Hand Tracking Glove

Real-time sign language translation remains a challenge, as traditional methods rely on interpreters or camera-based systems that raise privacy concerns, restrict mobility, and require users to remain within a fixed field of view. These limitations hinder natural, dynamic communication and often fail to capture nuanced gestures. This project proposes a wearable, sensor-based alternative for American Sign Language (ASL) recognition.

The system consists of a lightweight, ergonomic sleeve embedded with flex sensors and electromyography (EMG) sensors to capture detailed hand, wrist, and arm movements. These sensors collect data on finger bending, motion, and muscle activity, which is processed through an embedded microcontroller using filtering, normalization, and segmentation techniques. The processed data is then transmitted to a central device for classification.

A machine learning model—ranging from classical algorithms to deep learning architectures—will be trained on labeled gesture data to accurately recognize ASL signs. By leveraging multimodal sensor inputs, the system improves robustness and accuracy. This project demonstrates a portable, privacy-preserving approach to sign language recognition, with potential applications in assistive communication and gesture-based humanñcomputer interaction.