: Millions of speech-impaired individuals rely on sign language for communication, but most people are unable to understand it, creating a significant communication barrier. This paper presents an AI- powered smart glove that translates sign language gestures into text and speech in real-time. The system integrates flex sensors and an MPU6050 IMU sensor to capture hand movements, which are processed using Random forest model deployed on an ESP32 microcontroller. The recognized gestures were transmitted to a smartphone via Bluetooth, where a Text-to-Speech (TTS) system converts the recognized text into speech. The proposed system improves accuracy using TinyML optimization and enhances battery efficiency with low-power ESP32 modes. Experimental results demonstrate an accuracy of over 90% with optimized classification models.
E.THANGADEEPIGA et al. (Thu,) studied this question.