Hand Gesture Recognition Using Deep Learning for Deaf and Dumb Community

Sign language plays a fundamental part in encouraging communication inside the hard of hearing community. This paper focuses on developing a precise hand motion recognition system using Convolutional Neural Networks (CNN) and MediaPipe hand tracking technology. Our approach utilizes MediaPipe’s advanced hand keypoint technology to extract accurate data about hand movements, orientations, and keypoints from images. These features are fed into a trained CNN model for classification. Our system excels in handling variations in hand movement and orientation, enabling real-time and accurate recognition of signals. Rigorously tested with curated datasets from the deaf-mute community, our work has the potential to revolutionize communication for the hard of hearing. It seamlessly integrates with smartphones and tablets, making it versatile in classrooms, workplaces, and public spaces. Through this research, we contribute to the field of hand signal recognition, enhancing communication and promoting social inclusion for the hard of hearing community worldwide. By bridging the communication gap between the hard of hearing and non-deaf individuals, our framework opens up new avenues for meaningful interaction and improves the lives of the hard of hearing globally.