Study Of Color Image Digital Watermarking

With the development of computer network and multimedia techniques, especially the popularization of Internet, digital multimedia such as digital image, audio, video, has become main means of information communication. People have a quite easy access to digital products through Internet. Because digital products can be easily duplicated and edited illegally, a very critical issue of digital media's information security and copyright protection has emerged. A new technology named digital watermark which is able to protect the information security and copyright efficiently has been proposed in the world. Digital watermarking especial digital watermarking of color image has drawn extensive attention, been studied extensively and been widely applied.First of all, this thesis introduces the development background, basic principle, embedding method and system model of digital watermarking technique. Secondly, this thesis introduces the human visual system (HVS), the frequency response function of human visual sense, the discrete cosine transform (DCT) and the discrete wavelet transform (DWT), as well as the pretreatment transformation of image. Based on which DCT domain and wavelet transform domain watermarking algorithm is presented.A new digital watermarking algorithm for color images is presented, in which the binary watermark image is encrypted with chaotic sequence generated by Chebyshev map. And then one of the color components is selected to embed watermark information according to human visual system (HVS) model. Experimental results demonstrate that the watermark is invisible and robust against common noise, crop and JPEG compress etc.This thesis also presents an image digital watermarking algorithm for color images based on discrete wavelet transform. The algorithm takes advantage of the characteristics of discrete wavelet transform and HVS to select the low frequency domain to embed the watermark which has been changed from meaningful binary image into chaos sequence into color image. In dong this, we dramatically improve the robustness of the watermark and realize blind-detection. The results of experiments confirm that the algorithm is robust against the JPEG compression, clipping and median filtering.