Geometric distortion correction through image normalization

Towards robust image watermarking: combining content-dependent key, moment normalization, and side-informed embedding

“…Nomenclature w ori the watermark before ECC w ori len =jw ori j the length/magnitude of w ori K the secret key used to generate w ori w the embedded watermark signal after ECC w len the length of w w e the extracted watermark signal w ecd the extracted watermark after error correction decoding of w e jw ecd j the magnitude of w ecd QðiÞ the magnitude of an embedded watermark value wðiÞQðiÞ Q the constant watermark magnitude the threshold used to indicate the presence/absence of a watermark CDK the content-dependent key CDW the content-dependent watermark 1 There existed another version that was based on log-log map, which could resist change of aspect ratio but is sensitive to rotation. Furthermore, it should be noted that FMT cannot recover from general affine transformations.…”

“…In general, attacks can be roughly categorized into four classes [37]: (1) contain non-geometrical modifications, including filtering, lossy compression, denoising, and sharpening; (2) geometrical attacks that contain local/global transformations, warping, and jitter; (3) protocol attacks that are mainly composed of the copy attack and the watermark inversion attack; (4) cryptographical attacks that are related to the security of keys, such as brute force key search and oracle attacks. Removal attacks are considered to be less challenging because earlier robust watermarking methods [12] could resist them to a certain degree.…”

A geometric robust watermarking algorithm based on DWT-DCT and Zernike moments

“…In Ref. [9], a watermarking scheme was proposed using moment based image normalization which is well known in computer vision and pattern recognition applications [10] . In this approach, both watermark embedding and extraction were performed in a normalized image.…”

Confronting the Synchronization Problem of Semantic Region Under Geometric Attacks

“…Without synchronization the watermark cannot be extracted. Since the recognition of the need for watermarking schemes resilient to geometric distortions, many researchers [1]- [5] try to design watermarking techniques resistant to rotation, scaling, translation and shearing of images.…”

“…Initially, the estimated distortion was corrected and afterwards watermark detection was performed. In [5], a watermarking scheme was proposed using moment based image normalization with a standard size and orientation. Thus, it is suitable for public watermarking where the original image is not available.…”