The detection and classification of EMG signals with using artificial intelligence

Electromyography (EMG) signals has a low amplitude, and these signals are effected by the others physiological noise which is created by organs. EMG signals is measured with needle electrode or with surface electrodes, in their characteristic frequency spectrum, they are filtered, amplified and the result of EMG signal can be observed. During the measurement and filtering, with wide frequency spectrum it cause unwanted noise and this noise effect and change signal characteristic of EMG signal. However, the limitations to frequency spectrum cause trimming the propertied of the measured EMG signal, and it makes classification of EM signals as hard. Therefore it is necessary to use intelligent algorithm to utilize and classification EMG signals. There are a lot of methods to utilize EMG signals. One of them is segmentation algorithm. With the segmentation of EMG signals it can be generated an analysis areas, segment to segment transition area, and can be created node points on these segments. Thus, the basic particles can be created to solution the problem of characterization of EMG signals. In this study, four pre-determined finger movements are measured and then these signals are separated as segments. After all segmentation process these segments are compared each other, detected same properties and node points are created with using these properties. With the help of node points, as using artifical intelligence algorithm Hidden Markov Model is run, and the movements of finger is tried to detect as correct.