Design and control of intelligent bionic artificial hand based on image recognition

Area of Science:

• Biomedical Engineering
• Robotics
• Computer Vision

Background:

• Electromyography (EMG) control for intelligent bionic prosthetic hands suffers from low accuracy.
• Integrating computer vision presents a promising advancement for prosthetic hand control.

Purpose of the Study:

• To design an intelligent prosthetic hand utilizing image recognition for enhanced control.
• To improve prosthetic hand accuracy and the quality of life for individuals with disabilities.

Main Methods:

• A convolutional neural network (CNN) was employed for object recognition to trigger prosthetic hand actions.
• A novel mechanical design featuring a four-bar linkage and side swing mechanism was developed for comprehensive finger and thumb movement.

Main Results:

• The image recognition-controlled prosthetic hand successfully performed five distinct human actions: grasp, side pinch, three-finger pinch, two-finger pinch, and pinch between fingers.
• The system demonstrated excellent precision in image recognition for controlling prosthetic hand tasks.

Conclusions:

• Image recognition provides a highly precise control method for intelligent prosthetic hands.
• The developed intelligent prosthetic hand is capable of completing targeted tasks with excellent accuracy, validating the system's effectiveness.