Abstract: A novel 3D ear recognition method is proposed in this paper. Firstly, using the PCA and
SVD algorithm to normalize 3D ear point clouds model, and adjust the position and posture of all ear
point cloud models in the database. Then, based on the Iannarelli system, we extract four local feature
regions of 3D ear model, and match them with Sparse ICP algorithm. Finally we match 3D ear
models according to the distance between their corresponding points. The experiments show that our
algorithm has higher recognition accuracy and efficiency compared with other algorithms.

Key words: ear recognition, PCA, Iannarelli, local feature, Sparse ICP