Coordinate direction normalization using point cloud projection density for 3D ear

Human ear has been long proved to be a promising candidate for personal identification. Human ear, with its 3D structure stability and detail expressive, has drawn researchers´ attention recently. The posture of ear always changes to the motion of head, it is necessary to normalize an ear to a standard posture to ensure the accuracy of feature extraction. This paper proposed a new coordinate direction normalization schema using 3D projection. It solves the problem of posture´s variation without initial estimation. The experiments show that the method is very efficient, stable, fast, and it can also eliminate uncertainty existing in traditional method.