Optimal sub-band and strength selection for blind watermarking in wavelet domain

Abstract

With increasing use of digital media, need for digital rights management has arisen. Watermarking is used to hide copyright protection information in the host medium. Hiding information to ensure digital right protection must ensure high imperceptibility and an acceptable level of robustness. In the watermark embedding, appropriate watermark strength and place selection in the host image is the most critical aspect of the whole process. Both watermark strength and place selection are considered as optimisation problems and are optimised using genetic algorithm (GA) and particle swarm optimisation (PSO). The watermark is embedded in the wavelet domain. With the proposed method optimal wavelet family, band, watermark strength and wavelet depth level are selected to ensure higher robustness and imperceptibility. The watermark is embedded in the selected bands of the wavelet packets. The band and wavelet depth is selected using GA and watermark strength is optimised using PSO method. The proposed method shows promising results against attacks on a variety of filters, i.e. low pass, high pass and median filters. Robustness results on JPEG compression and gaussian noise are also improved compared with the current approaches in practice.