Head Pose Estimation Based on Face Symmetry Analysis

International audienceThis paper address the problem of head pose estimation in order to infer a non-intrusive feedback from users about gaze attention. The proposed approach exploit the bilateral symmetry of the face. We use the size and the orientation of the symmetrical area of the face to estimate roll and yaw poses by the mean of Decision Tree model. The approach does not need the location of interest points on face and is robust to partial occlusions. Tests were performed on different datasets (FacePix, CMU PIE, Boston University) which give variability in illumination and expressions. Results demonstrate that the change in the size of the regions that contain a bilateral symmetry provides accurate pose estimation