Research On Image Watermarking Technology Against Geometric Attack Based On Transform Domain

Image watermarking is an important object in information security researchfields, it plays a very important role in Multimedia content authentication and digitalworks confirmation. Currently, many algorithms of image watermarking have goodrobustness against normal signal attack, but are insufficient against geometric attack,even small rotation, scaling, etc. Therefore, it has the vital significance of studyinggood robustness of image watermarking technology to resist geometric attack.Because of the robustness is poor based on space domain, now the technologiesof image watermarking are mainly implemented in transform domains. Wavelet ismostly used as one of them. However, the watermarking is processed merely bytraditional wavelet has better robustness against normal signal attack, is short ofperformance in orientation recognition of geometric attack. Finite ridgelet transformhas an orientation recognition function compared with wavelet. So it has goodproperties of orientation recognition. The fratfd transform is an improvement of finiteRadon transform. The fratfd transform also has good properties of orientationrecognition. According to the advantage of orientation recognition in finite ridgelettransform and fratfd transform, the two transforms are combined with wavelettransform in this paper. The main research results are as follows:(1)The method of image watermarking against geometric attack based on finiteridgelet transform is proposed. First the image has1D finite ridgelet transform.Second after3D wavelet transform, the watermarking is embedded to the coefficienceof ridgelet transform. Then finish the image reconstruction. Experimental resultsdemonstrate good robustness against rotation, scaling, shearing, translation ofgeometric attack and JPEG compression, median filter, Gaussian noise, salt peppernoise of normal attack.(2)A method of image watermarking against geometric attack based on fratfdtransform is proposed. First the image has fratfd transform, then the watermarking has3D wavelet transform. The coefficience of low-frequency after decomposing isembedded to coefficience of fratfd transform. Then finish the image reconstruction.Experimental results demonstrate good robustness against rotation, scaling, shearing,translation of geometric attack and JPEG compression, median filter, Gaussian noise,salt pepper noise of normal attack. The performance of translation attack has been improved compared with the method of finite ridgelet transform. So it achieves theexpected results.