Robust Digital Image Watermarking based on spread spectrum and convolutional coding

Abstract

Digital watermarking is a promising technology to embed information as unperceivable signals in digital contents. A copyright protection method for digital image with convolutional coding is proposed in this paper. In this method, the watermark logo is coded with convolutional encoder and fused with noise bits to improve the robustness to malicious attacks including JPEG, noise and geometric attacks such as, cropping and rotation. During extraction, the watermark bits are decoded by Viterbi algotithm decoder, and the extraction procedure needs neither the original image nor the watermark logo. Simulation results show that the proposed method based on the convolutional coding can effectively resist the common malicious attacks, the embedding and extraction performances are quite improve compared to the method based on the direct coding.