Wavelet-Based Digital Watermarking Technology For Image

Digital watermarking is a new technique for information security. It embeds a special invisible mark into digital multimedia data, and then carries out the functions such as information hiding, storing, and copyright protection by tracking the use of the image by detecting the watermark. At present, watermarking technique has been a kind of effective means to multimedia copyright authentication and its integrity protection.This paper takes static digital images as the study object. On the basis of analyse the status quo of the domestic and international digital watermarking technology. We propose two improved watermarking algorithms. Then analyse the advantages and disadvantages of the two algorithms. The details of the work are as follow:(1) The watermark algorithm based on combining the four band wavelet with the singular value decomposition.The algorithm used four band wavelet transform multi-band characteristics and the stability of singular value decomposition, as well as the similarity of approaches pixels to achieve the watermark data embedded. Firstly, the four-band wavelet transform is applied to the host image. Secondly, the low-frequency area is decomposed into small blocks, and the mean value of these blocks, which stored in a matrix, is calculated. Finally, this matrix is decomposed by mean of singular value decomposition, watermark data are embedded in the decomposed matrix. The watermark extraction does not need the original image and the original watermark image. Experimental results show that the algorithm is very good invisibility and common attacks is very robust.(2) The algorithm gives full consideration to dual tree complex wavelet transform decomposition proceeds of regional diversity. The different watermark embedding strategies are adopted in all regions. First, the dual tree complex wavelet transform is applied to the host images. Then, the different embedding strategy for all the transform directions subband are adopted, that is, in the direction of high-frequency band, the adjustment factor strategy is used, and in low-frequency subbands, watermarking data is embedded using distributed wavelet transform. The watermark extraction does not need the original image and the original watermark image. Experimental results of a mock attack to test the validity of the algorithm.(3) Through horizontal comparison of the two algorithms proposed in this paper, then analyze the performance of these two algorithms.