The Study And Analysis Of Image Watermarking Algorithm Based On Wavelet Multi-resolution

Digital watermarking technique has been a frontier research field in academic circles in recent years and still has been in development stage. It is widely used in all kinds of digital products to protect the copyright and intellectual property of authors or owners. Therefore, the research of digital watermarking is of important significance.Based on the introduction of the basic knowledge of digital watermarking and application of wavelet analysis in digital watermark, the thesis is aimed at the algorithm of high and low frequency embedding for watermarking of wavelet domain, which has the problem of poor image quality and small capacity, and proposed a new method built on wavelet multi-resolution analysis and evaluation of the watermarking algorithm. Through theoretical analysis and simulation experiments, the existing watermark algorithm is studied and the results obtained as follows:(1) The thesis is aimed at the commonly adopted algorithm of high and low frequency embedding for watermarking of wavelet domain, which has the problem of poor image quality and poor attack capability, and adopts an improved algorithm of mid-frequency position with the way of fusion. The new algorithm is better at solving the problem of watermark image of invisible and robust needle, and is simpler and more practical.(2) For small watermark capacity problem, the thesis suggests making full use of the mathematical properties of generalized cyclic matrix, and generates the idea of a generalized cyclic matrix digital watermark embed watermark to the carrier image. Experiments show that the watermark is not only good for cover but also good for robust of noise, image filtering, geometry transform, the image file format conversion processing.(3) Two kinds of methods are employed in the validation image watermark. Using the generalized inverse matrix of the original image watermark, the thesis puts forward a new method of the grey relational watermark transparency to overcome the human evaluation based on its subjectivity, fatigue and weakness.