Fig. 15From: Facial expression recognition using optimized active regionsThe results of the three CNNs and their fusion on different expressionsBack to article page