Facial Identification and Macro Expression Recognition with a New Textural Featuring Approach

This work proposes a new feature, based on a textural approach, and targeting for facial identification or macro expression recognition. Approach is similar to classical LBP (Local Binary Pattern) considered in this domain as basis in state of the art by textural featuring approaches. Proposal is developed and, comparatively to LBP, evaluated with success on 2 different databases using an SVM classifier. Experiments are lead on both applications of facial identification and macro expression recognition, and our obtained results overpass LBP.