Optimized QR Code Watermarking for Robust Digital Content Protection: A Compression-Aware Framework with Multi-Metric Evaluation

The proliferation of digital content has intensified the need for robust solutions to safeguard intellectual property and ensure data integrity. While QR code-based watermarking offers advantages such as high data capacity and error correction, existing methods often lack resilience to compression and adaptive embedding strategies tailored to image texture. This study introduces a backward-optimized QR code watermarking framework that iteratively refines embedding parameters to balance robustness, imperceptibility, and computational efficiency. By integrating adaptive spatial/frequency domain embedding (LSB substitution and DCT mid-band modulation) and a compression-aware validation cascade, our method achieves 98.3% extraction accuracy under JPEG (QF=50) and Gaussian noise (\(\sigma^2 = 0.01\)) attacks. Comprehensive evaluations across RGB and grayscale images (Lenna, Baboon, Fruits) demonstrate that shorter QR payloads (e.g., "HI") preserve image quality (PSNR > 43 dB), while high-texture images like Baboon mask distortions more effectively than smooth-textured ones (MSE difference: 1.0–1.5). Compared to traditional LSB and DNN-based techniques, our framework reduces bit error rates by 62% and accelerates embedding by 80%. The results underscore the viability of texture-aware QR embedding for applications ranging from medical imaging to anti-counterfeiting, with future extensions proposed for blockchain-integrated traceability and generative AI watermarking.