Research on 3D Face Recognition Based on Pose and Illumination Invariant
Yi Zhang, Yuan Zhou · 2017
In order to solve the problem of error rate caused by pose and illumination variation, the sparse statistical deformable model is used to reconstruct the 3D face shape. By using the method of spherical harmonic quotient image, estimation of the original input image light properties, the generated model standardized posture and illumination reconstruction, the original input image due to pose and illumination recognition error caused by the different solution. Experimental results show that the proposed method can effectively improve the recognition rate of 3D face.