Dissertation > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > General issues > Security and confidentiality > Encryption and decryption

Wavelet-based image watermarking algorithm

Author BaiLinXue
Tutor ChenGuangXi
School Guilin University of Electronic Science and Technology
Course Applied Mathematics
Keywords Discrete wavelet transform (DWT) Digital Image Digital watermarking Human visual system (HVS) Robustness Invisibility
CLC TP309.7
Type Master's thesis
Year 2008
Downloads 308
Quotes 1
Download Dissertation

With the increasing popularity of computer networks and the rapid development of communication technology as an effective means of copyright protection, digital watermarking of multimedia signal processing technology has become a hot research topic areas, but also the field of information security technology is an important branch. The wavelet analysis, also known as multi-resolution analysis, Fourier analysis is developed on the basis of a new subject, has profound theoretical significance and wide range of applications, is the international scientific community highly concerned about the frontier. In this paper, the still image as the research object, the first detailed analysis of the summary of the characteristics of digital watermarking technology, classification and general model, the current typical digital image watermarking algorithm are summarized and discussed in-depth study of the theory of wavelet and wavelet domain digital watermarking technology, then the Hilbert scanning the human visual model and chaotic encryption technology into the digital watermarking technology in the past, made several more efficient wavelet-based digital watermarking algorithm: 1. propose a new DWT-based image block mean Adaptive quantization blind watermarking algorithm that combines the wavelet domain and Hilbert scanning advantages detail by adjusting the adjacent three sub-bands corresponding relationship between the mean size of the watermark from the adaptation, making invisible watermark has good robustness and resistance, and the watermark extraction process without the involvement of the original image. The experimental results are given their anti-attack performance analysis show that the algorithm is indeed effective. 2 frequency coefficients of the wavelet transform signal is concentrated most of the energy of the signal is more important, and usually has a low coefficient larger values, the watermark signal is relatively weak, little effect on the image after embedding. According to this principle, this paper, an adaptive robust watermarking algorithm has been improved in the sub-block of the image after each band are embedded watermark information, making the watermark on a variety of compression and common image processing operations robust sex. 3 based watermark scrambling pretreatment, presented two public watermarking algorithm. First, a detailed analysis of the original problems of the quantization of parity algorithm, based on this paper presents a robust adaptive wavelet-based image watermarking algorithm. The algorithm uses hashing chaotic sequence twice after positioning the embedded watermark image blocks and coefficients location in the wavelet transform frequency sub-band embedded watermark signal, while taking full account of the HVS masking characteristics, based on adaptive generation of quantization step size, thus ensuring the adaptive algorithm, so that a higher invisible watermark and robustness. Secondly, we propose a robust block based DWT image watermarking algorithm publicly, will block layer carrier image after wavelet transform, the control signal is scrambled watermark for each image block low frequency sub-band maximum coefficient changes adaptively watermark embedding complete. Finally, the two methods are given simulation results proved their effectiveness, and are compared with other algorithms. Algorithms mentioned in this article, and experiments were carried out under MATLAB7.0 environment, all the experimental results show that: the proposed algorithm can withstand noise, filtering, image compression, shear, rotation and other common image processing, with a more strong robustness and invisibility.

Related Dissertations
More Dissertations