Hand Gesture Recognition for Multi-Culture Sign Language Using Graph and General Deep Learning Network

Abstract

This project presents a strong and efficient system for real-time recognition of American Sign Language (ASL), aiming to improve accessibility and communication. The proposed system utilizes a custom dataset that includes ASL gestures, numerical signs from 0 to 9, and essential functional signs such as “Delete” and “Space.” By leveraging Media Pipe, hand landmarks are extr“cted t” prov“de a l”ghtweight yet effective representation of gestures, ensuring an efficient preprocessing pipeline. The extracted hand landmarks are then processed by an Artificial Neural Network (ANN), which is trained to classify gestures with high precision. The system is designed to function in real-time, seamlessly integrating with a web-based platform to enable live gesture detection and interpretation. Through meticulous data preprocessing, landmark extraction, and ANN-based training, the model achieves both scalability and high accuracy A dependable and efficient system developed for real-time American Sign Language (ASL) recognition. A key aspect of this project is its emphasis on multi-cultural sign language support, laying the foundation for future expansions beyond ASL. Integrating deep learning techniques strengthens the system's reliability and performance, ensuring reliable recognition across different environments. Additionally, the integration of Media Pipe ensures computational efficiency, making the system practical for deployment on various platforms, including web and mobile applications. Overall, this project offers a scalable, real-time, and accurate solution for gesture recognition, contributing Contributing to the progress of assistive technologies designed to improve communication for individuals who are deaf or hard of hearing continues to evolve. Future developments may involve integrating more sign languages, implementing advanced gesture recognition, and optimizing performance for low-power devices, thereby enhancing the system’s functionality and reach.