Prototype based feature learning for face image set classification

Recognizing human face from image set has recently seen its prosperity because of its effectiveness in dealing with variations in illumination, expressions, or poses. In this paper, inspired by the prototype notion originating from cognition field, we obtain discriminative feature representation for face recognition by implementing prototype formation on image set. The contribution of this paper is twofold: first, we propose to use prototype image sets as a common reference to sufficiently represent any image set with the same type; in addition, we propose a novel framework to extract image set's features through hyperplane supervised by max-margin criterion between any image set and prototype image set. The final features are summarized through pooling technique along the prototype image sets. We experimentally prove the effectiveness of the method through extensive experiments on several databases, and show that it is superior to the state-of-the-art methods in terms of both time complexity and recognition accuracy.