An Efficient 3D Ear Recognition Method Using Point Set Registration Approach

The paper presents 3D ear recognition which is based on the coherent point drift (CPD) algorithm. It primarily does the tasks of rigid point set registration followed by EM (expectation maximization) optimization. Automatic detection of pit points and nose index from the 3D ear images are the major concerns of this algorithm. The coherent point drift (CPD) algorithm performs well with numerous rigid point sets present in the anatomical structure of the ear. With the help of CPD algorithm, we recover the transformations between two points with perfect alignment between two-point sets. The UND collection J2 ear dataset has been used to validate the performance of the proposed work. Some challenges like occlusion by earrings and hair, pose variations, illumination changes, and rotations are present in the dataset. The proposed work has solved these significant problems The experimental findings on this dataset show that the suggested approach is feasible and effective under different environmental conditions.