Research of 3D ear recognition based on local feature matching

To overcome the drawbacks of 3D local binary patterns: high dimensional vector and being sensitive to noise, we proposed a new method based on 3D center-symmetric local binary patterns (3D CS-LBP) for feature extraction, then the distance of feature vectors is calculated for 3D ear recognition. Compared to the traditional 3DLBP, the 3D CS-LBP method for feature extraction gives the reduced dimensional vectors, which are insensitive to noise. The experiment results show that the method achieves the comparable performance and the speed is improved significantly.