Project Title: Design and Implementation of a Low-Cost Vision-Controlled Smart Prosthetic Hand Using Real-Time Hand-Gesture Recognition
Description:
This paper presents the design, implementation, and evaluation of a low-cost smart prosthetic hand controlled in real time through hand gestures and computer vision. The work addresses the limited accessibility of advanced prosthetic solutions, which remain expensive and dependent on complex physical or biological sensors such as surface electromyography electrodes. Instead of wearable sensing, the proposed approach uses a single standard webcam to capture and interpret human hand movements. The software employs OpenCV and the MediaPipe Hands framework to detect twenty-one hand landmarks and to track the thumb and index fingertips; a scale-normalized Euclidean distance between these fingertips is computed and compared against an empirically tuned threshold to classify the hand as open or closed. The resulting command is transmitted from Python to an Arduino Uno microcontroller through serial communication, driving micro servo motors that reproduce the corresponding motion in a 3D-printed prosthetic hand. A structured evaluation protocol is defined based on classification accuracy, precision, recall, F1-score, end-to-end latency, and processing frame rate, measured across repeated trials and different lighting conditions. The prototype demonstrated reliable real-time performance, low latency, and simple software–hardware integration, offering an affordable, contactless, and reproducible alternative to conventional prosthetic-control methods. The system also provides a strong foundation for future work, including independent multi-finger actuation, wireless control, and machine-learning-based gesture recognition
Uploaded At: 2026-04-14 18:44:48