Adaptive Digital Watermark Based On Wavelet Packet And Edge Detection Algorithm

Digital Watermarking is an effective measure to guarantee the security and validity of information. Imperceptibility and Robustness are contradictory characteristics, if one is strong, the other will be weakened. Therefore, watermarking algorithms usually satisfy one characteristic while dissatisfy another. For the embedding position, low frequency area can satisfy robustness well, while high frequency area imperceptibility. So how to harmonize the two characteristics is key points of the watermarking research.On the basis of analyzing existing algorithms deeply, this thesis mainly study watermarking algorithm based on wavelet. It Improves flaws of existing adaptive algorithms, and brings forward two new adaptive algorithms based on wavelet packet and wavelet edge detection.The main achievements of this thesis are as following:1. The thesis introduces the development of digital watermark methods, in which emphasize particularly on methods based on wavelet. It analyzes and compares various of watermarking algorithms through reading a large number of literatures, and summarizes their advantages and disadvantages.2. The thesis studies Wavelet Analysis, Multi-resolution Analysis and Wavelet Packet Analysis. According to the above theory, locating suitable embedded position for watermark via ordering, then computes corresponding embedded intensity adaptively. Thus, this algorithm harmonize imperceptibility and robustness effectively.3. On the basis of analyzing watermarking algorithms based on wavelet packet, a new method based on wavelet edge detection is proposed according to its excellent characteristic. This method successfully resolves how to select fitting embedded position and intensity more accurately through comparing every sub-block with the "sample" root in edge detection. Consequently, the algorithm holds more steady watermarking capabilities.