Automatic facial expression recognition combining texture and shape features from prominent facial regions

Facial expression is one form of communication which being non-verbal in nature precedes verbal both origin and conception. Most the existing methods for Automatic Expression Recognition (AFER) are mainly focused on global feature extraction assuming that all facial regions contribute equal amount discriminative information to predict class. The detection localization have significant contribution recognition highly distribution from those not fully explored. key contributions proposed work developing novel upon combining power shape texture feature; determining identifying prominent hold abstract recognition. features taken into consideration Local Phase Quantization (LPQ), Binary Pattern (LBP), Histogram Oriented Gradients (HOG). Multiclass Support Vector Machine (MSVM) used while versus classification. implemented CK+, KDEF, JAFFE benchmark datasets. rate 94.2% CK+ 93.7% significantly more than handcrafted feature-based methods.