엔트로피 기반 불균형 광전극 결정 규칙 및 웨이블릿 회귀 인공 신경망을 이용한 동잡음 제거 기법

Abstract

Disclosed is a method of removing motion artifacts from functional near-infrared spectroscopy (fNIRS) signals. The method includes: disposing N optodes at a plurality of locations on a scalp, and forming a plurality of channels between the N optodes; calculating a correlation coefficient difference index of a neighbor channel around each receiver optode, and detecting a receiver optode in which a motion artifact has occurred based on the correlation coefficient difference index; and removing motion artifacts based on the detected receiver optode and an arrangement structure of the N optodes.