Chaos-Based Renewable and Privacy Preserving Binary Palmprint Phase Templates for Cancellable Palmprint Recognition

This paper presents a chaos-based renewable and privacy preserving binary palmprint phase template for cancellable palmprint recognition. In the proposed approach, a novel 2D orthogonal Log-Gabor filter is firstly constructed by two 2D Log-Gabor filters with orthogonal orientations to extract binary palmprint biometric phase template. And then the binary palmprint phase templates are encrypted by chaotic stream ciphers encrypt run in integer domain to generate renewable and privacy preserving binary palmprint templates. During the matching stage, the Hamming distance between two encrypted palmprint templates is employed to measure the similarity of different palmprints. The experiments carried out on PolyU palmprint database confirm the effectiveness of the proposed approach.