Histogram of oriented gradient and multi layer feed forward neural network for facial expression identification

In this paper an automatic system of performance review has been proposed for facial expression recognition issue. The system is essentially based on histograms of oriented gradient features and multi layer feed forward neural network. The performance was evaluated on three different dataset by varying two parameters: face resolution and color space of images. Results demonstrate that good recognition rates can be obtained using small image resolution in both color spaces grayscale and RGB.