Proceedings of the
The Nineteenth International Conference on Computational Intelligence and Security (CIS 2023)
December 1 – 4, 2023, Haikou, China
A Zero-watermarking Algorithm Against Rotation Attack Based on DCT
College of information Engineering, Jiangxi University of science and technology, China.
ABSTRACT
To improve the anti-geometric rotation attack ability of the zero-watermarking algorithm, A zero-watermarking algorithm against rotation based on DCT is proposed. Firstly, the low-frequency region is extracted by a two-level Non-Subsampled Contourlet Transform (NSCT), the ring formula is used to extract the ring with different radii in the low-frequency region. These rings are recombined into a new sequence with a similar number, One-dimensional DCT was performed on the new sequences and the ave rage energy was calculated to extract the feature sequences. Then the dimension of the sequence is rai sed to form a feature matrix and the watermark binary image is performed by XOR operation to complete the zero watermark construction process. Before constructing a zero watermark, the watermark image is encrypted w ith Arnold scrambling to improve the security. Experimental results show that the algorithm is robust to geometric rotation attacks and common image processing attacks. Compared with SIFT and oth er correction algorithms, the proposed algorithm gets rid of the dependence on correction tools, can extract watermarks directly, and has lower computational complexity.
Keywords: Anti-geometric rotation attack, Non-subsampled contourlet transform, One-dimensional DCT, Ring, Zero-watermark.
Download PDF
College of information Engineering, Jiangxi University of science and technology, China.
ABSTRACT
To improve the anti-geometric rotation attack ability of the zero-watermarking algorithm, A zero-watermarking algorithm against rotation based on DCT is proposed. Firstly, the low-frequency region is extracted by a two-level Non-Subsampled Contourlet Transform (NSCT), the ring formula is used to extract the ring with different radii in the low-frequency region. These rings are recombined into a new sequence with a similar number, One-dimensional DCT was performed on the new sequences and the ave rage energy was calculated to extract the feature sequences. Then the dimension of the sequence is rai sed to form a feature matrix and the watermark binary image is performed by XOR operation to complete the zero watermark construction process. Before constructing a zero watermark, the watermark image is encrypted w ith Arnold scrambling to improve the security. Experimental results show that the algorithm is robust to geometric rotation attacks and common image processing attacks. Compared with SIFT and oth er correction algorithms, the proposed algorithm gets rid of the dependence on correction tools, can extract watermarks directly, and has lower computational complexity.
Keywords: Anti-geometric rotation attack, Non-subsampled contourlet transform, One-dimensional DCT, Ring, Zero-watermark.
Download PDF