APPARATUS OF BRAIN-COMPUTER INTERFACE, AND CLASSIFICATION METHOD THEREOF

摘要

PURPOSE: A brain-computer connection device and a classifying method thereof are provided to more accurately classify a test signal which is difficult to determine than an existed linear discrimination analysis when the test signal which is difficult to determine is classified. CONSTITUTION: A dictionary unit(110) designs a dictionary for classifying a test signal based on a training signal of each class. An explanation unit(130) finds a value of the test signal based on the dictionary. A classifying unit(150) classifies the test signal based on distribution of a coefficient which is not zero shown in each class of the value. A common spatial pattern(CSP) filtering unit performs common spatial pattern filtering of the training signal. The common spatial pattern filtering unit outputs the signal which the common spatial pattern filtering is performed. [Reference numerals] (110) Dictionary unit A; (130) Explanation unit y=Ax; (150) Classifying unit; (AA) Training signal; (BB) Test signal y