An adaptive robust watermarking scheme based on chaotic mapping
Yumin Dong,Rui Yan,Chenhao Yin
DOI: https://doi.org/10.1038/s41598-024-76101-w
2024-10-21
Abstract:Digital images have become an important way of transmitting information, and the risk of attacks during transmission is increasing. Image watermarking is an important technical means of protecting image information security and plays an important role in the field of information security. In the field of image watermarking technology, achieving a balance between imperceptibility, robustness, and embedding capacity is a key issue. To address this issue, this paper proposes a high-capacity color image adaptive watermarking scheme based on discrete wavelet transform (DWT), Heisenberg decomposition (HD), and singular value decomposition (SVD). In order to enhance the security of the watermark, Logistic chaotic mapping was used to encrypt the watermark image. By adaptively calculating the embedding factor through the entropy of the cover image, and then combining it with Alpha blending technology, the watermark image is embedded into the Y component of the YCbCr color space to enhance the imperceptibility of the algorithm. In addition, the robustness of the algorithm was further improved through singular value correction methods. The experimental results show that the average PSNR and SSIM of the watermarking scheme are 45.3437dB and 0.9987, respectively. When facing various attacks, the average NCC of the extracted watermark reaches above 0.95, indicating good robustness. The embedding capacity of this scheme is 0.6667bpp, which is higher than other watermarking schemes, and the average running time is 1.1136 seconds, which is better than most schemes.