An extensive review of facial expression recognition using salient facial patches
Elsa Sabu, Paul P Mathai · 2015
Derivation of discriminant features from outstanding face patches has a major role in accurate identification of face expressions. The precise discovery of fragments in the face image improves the confinement of the outstanding patches on face images. Some methods uses a special framework for expression identification by the use of appearance related features of important face patches. A few noticeable face patches which are alive for each expression classes are extracted considering the location of fragments in the face image. Then further process these alive patches in order to obtain the patches which are discriminant and contain features for grouping of each pair of expressions. Thus it selects different facial patches as important for each pairs of expression classes. Then use these features for performing one against one categorization method. Also, an automated learning free method for fragment identification can be adopted, which achieves similar performances as that of other related fragment detection methods, and requires less execution time compared to others.