RFID and IoT-Enabled Sign Language to Speech Conversion Using Deep Learning

Our research presents an RFID and IoT-enabled solution for converting sign language into spoken language using deep learning. This tool aims to address communication barriers faced by deaf and mute individuals due to the limited understanding of sign language in the general population. The system employs a hybrid model integrating IoT sensors, RFID technology, and Long Short-Term Memory (LSTM) networks with Transformer architectures. RFID-enabled IoT sensors capture real-time gestures, while Transformer encoders analyze complex contextual relationships between gestures. The LSTM layers handle the temporal aspects of sign movements, enhancing both gesture recognition and contextual understanding. The model, trained on a large dataset of sign language gestures, uses ReLU activation and Adadelta optimization. It achieved a training accuracy of 99%, testing accuracy of 93%, a Recall of 98.76%, a mean Average Precision (mAP) of 98.89%, and a Sensitivity of 97.84%. This innovative approach not only delivers high accuracy but also demonstrates significant potential for real-time sign language interpretation, enabling improved communication accessibility for the deafmute community through RFID and IoT technologies.