Classifying Facial Expression using Convolution Neural Network

Abstract

The human-computer interaction system is a success by deriving an effective facial expression recognition function. But it remains a difficult activity to understand facial speech. This paper sets out a novel Recognition of facial expression approach to the task. The approach proposed is motivated by the performance of the Convolutional Neural Networks (CNN) on the face trouble with identification. Unlike other plays, we focus on having good accuracy while requiring only a small sample data for training. The proposed approach is tested on Japanese Female Facial Expression (JAFFE). The accuracy increased compared with state-of-art results on the JAFEE dataset, where it achieved 95%.