Face feature extraction for emotion recognition using statistical parameters from subband selective multilevel stationary biorthogonal wavelet transform

Facial expression recognition is an extensive aspect in the field of pattern recognition and affective computing. Recognizing emotions by facial expression is an imperative action to design control-oriented and human computer interactive applications. Facial expression recognition is probable by the motion of facial muscles resulting in the appearance variation of face features. Accurate feature extraction is one of the extreme challenges that should be scrutinized for an admirable facial expression recognition system. One of the extensive key techniques used for feature extraction mechanism in facial expression recognition is wavelet transform. The features extracted from the wavelet transform incorporate both spatial and spectral domain information which is best adequate for identifying human emotions through facial expressions. In this paper, the statistical parameters from the proposed subband selective multilevel stationary biorthogonal wavelet transform are estimated and are used as features for effective recognition of emotion. The potency of the feature extraction algorithm is boosted by calculating the mean and maximum local energy wavelet subband of stationary biorthogonal wavelet transform. SVM classifier is used for classification of emotion using the preferred chosen features. Protracted experiments with well-known database for facial expression such as JAFEE database, CK + database, FEED database, SFEW database and RAF database demonstrates a better promising results in emotion classification.