Algorithm Research For Image Digital Watermarking Based On Wavelet Transform

With the popularization of Internet and the rapid development of multimediatechniques, digital watermarking, an e?ective way of copyright protection, has been aheated topic in multimedia and signal processing domain. In fact, digital watermarkinghas been an important branch in information security field in recent years.This thesis is devoted to still image. The characteristics, types and general modelsof digital watermarking technique are analyzed systematically. Some current maindigital image watermarking algorithms are discussed. Based on the wavelet transform,three kinds of e?ective digital image-watermarking algorithms are put forward withtechniques of Hilbert scanning, human visual model and chaotic encryption. The maincontributions are the following:1. Basing on the quantification of image blocks'average value, A new self-adaptiveblind watermarking algorithm is proposed. Taking full advantage of the wavelet do-main and Hilbert scanning, the watermark signal is embedded by modifying the valueof corresponding bands of 3 neighboring blocks. The watermarking is invisible androbust. At the same time, the watermark can be extracted without original image.The experimental results and the capability analysis show that the proposed algorithmis e?ective.2. The low frequency coe?cient concentrates the majority energy of the signalafter the wavelet transform, and is very important to the signal. Because watermarksignal values are much smaller than low frequency coe?cients, the e?ect to the imageis little. According to this theory, a robust image-adaptive watermarking algorithm isimproved. The watermark is embedded into every band of image blocks. The proposedalgorithm is robust to every compression and common image processing operations.3. Two kinds of public watermarking algorithms based on the false pre-process ofthe watermark signal are proposed. Firstly, we analyzed the problem of the originalodd-even quantification algorithm in detailed. And on this foundation, a robust image-adaptive watermarking algorithm is proposed based on the wavelet transform. Theposition of image blocks and wavelet coeffcients are determined by chaotic processedsequences. The watermark signal is embedded into the middle band of the image afterDWT . By taking full advantage of the masking characteristics of the human visualsystem and the adjustment of quantification steps according to the original image, theadaptation was ensured. The proposed algorithm is invisible and robust. Secondly, arobust public watermarking algorithm based on image blocks in DWT (discrete wavelettransform) domain is presented. Corresponding to the watermark signal, the maximalcoe?cients in the low frequency band of every image blocks are modified. Finally, the simulate experiment results of two algorithm show that they are effective.This thesis's watermarking algorithm is simulated in Matlab7.0 environment. Allthe experimentation results show that this thesis's algorithm is suitable for many fa-miliar image processing such as the noise, filter, image compression and shear, slice,and so on, and it has strong robustness and excellent invisibility.