A Novel Print-Scan-Resilient Binary Image Watermarking Method Based on Adaptive DCT Coefficients Modification

Currently, most existing binary image watermarking methods are vulnerable to print-scan attacks. Then, this paper proposes a novel adaptive binary image watermarking algorithm resistant to print-scan attacks. Specifically, the binary image is first divided into disjoint blocks, and the black pixels in each block are counted, thereby transforming the binary image into a grayscale-like image. Then, DCT is applied to the derived grayscale-like image. Subsequently, the adaptive algorithm based on an optimization problem is proposed to modify suitable DCT coefficients for watermark embedding. A visual quality constraint is set to ensure image quality and optimize watermark strength. After modification, the transformed image is reverted to the spatial domain to determine pixel changes in each binary image block. Finally, edge detection is utilized to select and modify cover binary image pixels. Experimental results show that the proposed method maintains excellent image visual quality while effectively resisting print-scan attacks.