A Digital Watermarking Method Based on Particle Swarm Optimization and Contourlet Transform

Abstract

As a key technology for digital content copyright protection, digital watermark often sees a dilemma between imperceptibility and robustness. In view of the problem, this paper suggests a CT-SVD-based intelligent watermarking scheme that embraces three major algorithms including the watermark preprocessing, the watermark embedding, and extracting. The host image is first decomposed by contourlet transform to get the low frequency subband images on which the SVD decomposition is then implemented before embedding the preprocessed watermark. The watermark extracting is an inversion of watermark embedding. To maximize the watermark robustness while maintaining its imperceptibility, particle swarm optimization (PSO) is applied to have heuristic searches for scaling factor. The experiment has demonstrated that this scheme is capable of resisting frequent image attacks with better imperceptibility and robustness.