A Novel Riesz Transforms Based Coding Scheme for Finger-Knuckle-Print Recognition

Abstract

It is well recognized that in biometric systems feature extraction and representation are key considerations. Among various feature extraction and representation schemes, coding-based methods are most attractive because they have the merits of high accuracy, robustness, compactness and high matching speed, and thus they have been adopted in many different kinds of biometric systems, such as iris, palmprint, and finger-knuckle-print (FKP) systems. However, how to devise a good coding scheme for FKP recognition is still an open issue. In this paper, based on the finding that Riesz transforms can well characterize the visual patterns, we propose to encode the local patches of an FKP image by using 2nd-order Riesz transforms. Specifically, we propose a 6-bits coding scheme, namely RieszCompCode, which consists of 6 bit-planes. In RieszCompCode, three of the bit-planes are obtained by binarizing the image's responses to the three 2nd-order Riesz transforms, and the other three ones are from the classical CompCode scheme. Experiments conducted on the benchmark PolyU FKP database corroborate that the proposed RieszComopCode can surpass all the other coding schemes evaluated for FKP verification in terms of verification accuracy.