| The digital media has become a main way for information communication along with the wide use of network. At the same time, multimedia information security problems including copyrights piracy emerged as side effects of the popularity of digital representation and distribution over network. Information hiding, also called Steganography, digital watermarking technology and multimedia information encryption methods provide novel ways to solve these security problems for multimedia information.In this dissertation, we study the technologies and application of multimedia information security based on chaotic sequence and wavelet transform in the area of digital image encryption and digital watermarking.Image encryption derives from the early classic encrypted theories. Its purposes is to encrypt a color image. Any scene and outline of the old image doesn't exist in the encrypted image and the deencrypted image is the same as the old one. And Digital watermarking is to embed some digital labels or signatures into host multimedia data in such a way that it causes no noticeable perceptual distortion to the original data. It has important potential applications in copyright protection of digital media, secret communication and hidden annotation of host data.Facing with this problems, aims of this paper will mainly focus on following aspects:1. Combined with the Logistic mapping in non-linearity dynamics system, an image-encrypting algorithm based on modular arithmetic in wavelet transform is designed. Firstly, a natural number chaotic sequence is created with the key. Then the image matrix is transformed with this chaotic sequence, using the Magic Cube Transformation. The results indicate that the proposed algorithm can get a satisfying effect. Finally, the characteristics of the algorithm are summarized and the aspects of the subsequent work are prospected. Similarly, another new method image-encrypting algorithm based on Magic Cube Transformation is proposed.2. By analyzing the characteristics of visible watermarking, an adaptive visible watermarking scheme based on texture features in wavelet transform is proposed.In order to adapt the scheme to both the color image and the gray image , While embedding the watermarking into the image, the scaling factors are determined by the local luminance feature and texture features of the image. It can improve the embedding strength and the watermarking robustness. From the experimental results, the stego-images are extremely robust to common image processing such as median filter, JPEG compression and denoising after adding the stochastic noises. The distributions of those gray value histograms of stego-images are more concentrated than those of cover2i2mages. It seems that removing the visible watermarking without damaging the cover2image is very difficult and time consuming.3. We propose a color-level adaptive blind watermarking arithmetic based on wavelet transform and visual perception character. This arithmetic divides the wavelet coefficients of carrier image into 27-classes by the background luminance, the texture and the edge mask character of visual system; then after encrypting watermark image with Magic Cube Transformation, the YCbCr model is employed, it embeds the watermark sequences in the carrier image with different intensity according to the coefficients' classes of carrier image in Y channel. This arithmetic can detect the digital watermark without the original carrier image and the original watermark. Experimental results prove that this arithmetic can hide the meaningful watermark efficiently in the carrier image, and it is robust enough to resist some image degradation processing such as clipping, JPEG compression, adding noise, image enhancing, filtering and so on. |
