Digital watermarking is an emerging research area for copyright protection and tamper-proofing of electronic documents and media, which are two of the most important problems to be solved in the cyberspace society. A window-based watermarking approach for tamper-proofing is proposed in this paper. The embedded watermark can be used as the proof of whether the stego-image has been modified. This approach is based on the realization that it is not necessary to verify whether the whole image has been altered. However, on the contrary, it is quite often that only some particular objects, e.g., persons or flowers, in an image are concerned by the image-holder or the image-receiver. This approach first allows the user to select several windows from an image, in which the concerned objects of the image are included. Then, a watermark is stretched in order to cover all the selected windows. But only the corresponding parts of the watermark, which have the same locations as the selected windows, are embedded into the image as the tamper-proofing seal. Even though this embedded watermark is visible, but it is difficult or troublesome to remove the visible watermark or to split the watermark away from the stego-image. Since all the concerned objects in the stego-image are watermarked, thus, if a forger tries to modify this stego-image, the visible watermark should be changed as long as the concerned objects are falsified. We give the features that a visible watermarking for tamper-proofing should be fulfilled and we also design a mechanism to satisfy these features. To verify whether the stego-image has been modified, it is simply to superimpose the stretched watermark on the stego-image. If the partial visible watermarks shown in the selected windows correspond with those in the watermark, we can conclude that the stego-imae is not fake.
展开▼
