Research On The Digital Watermarking Technology In Wavelet And Contourlet Transform Domain

With the development of the multimedia and the network technology, the application of the digital product is more and more wide. The security of digital information becomes the concern in people’s life. Especially, the copyright protection of digital product is more important and urgent. Of all the solutions to the copyright protection, the watermarking technology is the more effective one. Therefore, the study of this technology becomes a focus in the academic world.The dissertation has mainly studied salt-and-Pepper noise denoising, digital watermarking in wavelet domain and Contourlet domain for still images. The main work shows as follows:1 .Image pre-processing. The digital watermarking images are corrupted easily by the noise. To solve the problem of salt-and-Pepper noise in gray image, according to characteristics of the noise, a new denoising algorithm is proposed based on bi-directional prediction algorithm. The experimental results demonstrate that the algorithm presented is good both at preserving the details of the image and at denoising, which provides conveniency for the watermark extracting from salt-and-Pepper noise corrupted image.2.Wavelet transform is a good tool to analyze image because of its time-frequency analysis ability and multiresolution. A digital watermarking algorithm based on block energy analysis in wavelet domain is proposed. Horizontal and vertical sub-images of the third level detail sub-images decomposed from the original image using three levels discrete wavelet transform are dividedinto non-overlap blocks, and the meaningful watermarking information is embedded into the larger energy blocks. The experimental results demonstrate that the proposed watermarking algorithm is not only invisible, but also robust to many attacks, such as JPEG, noise and filtering etc.3.A novel watermarking algorithm based on Odd-Even Quantization in Contourlet domain is proposed. Contourlet transform is a flexible multiresolution, local and directional image expansion using contour segments, which is a true two-dimensional transform that can capture the intrinsic geometrical structure. According to the characteristics of Human Visual System, insensitive to the image’s texture and edges, the meaningful watermarking information is embedded into the bandpass directional subbands by the rule of even or odd adjudging. The experimental results demonstrate that the proposed algorithm is not only invisible, but also robust to many attacks,such as JPEG, noise and filtering etc.