Sign Language Detection to Enhance Online Communication for Deaf and Mute Individuals through Deep Learning Models

Communicating online presents significant hurdles for deaf and mute people, who frequently confront challenges due to speech and hearing limitations. These limitations can make their digital interactions difficult and irritating, as many online platforms and communication tools are not designed with their special needs in mind. This study intends to improve these people's online communication experiences by creating an object detection and classification model that detects Sign Language using computer vision. Three deep learning models, VGG16, ResNet50, and MobileNet, are evaluated using the SIBI Sign Language Dataset. The comparison focused on determining the most effective approach for accurate sign language detection. The results show that the VGG16 model outperforms the other models by 0.9680 accuracy and 0.9694 F1 score. By incorporating such technologies, a critical gap in digital accessibility can be closed, allowing these individuals to participate more completely in the digital world.