Multi-Biometric Watermarking Scheme Based on Interactive Segmentation Process

Abstract

Iris-based security systems are highly recommended because of their security and ease of use. Different watermarking techniques have been presented to provide security while exchanging or storing iris images. The previous iris image watermarking techniques have successfully reached their aims, however, they suffer from some limitations such as the distortions that are presented in the iris region because of the watermark embedding process, the limited embedding capacity, and the lack of robustness against unintentional attacks. On the other hand, nowadays, the biometric-based security systems have directed their interest towards multi-biometric techniques in order to improve the performance of the individual’s recognition process. This paper presents a new multi-biometric watermarking (MBW) scheme in which the features of the fingerprint image and some personal information are embedded in the iris image. To avoid the abovementioned limitations, the proposed scheme presents an interactive segmentation algorithm (ISA) and a Slantlet transform based watermark embedding method. The proposed ISA prevents any distortion in the iris region which is a beneficial feature for the iris image recognition process. The experimental results proved the efficiency of the proposed ISA in comparison with the Hough transform based methods in terms of accuracy, embedding capacity, and execution time. The proposed MBW scheme performs better in comparison with the state-of-the-art methods in terms of the intactness of the iris region, the robustness against unintentional attacks, and the watermark embedding capacity.