Multi-scale feature extraction algorithm of ear image

The algorithm based on local binary pattern algorithm, direct linear discriminant analysis and waterlet transform can extract the multi-scale feature of the human ear image. The irrelevant of the gray scale integer transform and the local texture feature extracting performance make the local binary pattern algorithm have great advantage in ear image feature extraction. The two dimension waterlet transform algorithm can extract the main feature of the ear image by twice transformation. The multi-scale feature extraction algorithm of ear image based on local binary pattern algorithm have distinct advance than the extraction algorithm using the original image in recognition rate.