Geometrically robust image watermarking based on Jacobi-Fourier moments

Abstract

Rotation, scaling and translation (RST) attacks can desynchronize the watermark detection so that many watermark systems failed. A geometrically robust image watermarking strategy based on Jacobi-Fourier moments (JFMs) is proposed. The Jacobi moments of the original image are first extracted as original moments; then the watermark image is embedded into the global or local area of the original image, and the Jacobi moments of the area are extracted. When the watermarked image is not attacked, the watermark can be retrieved by using the margin of the original moments and the moments of the embedded area. When it is attacked, the watermark can also be got in that way, and the original moments need to be transformed. It can be concluded that Jacobi-Fourier moments perform better than Zernike moments (ZMs) for small images. Meanwhile, the watermark is also robust to scaling and rotation as well as regular attacks such as added noises.

Keywords

Original Image / Watermark Image / Zernike Moment / Invariant Moment / Small Image

References

[1] RenH. P., PingZ. L., ShengY. L.. Acta Optica Sinica, 2004, 24: 5

[2] LiL.-d., GuoB.-l., SunL.. Journal of Optoelectronics · Laser, 2007, 18: 117

[3] KimH. S., LeeH. K.. IEEE Transactions on Circuits and Systems for Video Technology, 2003, 13: 766

[4] Chen Q, Yang X L and Zhao J Y, Canadian Conference on Electrical and Computer Engineering, 2005, 1340.