This research presents the design and implementation of a real-time sign language recognition system using computer vision and deep learning. The study integrates MediaPipe and TensorFlow-based Convolutional Neural Networks (CNNs) to interpret British Sign Language (BSL) hand gestures into text and speech without the need for a human interpreter. A custom dataset of gestures (A–E and 1–5) was generated for model training and evaluation. The model achieved an accuracy of approximately 94%, demonstrating robust performance and low latency on commodity hardware. The project contributes to the growing field of inclusive Artificial Intelligence, Human–Computer Interaction, and Assistive Robotics, promoting accessibility for individuals with hearing or speech impairments. The research also explores model optimization, dataset curation, and ethical implications of AI in accessibility technologies. Developed as part of the MSc in Artificial Intelligence and Robotics (with Advanced Research) at the University of Hertfordshire, this work reflects applied innovation at COA IMMENSITY LTD in developing socially responsible AI and robotics solutions. The findings have direct implications for real-world deployment of vision-based recognition systems, enhancing digital inclusivity and communication efficiency.