Head Pose Estimation based on Fuzzy Systems using Facial Geometric Features

Head pose estimation has many applications in the field of computer vision and it is a useful part of pose-invariant face recognition. In this paper, we propose a novel method to estimate head pose (yaw and pitch rotations) based on fuzzy systems by facial geometric features. Firstly, seven certain points are selected on face. These points includes some main properties. They are all visible even for large pose variations. Since no point is selected on mouth region, obviously this method is insensitive to facial expression. By these points, some ratios and angles are computed as the inputs of two fuzzy systems. The output data of these systems are the corresponding yaw and pitch angles. After training them and determining some parameters, they are used for head yaw and pitch estimation. This method is evaluated on two databases and the experimental results demonstrate that our proposed method is strongly accurate, robust, and beneficial for head pose estimation.