A Robust Image Watermarking Based on Image Restoration Using SIFT

Abstract

This paper introduces a novel robust watermarking scheme for digital images, which is robust against common signal processing and geometric distortion attacks. In order to be resistant to geometric distortion attacks, the matched feature points determined by the scale-invariance feature transform (SIFT) are used for image restoration to reduce the synchronization errors caused by geometric distortion attacks. An adaptive embedding scheme is applied in discrete Fourier transform (DFT) domain of each subimage. The watermark detection decision is based on the number of matched bits between the retrieved and original watermark in the embedding image blocks. Experimental results show that the proposed watermarking is robust to common signal processing attacks and geometric distortion attacks, including rotation, scaling, cropping, shearing and some combined attacks.