Analysis Of An Svd Watermarking Algorithm Contourlet Domain Watermark Embedding And Detection Algorithms

In this thesis, we summarize several fundamental issues of watermarking, includingbasic conceptions, applications, models, classical algorithms and the evaluation of itsperformance. We also discuss the relationships between watermarking and other fields ofimage processing. In addition, we list some of the most important international academicconferences.Due to the robustness of the singular values of the image, some scholars establishedSVD-based watermarking algorithms in both spatial domain and frequency domain, suchas DCT and DWT. There is, however, a disadvantage among all these algorithms, which isdiscussed through numerous experiments.The efficient representation of images and adaptive watermarking schemes are twointeresting research topics. We explorer the multidirectional and high redundancy of thenonsubsampled Contourlet transform, as well as the masking phenomena in human visualsystem. By computing the masking affect of the high frequency subbands, we design anadaptive image watermarking algorithm. Experiment results demonstrate the robustness ofthe proposed scheme.We finally implement multiplicative watermarks in the Contourlet transform domain.By modeling the coefficients of the Contourlet transform in the high frequency directionalsubbands using generalized Gauss distribution, we establish the uniformly most powerfuldetector, locally most powerful detector, and universal optimum detector.