2020/06/06 by Nibraas Khan, Khan, Nibraas, Ruj Haan +7
Computer Science · #Advanced Steganography and Watermarking Techniques #Cryptography and Security (cs.CR) #Digital Media Forensic Detection #FOS: Computer and information sciences #Generative Adversarial Networks and Image Synthesis
paper · pdf · doi:10.48550/arxiv.2006.04008
openalex publication_date 2020/06/06 · openalex created_date 2022/09/14 · openalex updated_date 2026/07/28
For as long as humans have participated in the act of communication,\nconcealing information in those communicative mediums has manifested into an\nart of its own. Crytographic messages, through written language or images, are\na means of concealment, usually reserved for highly sensitive or compromising\ninformation. Specifically, the field of Cryptography is the construction and\nanalysis of protocols that prevent third parties from understanding private\nmessages. Steganography is related to Cryptography in that the goal is to\nobscure information using some method or algorithm, but the most important\ndifference is that the information and the method of concealing information\nwithin Steganography both involve images--more precisely, the embedding of one\nimage or piece of information into another image. Ever since the creation of\ncovert communication methods, steps have been taken to crack cryptography and\nsteganography algorithms. The desire for this rises from both human curiosity\nand the need to counteract adverse uses, such as encoding harmful media in\ninconspicuous media (phishing attack). In this paper, we succeed in cracking\nthe Least Significant Bit (LSB) steganography algorithm using Cycle Generative\nAdversarial Networks (CycleGANs) and Bayesian Optimization and compare the use\nof CycleGANs against Convolutional Autoencoders. The results of our experiments\nhighlight the promising nature of CycleGANs in cracking steganography and open\nseveral possible avenues of research.\n