Research and Implementation of Gesture Recognition System Based on sEMG

In order to improve the life of handicapped people, this paper presented a two-channel gesture recognition system based on sEMG. This paper focused on how to improve the recognition rate of three kinds of gestures of ok, paper and rock, then presented two research points again: (1) Study the recognition rate of ok, scissors, rock and paper. (2) Study the relationship between location and recognition rate of two-channel acquisition system based on extensor carpi ulnaris and flexor digitorum profundus. Among the system, the algorithm processing included software de-noising, active segment detection, feature extraction and BP neural network classification through matlab. Finally, recognition rate of three kinds of gestures reaches 97.333%, which of four kinds of gestures reaches 87% and which of combining the extensor carpi ulnaris and flexor digitorum profundus as two channels is higher than that of the two muscles as channel alone.