A Secure and Robust Watermarking Method Based on Evolutionary Optimization for Embedding Parameters

The overall performance of the invisible watermarking (WMg) process is significantly influenced by the selection of spatial locations in the frequency domain and the strength of the embedding. The first part of the work thus presents a new blind watermarking technique based on selective modulation by employing automatic optimum embedding parameters that ensure high-quality watermarking based on an evolutionary algorithm adopting two embedding parameters. Based on the magnitude of the coefficients in the middle band, we propose two watermarking approaches: the averaging and the difference scheme that guarantee good imperceptibility (Ipy) validated using six hosts and two watermark images. Ownership rights for shared data are provided by watermarking which can also assist in transmitting secret information in secure communication realization. The blind digital watermarking scheme is tested for reliability and robustness using a wireless communication system based on 802.11a. The combination of watermarking and communication system is experimented with using Additive White Gaussian Noise (AWGN) to study the channel effects on the watermarked images (WIs) in terms of Peak Signal to Noise Ratio (PSNR), Structural Similarity Index Metric (SSIM), and Bit Error Rate (BER). Experiments showed that the proposed watermarking scheme offers better imperceptibility and robustness against AWGN.