Research On Algorithms And Applications Of Information Hiding

The fast development of Internet makes it convenient to access the multimedia data, but at the same time the Internet has brought the new challenge to the copyright protection. As an effective means for the protection of digital property rights, at present, the technique of digital watermarking is an international important part of the region of information hiding. Digital watermarking is a technique that embeds copyright information, such as figures, sequence numbers, words, image logos and so on, imperceptibly into multimedia data. It can be used for tracing the distribution and usage of multimedia data. The special application of digital watermarking technique requires that the embedded watermark should be not only transparent to human observers, but also robust enough so that it can resist kinds of attack.In this paper, the history of image hiding, research background and development status is first described and particularizes the basis frame of system of digital watermark, the classification and main application of watermarking technology. According to the different ways of information hiding, the performances of existing algorithms of digital image watermarking are compared. Aimed at the shortage of existing algorithms, the researched results of this paper are following:1) A novel joint spatial and frequency domains watermarking algorithm is proposed. Character watermark is an image that is combined from those image-blocks which are extracted from the original watermark image. It can denote the meaning of the original watermark best. As an image, watermark includes the different information. Traditional watermarking segmentation selects the middle region of watermark as the important part. It does not consider the actual content of the image, so it has little practical meaning. The character watermark is extracted through segmenting the watermark image by means of the content of the watermark image. Then the character watermark is imbedded to the robust region of the host image. The dither modulate embedding scheme in wavelet packet coefficients, which belong to texture-active region, not only promises the hiding of large capacity of robust information but also achieves the goal of watermark blind-extraction.2) A novel digital image watermarking algorithm based on singular value decomposition and wavelet packet decomposition is proposed. Each cover image has different luminance information and character. The noise sensitivity of different luminance and texture region has different visual sensitivity according to human vision system. So we divide the image into non-overlapping blocks according to the texture character of the cover image to guarantee the controllable and invisibly.For a region of an image, the bigger variance it has, the more complex texture it possesses. The bigger singular value it has, the higher luminance it makes out. Therefore, the variance can be used to draw the contrast and smoothness of a region of an image and the singular value can draw the luminance information. Based on these features, we give the idea which combines the singular value and variance to segment and describe the regional feature of the host image and propose a novel method for processing the host image using these features. The results of experimentation show that this novel method is more flexible and robust than traditional methods, and can resist the compress with the format of JPEG, specially.3) We propose a novel color image watermarking based on genetic algorithm. The digital watermark algorithms mostly focus on gray level images with long time. With the fast development of multimedia, color images become more and more popular. Nowadays, most algorithms embed watermark in single channel image or intensity image of the color one, but they do not consider that the hiding ability of different color channel vary greatly in different images. The watermarking algorithms for protecting the color image effectively are in a dire need. For this, we propose a novel watermark algorithm based on genetic algorithm (GA). GA can analysis the influence of watermark on original image and power of resisting attack and select the best embedded weight value of each channel. GA can improve the imperceptions of the embedded color image, the robustness and security of the watermark, and fulfills a good compromise between the robustness and invisibility.4) We make use of the memory ability of BP neural network to renew the attacked watermark image, make sure it becomes more clearly than the one extracted directly.Information hiding technique becomes a hot research in recent years. It overcomes the shortcomings of encryption techniques in information security and multimedia authentication, and provides the complementary solution for encryption techniques. Based on summarizing modern cryptography theory, this paper gives an analysis on the current theories and technologies of information hiding, and presents two different steganographic methods for image hiding.1) One method makes use of the noise visibility function to analyze the local properties of host image and embed more secret bits in the textured or edge regions using the substitution table to improve the imperceptibility of stego-image. This method makes use of vision feature of human, and proposed a new classification approach of host pixels. It embeds more hiding information in imperceptible region, and less information in perceptible region. Through experiments, the vision results of testing images improved greatly.2) Based on the basis secret sharing, we propose a novel method of secret sharing. There is a common problem in current secret sharing algorithm: namely, we have to get sufficient r sub-images to get all information about the secret image. If this is not satisfied, nothing about the secret image can we get. But under some actually conditions, only the main information of the secret image we need to know, so the (r, n) threshold algorithm is not sufficient. Ja-chen Lin proposed a secret sharing based on error tolerance to solve this problem to some extent in 2005. But they only use the gray image, resulting in the limitation of steganographic method, and part of the shared images also show the information of the original one.We use the redundancy between the pixels of image to divide the secret image into some sub-images. Then the sub-images are hidden in different host images. This method can guarantee the correctness of secret image in the receiver through the separated transportation of network.It presents the future work in the end of this paper. At present, existing palm-print systems don't employ the mechanism of information hiding. And it can't ensure the security of the system and integrity of biometric information when transporting. We will combine the mechanism of information hiding with palm-print and signature identification system to propose the multi-model biometric identification approach based on biometric information, which can realize the mechanism of multi-protection for security and integrity of information of biometric between the localization and long-distance.