zfatmi13 / gesture-detection Goto Github PK

The objective of our project is to use electromyography (EMG) in order to classify different hand gestures. EMG is a technique for evaluating and recording the electrical activity produced skeletal muscles. We used double channel surface electrodes to record EMG, with 3 electrodes per channel, one for ground and 2 for the differential input. We have used two channels as we are acquiring EMG for two different muscles, the flexor carpi ulnaris and the extensor carpi ulnaris. We have achieved two goals in our project: non-real time as well as real-time gesture detection. For non-real time gesture detection, we used BIOPAC to do the EMG recording and then imported the files into MATLAB, where we designed an algorithm to detect the gestures based on the features of the signal. When doing real-time gesture detection, we record EMG using a circuit. The circuit preamplifies the input signal, filters the resultant signal to keep the frequencies that are in the range of EMG, and removes the 60Hz noise. The output signa