Double Watermark Algorithm Of Color Image Based On DCT

Digital multimedia has been become a main way for information communication along with the wide use of network. Thus, how to protect the intellectual property for digital media becomes an interesting issue on international. As an important embranchment of information hiding, watermark is effective means for protecting copyright. It can identify the signal's integrality or property through embedding additional information into multimedia signal. Watermark applied in color image is researched in this paper.Firstly, background and conditions about watermark is introduced. Definition, sorts and applications are also talked about. Some key technologies about image watermark such as encrypting, color space, transform domain are introduced in detail.Through studying existing watermark technologies, a new double-watermarking algorithm based on DCT for color image is presented to authenticate property and make tracks for peddler. Both a binary image and a logistic chaotic sequence are used as watermark in this algorithm. After transformed the original image from RGB to YCbCr, the binary image which had been encrypted through Arnold transforming is embedded into the DCT low frequency coefficient of the Y component of the image. And then, considering the HVS characteristics of human eyes, the logistic chaotic sequence is adaptive embedded into the DCT AC coefficient of the Cb component of the image.Experimental results show that watermarks are invisible after embedded into image and are robust against various signal processing such as compressing, zooming, cropping, rotating, and noise adding. But when color image is changed into gray level image, the logistic chaotic sequence can not be found out.To avoid the logistic watermark is lost, a little improvement is come up to the algorithm. Both the binary image and the logistic chaotic sequence are embedded respectively into the DCT low frequency coefficient and AC coefficient of the Y component of the image. Experimental results show that after improvement the watermarks are still invisible and robust. Moreover, when original color image is changed into gray level image, both of the watermarks are still existent. So after improvement, the algorithm can be applied to gray levelimage.The results show that the double-watermarking algorithm based on DCT for color image is feasible.