An eigen value based watermarking scheme for tamper detection in gray level images

Abstract

Fragile watermarking is used to determine if a digital content has been tampered, and localize tampered regions without referring to the original content. In this paper, we present a novel fragile watermarking technique for tamper detection in gray level images. The proposed method is efficient for simple and intelligent attacks such as collage attack and Vector quantization attack. The embedding process of the watermark starts from computing the Eigen values of the chaotically mixed image and that stands for the watermark, followed by inserting it into the LSBs chosen using the logistic map. The chaotic map is used to improve the security of our algorithm and to contribute with the Eigen values computation by making the watermark as unique as possible in order to face off intelligent attacks. Simulation results show that the presented method is fast, secure and performant in detecting and localising tamper.