Palmprint Texture Analysis Using Derivative of Gaussian Filters

Abstract

This paper presents a novel approach of palmprint texture analysis based on the derivative of Gaussian filter. In this approach, the palmprint image is respectively preprocessed along horizontal and vertical direction using derivative of Gaussian (DoG) filters. And then the palmprint is encoded according to the sign of the value of each pixel of the filtered images. This code is called DoGCode of the palmprint. The size of DoGCode is 256 bytes. The similarity of two DoGCode is measured using their Hamming distance. This approach is tested on the PolyU Palmprint Database, which containing 7605 samples from 392 palms, and the EER is 0.19%, which is comparable with the existing palmprint recognition methods