A Study on Countermeasures against Steganography: an Active Warden Approach

Digital steganography is a method used for hiding information in digital images. It can be used for secure communication. There have been many robust digital steganography methods invented in recent decades. The steganographic message can be inserted in multimedia cover signal such as audio, image and video. However, this technique also may be used by malicious users to transmit dangerous information through the Internet beyond the control of security agencies. How to detect and/or block potentially dangerous information transmission on the Internet through billions of multimedia files while not affecting innocent multimedia communications becomes a challenging problem. Existing steganalysis methods or steganography attacking methods which are mostly passive methods cannot be used for analyzing a large volume of digital images in a short time. In addition those passive methods also cannot be generic enough to defeat various steganographic algorithms on the Internet. In this paper, we propose an active attacking model to defeat the rising threat of steganography. The active protection mechanism is proved to be more effective to protect the integrity of the multimedia data. Based on the active attacking model, a steganography attacking method which is not limited by the types of the steganography methods is proposed. The proposed method can process the digital multimedia data to remove the potential dangerous hidden information while keeping the digital data in a high visual quality. This attack method is based on a proposed transform called Discrete Spring Transform. Some implementations of the Discrete Spring Transform in audio, image and video signals are proposed. The proposed transform causes that the numerical values of the image to be changed dramatically and then the hidden information is not able to be recovered, while at the same time the visual image quality can be maintained. This method is a generic approach for multimedia signals and contains theoretical advantages over similar methods. Our experimental results have demonstrated that the quality of the multimedia signal can be guaranteed while the stego-data are considerably destroyed. iv To Mingrun, my parents and grandparents. v ACKNOWLEDGMENTS I would like to express my sincere gratitude to my advisor Dr. Dongming Peng and Dr. Yaoqing Yang for the continuous support of my Master study and research, for his patience, motivation, enthusiasm, and immense knowledge. Besides my advisors, I would like to thank the rest of my thesis committee: Dr. and hard questions. Last but not the least, I would like to …