Application of active appearance model for dual-camera face recognition

Face recognition is a very important topic in the field of pattern recognition. Traditional two-dimensional face recognition technologies using images taken by a single camera are easily influenced by expressions and poses resulting in low recognition accuracy. In this paper, a new three-dimensional face recognition technique is proposed. We apply a dual camera module to extract two images of simulated human eyes. The active appearance model is applied to find facial feature points. The disparity between the images of the left eye and the right eye is calculated and used to reconstruct a 3D face model. Twenty-four geometric features are extracted from the 3D face models and a multi-class support vector machine is then applied to face recognition. The experimental results show that the proposed method can reduce the influence of facial expressions and the risk of photo fraud.