Study On Color Image Digital Watermark Algorithm Based On Wavelet Packet Decomposition

With computers and networks are becoming popular and widely used, the protection, authentication, security and intellectual property of digital media are becoming prominent problems. Therefore, how to effectively protect the copyright of digital works has become an important and urgent topic in the field of academic research. The emergence and development of digital watermarking technology provides a better solution to solve all the above problems.Firstly, this thesis introduces the background and state-of-art technology of the watermarking technology, describes the basic concepts, the general model, the common attacking methods and the performance evaluation, etc. of digital watermarking. Secondly, it discusses the wavelet and the wavelet packet transform and gives a systematical study on the spatial and transform domain digital watermarking algorithm systems.Lastly, based on the study of the wavelet packet coefficients distribution, a wavelet packet-based watermarking algorithm for color image is proposed. The algorithm chooses an gray image with intuitive visual property as the original watermark, and scrambles it before embedded in order to enhance the safety and concealment. The algorithm selects the optimal wavelet basis with the Shannon entropy criteria, decomposes the host scrambled and the gray-scale image respectively with wavelet packet, selects the embedding location according the energy of subbands. Using HVS model, different embedding strength can be chosen according to the coefficients of the Y, I components of color images at low and middle frequency. The watermarking then can be added to the gray image adaptively.The experimental result shows that the algorithm has better robustness for common attacking such as JPEG compression, noise, filtering, cut, etc. and shows it has good invisibility.