New Paradigm for Self-embedding Image Watermarking with Poisson Equation

Abstract

In this paper, we propose a new self-embedding image watermarking scheme based on reference sharing and Poisson equation. With Laplacian operator, the relationship of each pixel and its neighborhood in original image is established and can be converted to compression bits. Then, after scrambling, compression bits are interleaved through the reference sharing mechanism, which can introduce more redundancy into the reference bits to be embedded for future content recovery. Thus, the relationship between each compression bit and each reference bit is constructed so that the recoverable area for tampered image can be increased effectively. Tampered contents can be recovered with the Laplacian values of tampered blocks and the boundary values around tampered blocks based on tampering localization and Poisson equation solver. Experimental results demonstrate the effectiveness of the proposed scheme.