Geometrically Robust Image Watermarking Based on SIFT Feature Regions
Gao Xin-bo · ACTA PHOTONICA SINICA · 2009
A novel robust image watermarking method using local feature regions is proposed.Scale Invariant Feature Transform(SIFT)is first adopted to extract feature points and to construct local circular feature regions which are invariant to scaling and translation.Then,these regions can achieve rotation invariance by normalizing their corresponding domiant orientations.Finally,watermarks are embedded in the mid-frequency DFT coefficents of these non-overlapped local regions.Experimental results show that the proposed scheme not only gains high image quality,but also is robust against geometric attacks as well as common signal processing attacks,even some combined attacks.