Modified HMAX models for facial expression recognition
Wenfei Gu, Cheng Xiang, Hai Ting Lin · 2009
Three variants of HMAX model are proposed in this paper for recognizing facial expressions of novel faces. The expressions under consideration are: happy, sad, angry, surprised, disgusted, scared and neutral. The modifications are based on recent biological findings about face processing procedure in human brain. Computational results on two different facial expression databases show the efficiency of modified HMAX models compared to the original one.