A PDF watermarking scheme utilizing invisible QR code embedding in the spatial domain

In today’s digital era of constant information sharing, ensuring the authenticity and integrity of digital documents—particularly portable document format (PDF) files—has become increasingly important. This paper introduces a watermarking scheme for PDF files by embedding quick response (QR) codes into the spatial domain of text color. The method takes a PDF and a QR code as input, resizing the QR code to match the length of the text on each page. The embedding process incorporates Arnold’s transformation and a permutation equal to the QR code’s size to secure both embedding and extraction. The QR code is applied by slightly altering the color of specific text characters that correspond to black pixels in the QR code image, achieving spatial-domain watermarking without degrading visual quality. Experimental evaluation demonstrates the method’s robustness against various attacks, with the watermarked PDFs achieving an average Peak Signal-to-Noise Ratio (PSNR) of 57.6 dB and Structural Similarity Index (SSIM) of 0.99. The extracted QR codes maintain PSNR values above 68.8 dB and SSIM equal to 1 after common processing attacks. The scheme also effectively detects malicious tampering, proving the potential of the proposed watermarking scheme in the protection of PDF files.