A Study Of Digital Watermarking Based On Digital Halftoning Technologies

Digital watermarking is an efficient measure to protect copyright. Because of the complexity of the print and scan manipulations, the current digital watermarking methods can not availably resist print-and-scan attack, and it is difficult to protect the ownership rights of the printed products. To solve the problem, the influence which the print and scan manipulations impose on an image is analysed, and the hardcopy watermarking methods based on halftoning technologies are investigated. On the other hand, due to image data is large, it is necessary to compress an image for saving the storage space or the transmission time. In order to compress image and protect copyright meanwhile, the robust watermarking methods jointing compression & watermarking are discussed.Firstly, the influence which the print and scan manipulations impose on an image is analysed, and the various attacks which the hardcopy watermarking must resist are summarized; A new standard for categorization is set, by which the hardcopy watermarking methods are divided into four categories,and the characteristics of the every category is summarized. Secondly, the watermarking methods based on compression domian are studied and classified, and the shortcomings of the current watermarking methods based on compression domian are pointed out. On the bases, we design two hardcopy watermarkings and a watermarking based on compression domian starting from three different aspects.To improve the robustness of the hardcopy watermarking, a halftone watermarking method with block-directed parity-matched error diffusion is proposed.The method is developed in parity domain based on pixel block. Especially, the parity sum of a pixel block is defined by comparing the average of the pixel block with an image-dependent threshold. By altering the pixel block’s parity based on noise-balanced block error diffusion, watermark is spread into the host image. Watermark is retrieved by employing each pixel block’s parity and majority voting strategy, not referring to the original image. Compared with the state-of-the-art method in parity domain, the results indicate that the proposed method has high watermark rate and watermark rate flexibility. Moreover, it is capable of extracting watermark directly from the attacked watermarked image without quantizing it into a halftone image. And it can achieve high robustness against common attacks and print-and-scan attack of different types of printers and scanners.To utilize the different advantages of embedding watermark in transformation domain and embedding watemark in spatial domain for improving the watermarked image quality and the robustness, an attribute-based halftone watermarking method is proposed. An attribute image as a good connection between original grayscale image and its halftone image is defined firstly. Then the attribute image is used as watermark carrier, the watermark is embedded by changing the DCT(Discrete Cosine Transform) coefficients of the attribute image. The watermarked halftone image is found by solving a modified direct binary search problem. The watermark decoding is the inverse process of the watermark encoding, not referring the original host image or the original watermark. Experimental results demonstrate that the proposed scheme in comparison with other similar methods maintains high watermark capacity with good image quality, high robustness. Moreover, the framework of the proposed method is universal, and the three stages include attribute image extraction, watermark embedding in attribute image, watermarked halftone image searching can be replaced with other operations according to different application.To overcome the shortcomings of bad image quality or robustness of the current methods, a robust watermarking method based on adaptive order dither block truncation coding is proposed. Firstly, the image is adaptively divided into smooth regions, texture regions and edge regions using a quadtree segmentation, according to the information entropy and the variance, and then these blocks are coded separately. Secondly, we exploit the bit plane obtained from adaptive order dither block truncation coding, and the watermark is embedded by changing the parity sum of the pixels within a window in the bit plane. Lastly, the watermark is retrieved with the parity sum and the majority voting strategy, not referring to the original image. Compared with the similar algorithm, the results indicate that the proposed method overcomes the inherent conflict that the larger block accompanies the more obvious block effect and the smaller block accompanies the lower coding rate in traditional block truncation coding. The proposed method can achieve good image quality with low bpp(bits per pixel), as well as high watermark rate and watermark rate flexibility. Moreover, it is robust to common attacks, especially to Salt & Pepper noising, contrast adjustment, rotating and cropping attacks.