A Novel Mechanism Based on Gabor Convolution Network for Emotion Recognition in an Un-constrained Environment

Abstract

In the real-time scenario, the performance of the facial expression recognition algorithms is diminished by the factors like quality of the image, occlusion and head pose variations which are considered as open challenges in the field of image processing. This paper presents a novel facial expression recognition algorithm MCNNR based on the modified Gabor convolution network with enhanced random forest to detect expressions in a unconstrained environments. The simulation results of the proposed algorithm outperforms various traditional expression recognition algorithms in the context of accuracy while simulated on the JAFFE, CK+ and LFW datasets.