Palmprint Biometric System Modeling by DBC and DLA Methods and Classifying by KNN and SVM Classifiers

Abstract

Biometric technology is an automatic personal identification method based on physical or behavioral characteristics of the individuals. Among of the physical characteristics, palmprint is useful in various applications such as forensic science access control, thus resulting in an increasing of research interest. In this paper, we explore a new methodology focused on integrating the fractal and Multi-fractal techniques for human identification based on extracting the texture pattern features. Therefore, we extract the palmprint texture information based on the calculation of the fractal dimensions using the Differential Box Counting (DBC) and the Diffusion Limited Aggregates (DLA) methods corresponding to the Fractal and Multi-Fractal techniques respectively. These methods have been broadly applied in image processing fields to estimate the fractal dimensions of an image as important parameters for analyzing the irregular shapes of the texture image. The proposed method produces encouraging recognition rates by 94.02 % and 93.44 % when tested on benchmark databases “CASIA-Palmprint” and “IITD-Palmprint” respectively. The performance of our method is compared with palmprint recognition accuracy gained from well-known state-of-the-art palmprint recognition, producing favorable results.