Research On Adaptive Digital Watermarking Algorithm Based On Balanced Multi-Wavelet

With the rapid development of digital technology and the increasing popularity ofthe Internet, network security has become increasingly serious day by day. As aneffective method to protect the copyright of digital products, digital watermarking hasbeen receiving wide attention. In recent years, most studies in the field of digitalwatermarking technology focused on discrete wavelet transform. With the emergence ofbalanced multi-wavelet transform, which has some good properties such asorthogonality, symmetry and pre-filtering being no need, so the balanced multi-wavelettransform is a trend that can resolve the conflict between watermark invisibility androbustness.Based on the characteristics of the balanced multi-wavelet transform, combiningwith the human visual system, and this paper obtains the improved model and justnoticeable difference (JND) formula, with the improvement of the three aspects ofvisual masking model cover factor, brightness and texture cover factor. For theproblems in the current digital watermarking model, a new balanced multi-wavelettransform based adaptive blind watermarking algorithm is proposed. The original imageis processed by balanced multi-wavelet transformation, the JND threshold of the detailsubband coefficients frequency of sub-block area image can be calculated through theimproved human visual model. The watermark image which is about to be embeddedwas processed by Arnold transformation, and the watermark will be adaptivelyembedded into the sub-block transform domain of the original image, in accordancewith strength and position, and ensure the invisibility, robustness and safety of thewatermark.The algorithm proposed in this paper is blind watermark; watermark can beextracted without the original image, simple and has high applicability. Experimentalresults show that in premise that the information of watermark is invisible and thealgorithm has good robustness for the common JPEG compression, filtering, addingnoise and other attacks. It will study the image watermark is embedded into the othertransformation areas and the algorithm is extended to multimedia in the future.