Biometric Watermarking Based on Affine Parameters Estimation

Yang Li,Sidan Du
DOI: https://doi.org/10.1109/mmcs.2009.5256718
2009-01-01
Abstract:In this paper, the architect of biometric watermarking is researched. A feature embedding watermarking (FEW) system is resented firstly. A robust biometric watermarking method for iris image based on affine parameters estimation (APE) is presented. The determination of the general affine transform applied to an iris image relies on the determination of the regular grid of points for estimation of the Radon transformation. the entire approach lies on the regularity of the reference grid, as extracted form an input and possibly distorted image, and by comparing it with a known reference grid. Improving Voloshynovskiy's methods, we showed that this problem can be efficiently solved by Radon transform in the case of general affine transforms, avoiding the exhaustive search. Experiment shows that the watermarking method is robust to affine transform. An efficiency security way to protect biometric data is presented and researched. The proposed approach can also be used in any application where images need to be resynchronized, such as satellite imaging, or interactive digital maps, avoiding the need to embed visible marks.
What problem does this paper attempt to address?