Facial expression recognition in video with multiple feature fusion

Abstract

Video based outward appearance affirmation has been a long standing issue and pulled in creating thought starting late. The way to a productive outward appearance affirmation system is to manhandle the conceivable outcomes of shifting media modalities and structure incredible highlights to effectively depict the facial appearance and course of action changes achieved by facial developments. We propose a powerful system to address this issue in this paper. In our examination, both visual modalities (face pictures) and sound modalities (talk) are used. Another part descriptor known as bar chart of familiarized Gradients from 3 Orthogonal Planes (HOG-TOP) is projected to free dynamic surfaces from video courses of action to depict facial look changes. In addition, another effective geometric element got from the curve change of facial places of interest is proposed to get facial plan changes.