Automatic system for facial expression recognition based histogram of oriented gradient and normalized cross correlation

In this paper we built an automatic system for performance reviews of a new approach of facial expression recognition FER which is essentially based on Histogram of Oriented Gradient and Normalized Cross Correlation. The performance was evaluated by varying two parameters: face resolution and colour space of images. Results show that good recognition can be attained using low resolution, in particular 64 by 64 pixels in RGB images.