Harris-based digital watermarking method with angular point geometric attack resistance

Abstract

The invention relates to a Harris-based digital watermarking method with angular point geometric attack resistance, and discloses a Harris-based watermark embedding and extracting algorithm. The method comprises the following steps of: firstly searching watermark embedding areas from a carrier image embedding rotary flag area and a rotation angle detection and water mark region flag area, and resisting rotary attacks and translation attacks through carrying out position adjustment and angle correction on the flag areas so as to complete image preprocessing; after the image preprocessing is completed, extracting feature points by using a Harris operator, screening stable feature points, dividing the stable feature points into different small rectangular areas, selecting feature areas matched with watermark information of corresponding watermark image blocks according to sizes and a quantity of feature areas screened according to a watermark information size requirement, and embedding watermarks to the selected feature areas after carrying out wavelet transformation, so as to enhance the shearing attack resisting robustness of the algorithm and complete the watermark embedding and extraction.

Description

technical field [0001] The invention relates to a digital image watermark embedding method, in particular to a digital watermark method using Harris corner points to resist geometric attacks. Background technique [0002] Nowadays, the era of digital information has already arrived, and multimedia technology is developing rapidly. Whether it can effectively protect the information security of the Internet and the copyright of many digitally printed products has become one of the hot spots for researchers. The robustness of digital image watermarking against geometric attacks needs to be improved. Geometric attacks will have a certain impact on the synchronization of watermark extraction and embedding. Even a small attack will damage the watermark and cause the extraction to fail. Therefore, it is particularly important to find a watermarking method with good performance against geometric attacks. Among the many methods of watermark embedding, the most common is to embed the...

AI Technical Summary

Problems solved by technology

Among the many methods of watermark embedding, the most common is to embed the watermark in the space domain and the transform domain. The traditional method of embedding watermarks in the wavelet domain is usually to directly modify the coefficients in the wavelet domain. Although this method can ensure the invisibility of the watermark Good, but vulnerable to thresholding during image compression and transformation, causing watermark information to be destroyed

The traditional method of watermark embedding is to embed the watermark in the wavelet domain, which is usually to directly modify the coefficients in the wavelet domain. Although this type of method can ensure the invisibility of the watermark is better, it is easily affected by the threshold when the image is compressed and transformed, resulting in watermarking. information compromised

[0003] A new anti-rotation blind watermarking algorithm based on the logo area proposed by Zhou Guangzhou et al. The method of embedding the carrier image into the rotation logo area and the watermark logo area. This algorithm is robust to attacks at any angle of rotation and translation attacks. However, the invisibility analysis of watermark embedding is not done, and only a small amount of watermark information can be embedded due to the influence of the carrier image

[0004] The existing watermark embedding method has poor visibility and low similarity with the original watermark, and the embedding of watermark information is greatly restricted and only a small amount of watermark information can be embedded

Images