A wavelet-based genetic watermarking scheme for stereoscopic images

In this paper, we propose a genetic watermarking method for stereoscopic images. A stereoscopic image is stored as two images and combined by software while displayed. Thus, a pirate can only clip half of the image to remove the embedded watermark. To prevent the watermarks from being removed, the proposed watermarking method embeds an identical watermark into two half images of a stereoscopic image, respectively. To resist collusion attacks, the proposed scheme identify similar portions of the two images containing in a stereoscopic image and the wavelet coefficients used to embed watermark data are selected with genetic algorithm. Then, identical modification will be applied on corresponding wavelet coefficients of the similar portions. As shown in experimental results, test images with watermarks embedded are still have high quality. Moreover, It has been indicated that the embedded watermarks are still detectable after common image processing operations. Thus, the proposed method can increase the capability to resist some image processing methods if an appropriate fitness function of GA is adopted.