Researchers at the University of Berkeley have developed a device that can be used to control
This is far from the first recognition systemgesture, designed for human-computer interaction, but the new system presented unique advantages. For example, a fast algorithm for calculations in neuroscience, which is updated as new information is received.
“In gesture recognition, your signals over timewill change and this may affect model performance. We managed to significantly improve the classification accuracy by updating the model on the device, ”the scientists noted.
Researchers have selected material for theirdevice due to its flexibility to match the movements of the forearm muscles. The matrix consists of 64 electrodes, each of which detects electrical signals from different points on the arm. This data goes to an electrical chip, which uses an algorithm to associate signals with specific hand gestures.
The team trained the algorithm by winding the bandageon the user's forearm and instructing him to perform dozens of gestures. In testing, the system accurately classified 21 hand signals, including fist, thumb up, and finger counting.
Read more:
Check out the most beautiful pictures of Hubble. What has the telescope seen in 30 years?
Ancient people saved themselves from the cold by hibernation
Astronomers have discovered an "intriguing signal" from the closest star system to the Sun