This thesis deals with digital steganography and steganalysis. It explains the significance of both disciplines and gives a brief overview of the history in the given fields. The paper separates existing steganographic and steganalytic methods and describes the attributes of various branches of steganography. With respect to the chosen field of steganography (the image steganography), the paper narrows down the set of suitable steganographic and steganalytic methods, whose features are then described in detail. The results of the thesis inlcude experiments that verify the features of chosen steganographic methods and evaluation of steganalytic methods and their success in detection of steganography.

<p>Digital image steganography is the process of embedding information withina cover image in  a  secure,  imperceptible,  and  recoverable  way.The  three  main  methods  of  digital  image steganography  are  spatial,  transform,  and  neural  network  methods. Spatial  methods  modify  the pixel  valuesof  an  image  to  embed  information,  while  transform  methods embed  hidden information within the frequency of the image.Neural network-based methods use neural networks to perform the hiding process, which is the focus of the proposed methodology.</p> <p>This  research  explores  the  use  of  deep  convolutional  neural  networks  (CNNs)  in  digital image steganography. This work extends an existing implementation that used a two-dimensional CNN to perform the preparation, hiding, and extraction phases of the steganography process. The methodology proposed in this research, however, introduced changes into the structure of the CNN and   used   a   gain   function   based   on   several   image   similarity   metrics   to   maximize   the imperceptibility between a cover and steganographic image.</p> <p>The  performance  of  the  proposed  method  was  measuredusing  some  frequently  utilized image metrics such as structured similarity index measurement (SSIM), mean square error (MSE), and  peak  signal  to  noise  ratio  (PSNR).  The  results  showed  that  the  steganographic  images produced by the proposed methodology areimperceptible to the human eye, while still providing good  recoverability.  Comparingthe  results  of  the  proposed  methodologyto  theresults  of  theoriginalmethodologyrevealed  that  our  proposed  network  greatly  improved  over  the  base methodology in terms of SSIM andcompareswell to existing steganography methods.</p>

The major topic of this dissertation is how to improve the image quality and application of digital watermarking techniques. Most traditional digital watermarking techniques focus on the development and improvement of embedding algorithm. But fewer people really care about the degradation of the image quality of digital watermarking techniques. Meanwhile, the application of digital watermarking technique seems to be limited in the copyright protection. Therefore, the techniques used to improve the image quality and to expand the application of digital watermarking technique are proposed in this dissertation. The concept of the proposed technique used to improve the image quality of digital watermarking technique is different from the concept of traditional watermarking techniques which directly modify the features of image to embed digital watermark. Dissimilarly, the proposed technique uses the relation between the features to embed the digital watermark. In this way, the probability of modifying the feature can be efficiently reduced. Therefore, the image quality of watermarked image can be improved. Besides, the proposed technique works like a pre-process. The proposed technique can be easily applied to present traditional digital watermarking techniques without changing the original embedding and extracting process. The expansion of digital watermarking techniques application consists of two techniques. The first one uses the digital watermarking technique to provide the ¡§Image Refining¡¨ function for digital images. In the proposed technique, the RS (Reed-Solomon) codes of original image are used as a watermark and are embedded in the original image. The watermarking information can provide the error correction capability while the original image is distorted (refining function). If there is some alterations made to the watermarked image and they are within the error correction capability of the RS parities, the alterations will be corrected by the RS decoding to restore the original image. The second one applies the digital watermarking techniques on the Pay-TV system. Conventional key distribution methods used in Pay-TV system emphasize on the method¡¦s refinement in order to reduce the frequency of the keys distributed over network. The proposed method uses the digital watermarking technique to hide the cryptographic keys into routine encrypted video program without any increasing of video program size for the economy of network bandwidth and degradation in perceptual quality. The purpose of this thesis is to improve the practicability of digital watermarking techniques. The proposed techniques could supplement the incompleteness of the traditional digital watermarking techniques with additional functions and applications. In this way, traditional digital watermarking techniques can not only be used for copyright protection but also be the new choice of resolving problem in the future digital environment.