3D Ear Recognition Using Spherical Harmonic Analysis

In recent years, biometric technology has been widely applied to various fields, earrecognition is a biometric authentication technology based on the most active, the mostchallenging task, but also one of the most promising technologies of this century.3D earrecognition is expected to address the impact of the two-dimensional light ear recognition,gesture recognition and other factors on effect, become the focus of researchscholars.However, the current domestic and international research institutions on thepractical application of the system is small, not very full. Because of this reason, this articlewill do some research about3D ear recognition.In order to achieve the research requirements, this paper do the following work:(1) This paper introduces the research and development status of the significance of earrecognition system, as well as some of the mainstream of current research methods and keytechnologies it used, and thus come to the difficulty of the subject of research for the study ofthis article the ideas.(2)The traditional characteristics of the human ear recognition algorithm is tootime-consuming, the accuracy is not high, ear recognition algorithm is proposed based on3Dspherical harmonic functions, in order to increase the accuracy of the algorithm, using aconcave point to the helix Globe construct10concentric spheres structure.(3) According to the human ear characteristics of traditional recognition algorithm is tootime consuming, the problem of accuracy is not high, this paper proposes a3D ear recognitionbased on spherical harmonic analysis algorithm, in order to increase the accuracy of thealgorithm, adopted with helix concave points to construct its ten concentric structure of theball. The algorithm using spherical harmonic functions of rotation invariance,three-dimensional shape feature of the human ear by the airspace to frequency domaintransform, finally use to calculate the characteristic matrix of the3D ear recognition, so as toimprove the accuracy of the3D ear recognition. Experiments show that this algorithm caneffectively, the accuracy and real-time3D ear recognition, can get satisfactory effect.(4) In this paper, the second3D ear recognition algorithm is the use of harmonicmapping, mapping a3D model of human ear on the unit circle, combining with the gaussiancurvature construct harmonic mapping to the human ear shape images, by using harmonicmapping to the human ear shape image computing the similarity between the two ear model.Experiments show that the algorithm can effectively, accurate3D ear recognition.