Geometrically invariant watermarking for copyright protection and biometric authentication using feature detectors

Abstract

Protection of biometric data is gaining interest and digital watermarking techniques are used to protect the biometric data from either accidental or intentional attacks. Among the various biometrics, iris is very famous in the authentication area, as they are unique to each person and are mainly used for the establishment of instant personal identity. This paper proposes a feature-based robust digital image watermarking algorithm is proposed to achieve the goal of image authentication and protection simultaneously. The first step is to identify the feature key areas for embedding. The Hessian-Affine feature detector and Gillie's key detector are used for this purpose. In the next step, the copyright watermark is embedded into these regions and the authentication watermark is embedded in the remaining regions by using a block-wise fragile watermarking technique. The experiments were conducted with and without attacks and was proved that the algorithms are able to resist the attacks. A comparison between the two feature selection algorithms was also performed based on the number of feature points detected with and without attacks. Both the algorithms prove that they are efficient in terms of watermark embedding and extraction and can resist various image processing and geometric attacks.