Facial expression feature selection based on rough set

Abstract

An improved reducing algorithm for rough set attributes has invented for answering the question of the excessive features vector dimensions. It obtains the local feature vector through geometric feature points. By introducing the rough set and improved reducing algorithm that it is able to select optimally among the existing expression features, also clipping the redundancy and useless information for the selection of expression feature. The experiment has showed that, this method has demonstrated high level of validity for its more convenience, higher recognition rate and more efficiency. © 2012 Springer-Verlag.