Coverless Image Steganography: A Survey

With the digitalization of information, a lot of multimedia data are under attack, information security has become a key issue of public concern. Image steganography, aiming at using cover images to convey secret information has become one of the most challenge and important subjects in the field of information security recently. Different from the traditional image steganography, coverless image steganography does not need to employ the designated cover image for embedding the secret data but directly transfers secret information through its own properties such as pixel brightness value, color, texture, edge, contour and high-level semantics. Therefore, it radically resist the detection of steganalysis tools and significantly improves the security of the image. Its basic idea is to analyze the attributes of the image and map them to the secret information according to certain rules based on the characteristics of the attributes. This paper includes more than 50 key contributions to provide a comprehensive survey in this field, covers the main aspects of coverless image steganography research: the fundamental frameworks, preprocessing, feature extraction, generation of hash sequence and mapping relationships. The existing methods are evaluated and the prospect of future work is also summarized.