Photogram Classification-Based Emotion Recognition

This paper presents a method for facial emotion recognition based on parameterized photograms and machine learning techniques. Videos of people displaying emotions are by means feature-based emotional category association process to determine whether given photogram expresses comparing the action units displayed with findings in literature about emotion. To test proposed approach, two strategies adopted. First, gathered, then different classifiers applied check goodness obtained set categorized photograms. Second, trained sets were used emotionally classify all videos each database, using no preprocessing or selection. The presented was tested OpenFace parameterizer gathered from Multimedia Understanding Facial Expression (MUG) Cohn-Kanade (CK+) databases. outcomes achieved classification reached maximums 99.80% 99.63% MUG CK+ databases, respectively. classified voting regarding outcome video results reflecting rates 70.71% 66.36% up 72.55% 88.37% when classifier permutation used. work carried out opens door followup concerning data use combination methods recognition.