A Multiple Watermarking Algorithm Based On KICA And Wavelet Domain Transform

With the development of network and multimedia technology,how to apply watermarking technology effectively to improve the digital copyright protection will be a focus of information security.By analyzing the wavelet transform,in this paper,a novel multiple watermarking algorithm based on bandelet transform is proposed,and a feature extraction algorithm is performed to the process of watermark extract.Firstly,we get the frequency coefficients after performing several wavelet transforms,and embed the one watermark into the frequency.Secondly,we implement a bandeletization to higher coefficients,so that the geometric flow is traced.The texture and edge considered for feature parameters of the image are used to high frequency watermarking.And then we embed the other watermark into bandeletization frequency.At last,we use kernel independent component analysis(KICA) to extract the watermarks.Experimental results demonstrate that the proposed approach has the capability to resist strong attacks and a good invisibility.