Ear Identification Based On L0 Gradient Smooth

The basic principle of biometric identification technology is using a unique feature or characteristic possessed by the individual for identity authentication. Biometric identification technology is better the security and ease of use compared with the traditional identity authentication technology; it is gradually replacing the traditional identity authentication technology. Ear recognition technology is one of the biometric identification technologies becoming more and more extensive attention of scholars at home and abroad.Image preprocessing to a great extent, will affect the result of the later of the accuracy of image identification. This paper proposes a L0 gradient based image smoothing method for ear identification. First of all, using L0 gradient method dealing with the human ear image for smooth operation, and then to enhance the image, the process is the use of histogram equalization, finally using Canny operator to the human ear image edge extraction. L0 gradient smoothing method is proposed in this paper respectively with Gaussian filtering smoothing method and bilateral filtering smoothing method has carried on the contrast experiment. The enhanced image compared with using L0 gradient filtered image, makes the ear has better identification effect. Comparative experiments with an existing algorithm demonstrate that our method has better performance and is more suitable for ear identification.Ear identification under different lighting conditions is quite difficult because some of the ear structures may be indistinguishable. This paper puts forward a kind of based on image details to enhance the human ear recognition algorithm. Through to the image binary conversion, add white areas in binary image details of the image enhancement processing, then SIFT(scale invariant feature transform) features are extracted from the enhanced images. Ear recognition is done by calculating the Euclidean distance between different SIFT descriptors. Standard SIFT key point of ear image are extracted respectively, and histogram normalized after SIFT key point extraction method of extraction and the detail enhancement experiment. The experimental results demonstrate that the proposed detail enhancement algorithm can reserve most of the ear structures and our ear identification algorithm has better accuracies.