A Smart Glove Empowered by AI embedded on a Raspberry Pi

Abstract

This thesis presents the development of a smart glove system designed to translate ASL gestures into text and speech. The system uses a combination of flex sensors for finger movement detection, a module that combines accelerometer and a gyroscope for hand gesture capture, and a microcontroller for data processing. Data from all the sensors modules is acquired using an esp32, while the Raspberry Pi 4 is used for executing a machine learning model based on Random Forest algorithms. This model is trained with a dataset to recognize ASL gestures and convert them into text which is then synthesized into speech using a TTS module. That glove aims to help the deaf community to interact easily with the other categories of the society.Initial testings showed successful results,future work will focus on improving the device’s capabilities with other sign languages.