A secure color image dual watermarking combining block feature modulation and voting mechanism for authentication and copyright protection

In this paper, a color dual image watermarking is proposed by embedding a robust watermark to protect copyright, as well as a zero watermark for tampering detection. Block Feature Modulation Scrambling (BFMS) is proposed for security, which satisfies additive homomorphic encryption and reduces the numbers of key transmission. Moreover, Voting Mechanism with Majority Rule (VMMR) is proposed for tampering detection, and multiple watermarks are used to improve accuracy. In addition, Fibonacci Arnold Scrambling (FAS) and two-layer visual cryptography are also presented. In terms of red channel of RGB color space, a zero-watermark is constructed for the purpose of tampering detection, taking advantage of LT code, SVD (Singular value decomposition), Haar DWT and two-layer visual cryptography. In blue channel, a robust watermark is embedded by BFMS, FAS, SVD and QR decomposition, and it uses a specialized FOA (Fruit Fly Optimization Algorithm) and evolutionary optimization methods to balance the trade-off between robustness and imperceptibility. In green channel, no watermark is embedded due to its sensitivity, but BFMS is adopted to improve security. The authenticity of a suspected image can be verified in the absence of the original watermark and host images. According to the experimental and comparative results, the proposed scheme provides superior outcomes with high security, robustness, imperceptibility, and capacity along with a good accuracy rate in locating tampering.