A novel semi-blind watermarking algorithm based on fractal dimension and image feature

作者：Ni RR^*; Ruan QQ

来源：Chinese Journal of Electronics, 2004, 13(2): 293-299.

• semi-blind watermarking
• fractal dimension
• image feature
• discrete cosine transform
• Arnold iteration transform
• coefficients relations

This paper presents a novel semi-blind watermarking algorithm based on fractal dimension and image feature. An original image is divided into blocks with fixed size. According to the idea of the second generation watermarking([1]), the image is analyzed using fractal dimension to attain its feature blocks containing edges and textures that are used in the later embedding process and used to form a feature label. The watermark that is the fusion of the feature label and a binary copyright symbol not only represents the copyright symbol, but also reflects the feature of the image. Arnold iteration transform is employed to increase the security of watermark. Then, DCT (Discrete cosine transform) is applied to the feature blocks. The secure watermark that is adaptive to the individual image is embedded into the relations between middle-frequency coefficients and corresponding DC coefficients. The detection and extraction procedure is a semi-blind one which does not use the original image but the watermark. Only those who have the original watermark and the key can detect and extract the right watermark. This makes the approach authentic and have high security level. Experimental results show that this algorithm can get good perceptual invisibility, adaptability and security. And it is robust against cropping, scribbling, low or high-pass filtering, adding noise and JPEG compression.

• 出版日期2004-4
• 单位北京交通大学