Lastly, we learn the size of the salient facial blocks to represent the facial description with the notion that changes in expressions vary in size and location. We conduct person-independent ...