FRFT Domain Watermarking Algorithm Based on GA Adaptive Optimization

In the field of digital communication and copyright protection, digital watermark technology is extremely critical to ensure the security of secret communications and copyrighted information. This paper proposes an improved hybrid watermark optimization scheme. This scheme takes advantage of the multilevel wavelet transform (MDWT) to embed watermark information in the fractional Fourier transform domain by modifying the singular values of the image. At the same time, a genetic algorithm (GA) is designed to optimize system parameters. Ensure adaptive and efficient embedding of watermarks. To further enhance the security of the watermark, the Arnold transform (AT) is used to process the watermark, adding an extra layer of security to it. In addition, this solution conducts a comprehensive evaluation of watermarks of different sizes, and the experimental results confirm the efficiency of the system: the peak signal-to-noise ratio (PSNR) values are all above 48dB, and the structural similarity (SSIM) and normalized correlation coefficient (NC) are all close to 1. Compared with some existing algorithms, this scheme performs well in resisting common image processing attacks and has strong robustness while maintaining good invisibility and security.