Automated Hand Gesture Recognition for Enhancing Sign Language Communication

This paper presents a novel approach to improving communication between individuals with hearing difficulties, with people of normal hearing capacity and unfamiliar with sign language. The work combines advanced technologies, such as artificial intelligence (AI) and computer vision to address the communication barriers faced by this group. Recent advancements in AI have made this possible. By using techniques like deep learning and computer vision, the work aims to develop a model that can turn complex sign language gestures into comprehensible/ interpretable language. Sequential deep learning models, specifically Long Short-Term Memory (LSTM) and Gated Recurrent Units (GRU), are used to recognize sign language gestures using Tensor Flow, with the assistance of MediaPipe Holistic. This application aims to narrow the communication divide between individuals proficient in sign language and those unfamiliar with it. This paper highlights the potential for a beneficial outcome through the fusion of AI and computer vision, especially given the widespread availability of video accessibility on smartphones and the growing influence of AI. The result shows that GRU performs better than LSTM in recognizing the sign language gestures with an average accuracy of 97%.