Affiliations:
Faculty of Computer and Software Engineering, Huaiyin Institute of Technology, Huai’an, China
To improve both invisibility and embedding capacity, this study proposes a reversible watermarking algorithm based on diamond pixel pairs with complementary embedding. To avoid pixel value overflow, the carrier image is first preprocessed and divided into blocks. Each block is then separated into two half-planes, gray and white. The watermark is embedded in both horizontal and vertical directions. In the horizontal direction, the target pixel in a smooth block is compared with four reference pixels in its cross-shaped neighborhood to calculate their difference values and construct a histogram. The first embedding is carried out in the gray half-plane using the histogram shifting method, followed by a second embedding in the white half-plane. The same process is applied in the vertical direction to reduce the distortion introduced by horizontal embedding. Experimental results indicate that, under the same embedding capacity, the proposed method achieves an average PSNR (Peak Signal-to-Noise Ratio) that is 6.92 dB higher than that of the non-complementary embedding scheme. Moreover, compared with other related algorithms, this approach demonstrates superior embedding performance for the same PSNR value.
Reversible watermarking, Complementary embedding, Diamond pixel pairs, Histogram shifting, Image distortion reduction
https://doi.org/10.21833/ijaas.2025.12.002
Zhang, Z., Li, F., & Jin, S. (2025). A reversible watermarking algorithm based on complementary embedding of diamond pixel pairs. International Journal of Advanced and Applied Sciences, 12(12), 9–18. https://doi.org/10.21833/ijaas.2025.12.002