Sign Language Predictor System Using Human Activity Recognition

Sign language detection using LSTM is a rapidly evolving research area that focuses on developing intelligent systems capable of interpreting sign language gestures. This study proposes a deep learning approach to detect sign language gestures by using Long Short- Term Memory (LSTM) networks. The proposed system first extracts features from sign language videos and then feeds them to an LSTM network for classification. The LSTM model learns the temporal dependencies in the data and accurately identifies sign language gestures. Experimental results indicate that the proposed method achieved an accuracy rate of 94 ^% on the American Sign Language (ASL) dataset and 90% on the Turkish Sign Language (TSL) dataset, outperforming other state-of-the-art methods. The proposed approach also shows robustness against varying lighting conditions and camera viewpoints. We will be employing Media pipe framework to recognize faces and landmarks which will then enable us to recognize and predict sign language symbols effectively. This study presents a promising approach for sign language detection using LSTM, which has the potential to significantly enhance communication between hearing-impaired individuals and the rest of the population. The proposed approach can be further improved by incorporating additional data augmentation techniques and other deep learning architectures.