Image watermarking method in multiwavelet domain based on support vector machines

A novel image watermarking method in multiwavelet domain based on support vector machines (SVMs) is proposed in this paper. The special frequency band and property of image in multiwavelet domain are employed for the watermarking algorithm. After performed single-level multiwavelet decomposition on each image block of an image, a mean value modulation method, which modulates mean value relationship of multiwavelet coefficients in two approximation sub-bands, is used for carrying watermark embedding. The mean value modulation method can more effectively reduce image distortion than that of conventional single coefficient. At watermark detector, SVMs is used to learn the mean value relationship. Due to good learning ability of SVMs, watermark can be correctly extracted under several different attacks. The experimental results show proposed algorithm is robust to common attacks such as JPEG, low-pass filtering, noise addition, rotation and scaling, etc.