Research Area:  Machine Learning
Automated Facial Expression Recognition has remained a challenging and interesting problem in computer vision. The recognition of facial expressions is difficult problem for machine learning techniques, since people can vary significantly in the way they show their expressions. Deep learning is a new area of research within machine learning method which can classify images of human faces into emotion categories using Deep Neural Networks (DNN). Convolutional neural networks (CNN) have been widely used to overcome the difficulties in facial expression classification. In this paper, we present a new architecture network based on CNN for facial expressions recognition. We fine tuned our architecture with Visual Geometry Group model (VGG) to improve results. To evaluate our architecture we tested it with many largely public databases (CK+, MUG, and RAFD). Obtained results show that the CNN approach is very effective in image expression recognition on many public databases which achieve an improvements in facial expression analysis.
Keywords:  
Author(s) Name:   Abir Fathallah; Lotfi Abdi; Ali Douik
Journal name:  
Conferrence name:  IEEE/ACS 14th International Conference on Computer Systems and Applications (AICCSA)
Publisher name:  IEEE
DOI:  10.1109/AICCSA.2017.124
Volume Information:  
Paper Link:   https://ieeexplore.ieee.org/abstract/document/8308363