Chinese scientists develop gesture recognition method using muscle signals

Researchers at Shanghai Jiao Tong University have developed a new and innovative method for recognizing hand gestures. The technology works by generating images from muscle signals.

The basis of the technology — HD-sEMG
The research relies on high-density surface electromyography (HD-sEMG). It:

• Accurately records muscle electrical activity;
  
• Converts the signals into impulse sequences (cw-CST);
  
• Transforms the data into images mapped to electrode positions on the hand.
  

Neural network analysis: gesture recognition
The resulting images were analyzed through a specially designed neural network. It:

• Extracts relevant visual features;
  
• Accurately recognized 10 different gestures;
  
• Outperformed existing methods in many cases.
  

Why is this method important?
Project lead Yan Yu explained:

Where can this technology be applied?
Experts highlight that this method can be used effectively in the following areas:

• Controlling prosthetics;
  
• Post-injury rehabilitation;
  
• Augmented and virtual reality interfaces.
  

Yan Yu added:
Read 'Zamin' on Telegram!