Digital Watermarking Based-on Independent Component Analysis

This paper first introduces the digital watermarking in detail, gives a common framework for watermarking system, and demonstrateds the general methods of watermarking production, embedding and detection. Then some important research results of watermarking are presented and the attacks towords watermarking system are discussed.Progress and applications of independent component analysis are introduced. Principle and implementations of ICA are discussed. Some important algorithms and their relationships of ICA nowadays are listed systemically. FastICA which is one algorithm of ICA with higher running speed is analysed.This paper combines the ICA with the Discrete Cosine Transform and the Discrete Wavelet Transform, and analysises some new methods of digital wantermark. Computer simulations show the combine of these algorithms can improve the security of watermarking greatly.In this paper, a method to protect image by its eigenvalue is proposed. The method combines the independent components analysis with digital watermarking. And such schemes also produce new ideas for the development of watermarking and the application of ICA. And independent component analysis method is adopted to estimate the differentiation of eigenvalue, which performs the image protection. Experimental results show that this scheme improves the performance. The original image modification will not be a necessary condition; therefore, this approach is promising in practice.At last, perspective of digital watermarking Independent component analysis and image characteristic protecting are discussed.