| The digital watermarking technology, which arouses the international academicand business attention, is the fastest development hot technology in the multimediainformation security research field. This technology is widely used in the digitalproducts copyright protection and identity authentication.First, based on the studying of the text watermarking and the digital watermarkingtechnology, the model of the text digital watermarking technology and its performanceevaluation criteria are introduced. Meanwhile, the mainstream algorithms of the textdigital watermarking technology and the impacts of the print scanning process on theimage quality are discussed and analyzed in this thesis.Second, by researching on the text word space and line space distribution, a binarytext image watermarking algorithm based on the word space and line space is broughtforward, which is very robust for the print scanning channel. The algorithm ensures thatthe watermark information can be correctly extracted under the blind watermarkextraction mode. At the same time, it can achieve high watermark embedding capacity,high invisibility, the simplicity and security of the watermark embedding and extractionsystem.Finally, because of the shortcomings of one kind of the watermark algorithm basedon the pixel flip, which are low embedded capacity and bad image quality afterembedding the watermark, and through the introduction and experimental comparisonof three kinds of pixel flip method, a new pixel flip model is put forward. Then, byoptimizing the method of determining the critical value K and encoding for the binarywatermark image, another binary text image watermark algorithm is proposed, which isanti-print-scan and combines with the spatial and frequency domain. The experimentalresults show that this algorithm has very good capability of anti-hard-copy and also canincrease the embedding capacity and improve the robustness. |
