Study On Image Digital Watermark Algorithm Based On Contourlet Transform

As a new research focus in the field of informationsecuritytechnology,digital watermarkhas become an important means ofcopyright protection and content authentication for multimedia data. When itwasput forward, digital watermark technology aroused the concern andattention of researchers immediately; they also made a lot of good andmature watermarkalgorithm, especially in the field of image copyrightprotection.In the study process ofwatermark algorithmfor image, we find that a lotof watermark algorithms used discrete-wavelet transform, whichhas strongrobustness in terms of resistance to compression mainly attributed to itscompatibility with the image compression standard on JPEG.However,wavelet theory cannot achieve fine resolution to images, thus the selection ofembedding region is quietlylimited, resulting in the robustnessofwatermarkalgorithm based onwaveletis not strong.Aim atthe problem that robustnessof watermark algorithm based onwavelet transform is not strong,we put forward an image watermarkalgorithm based on Contourlet transform. Contourlet transform in thealgorithm has a dual filter bank structure constituted with the Laplacianpyramid transformand direction filter bank. The filter bank candecompose animage into a plurality of low-frequency sub-band images onmultiplescales and a plurality of high-sub-band images onmultiple directions. If applyingthetransform theory to thecarrier imagein watermark,we can decompose itintomultiple areas available for the embedding of watermark, which can wellcompensate for the insignificance of wavelet transform, making thealgorithm has strong robustness. The main work of this paper are as follows:(1) Before the watermark is embedded,we do various image-processingoperations including gray processing, median filtering processing, histogramequalization, binarization, gray encryption, position encryption of the pixel,etc., to ensure the watermark image is not disturbed and protect the securityof the watermark to the maximum extent.(2) Proposing a dual-factors encryption algorithm based on the sequenceof chaotic, the algorithm is one of the important research contents of thisarticle. Applying the modern cryptography system to the image encryption,which use two chaotic sequences as the watermark encryption keys. One ofthe keys is used in gray encryption to disrupt the information of the image’shistogram statistics and reduce the probability of being attacked,thus,providing the first layer of security protection for the watermark image;In addition,we apply another chaotic sequence as the key of watermark’slocation encryption, furtherly, to disrupt the attacker’s sight on the visual andposition,thus,providing the second layer of decoding protection forwatermark image.(3) Preprocessing of the original carrier image, to ensure the security of embedding position of the watermark and the robustness of encrypted carrierimage, this paper applies the invariant centroid theory of imagenormalization technique to the original image to get the important area of theoriginal image, which is the area where watermark information is to beembedded in.(4) Applying Contourlet transform to the key area of normalization onthe original image to obtain its high-frequency sub-band images and selectthe sub-band image with the most energy as the embedding area of thewatermark, then use a certain embedding rules to complete the watermarkembedding, finally use inverse Contourlet transform to get the encryptedimage, accordingly, completing watermark embedding.(5) Combining with this algorithm, we design and develop acorresponding digital watermark system. The System has five modules,mainly includes file, image processing, watermark embedding, watermarkextraction and attack test, etc. It nearly covers all of image processingfunctions used in this paper. Experiments show that this system can wellrealize the algorithm in this paper. It has some application value.