An Improved Blind Watermarking Method Facing Dual Color Images Based on Hadamard Transform

This paper proposes a blind color image watermarking method based on Hadamard transform to deal with the infringement of color images in the complex network environment. First of all, the color watermark image is channeled and each pixel in the channel image is converted into an 8-bit binary code. And then, the color host image is also channeled and expanded into 4 × 4 non-overlapping pixel blocks based on the selected embedding position. Finally, Hadamard transform is put on the pixel block, and 1-bit watermark information is embedded by the coefficient pair of the first line of Hadamard coefficient matrix. When extracting watermark, the watermark can be extracted only by comparing the size relationship of the coefficient pairs without the help of the original image. In this paper, a voting system of the minority obeying the majority is designed and implemented, which actually enhances the robustness of the method. In addition, a high security encryption system is designed by combining symmetric key and public key, and the correlation between RGB color channels is considered. Simulation results show that the proposed method has better robustness, real-time performance and security compared with similar methods, and meets certain requirements in terms of invisibility and capacity.