A blind robust watermarking scheme with non-cascade iterative encrypted kinoform.

Digital watermarking remains a challenge in secure data embedding. Existing methods often struggle with balancing imperceptibility and robustness. Prior research has shown that traditional watermarking techniques face limitations in handling complex attacks. The Human Visual System (HVS) has been used to guide embedding in perceptual regions. However, embedding efficiency and security remain unresolved issues. Non-cascade iterative algorithms have not been widely applied in watermarking. Random fractional Fourier transforms offer novel encryption possibilities. This gap motivated the development of a new watermarking approach combining HVS and advanced encryption.

The study aimed to develop a blind watermarking scheme with enhanced security and robustness. The researchers sought to reduce data embedding requirements while maintaining imperceptibility. They focused on using non-cascade iterative algorithms for watermark encryption. The goal was to create a watermark that could be extracted without the original image. They also aimed to test the method's resistance to common image attacks. The motivation was to address limitations in current watermarking techniques. The study's problem was the lack of secure, efficient watermarking with strong robustness. The solution required integrating HVS with advanced encryption methods.

The researchers used a non-cascade iterative encrypted kinoform as the watermark format. They applied a non-cascade phase retrieve algorithm to transform the watermark. A random fractional Fourier transform (RFrFT) was used to encrypt the kinoform. The HVS was integrated to guide embedding into specific DWT coefficients. The watermark was adaptively embedded into 2-level DWT coefficients of the cover image. The iterative algorithm ensured efficient embedding while preserving image quality. The method avoided using a cascade structure to simplify the process. The watermark could be extracted using only the correct phase key and fractional order.

The proposed scheme achieved high imperceptibility with minimal visual distortion. The watermark remained detectable even after noise addition. Compression resistance was confirmed with PSNR values above 30 dB. Filtering attacks had little impact on watermark extraction. Cropping up to 25% of the image still allowed watermark recovery. The kinoform required less data than traditional computer-generated holograms. Extraction success rate was 98% with the correct phase key and fractional order. The method outperformed existing techniques in both security and robustness.

The authors demonstrated a blind watermarking scheme using non-cascade iterative encrypted kinoforms. The method achieved high security through random fractional Fourier transform encryption. Imperceptibility was maintained using HVS-guided embedding into DWT coefficients. The scheme showed robustness against noise, compression, filtering, and cropping attacks. The kinoform required less data than standard computer-generated holograms. Extraction required only the correct phase key and fractional order. The method's performance exceeded expectations in both security and robustness. The findings suggest this approach offers a practical solution for secure watermarking.