Using EEG pattern analysis for implementation of game interface

In this paper, we present on EEG pattern recognition to adapt a serious game without using a game controller, which can be replaced by EEG(electroencephalography) signal research of BCI system. The feature extraction from EEG raw signal can be extracted the ERS and ERD. The features are compared frequency band by band-pass filtering after the above feature extraction steps. The average value of extracted feature signal, also it can divide two types of Support Vector Machine (SVM). Thus, the classes of divided support vectors can be input into the pattern recognition left and right direction. Also, the proposed SVM algorithm for classification will be compared with other algorithms an improved recognition rate. The recognition rate of SVM shows the increased correct rates. The highest of average recognition (success) rate is 82.45% for the discrimination of two support vectors. The experimental results of the prototype game system indicate that superiority of SVM by comparing recognition rates of the others, and it is able to apply for experimental serious games without any controllers.