Blind steganalysis method based on symmetric neighborhood information

Abstract

Disclosed is a spatial-domain characteristic-based digital image steganalysis method. The method comprises the following steps: for category information-marked digital images in a training set, calculating pixel grayscale difference value information on multiple directions in symmetrically adjacent domains of the digital images; processing and rotational invariance coding the pixel grayscale difference value information; then calculating a histogram of the coded information, normalizing the histogram to serve as characteristics; training and classifying the characteristics extracted, acquiring a classifier model; for an arbitrary digital image inputted, calculating adjacent domain information thereof to extract the characteristics; inputting the extracted characteristics into the classifier model, and acquiring classification information of the image inputted. The present method obviates the need for analysis of the characteristics of any specific steganography and is a universal method, the method therefore can achieve broad application in common (blind) steganalysis of digital images.

Description

technical field [0001] The invention relates to the fields of image forensics and network information security, in particular to a blind steganographic analysis method based on symmetrical neighborhood information, which is used for detecting information hiding of digital images using steganography. Background technique [0002] The 21st century is an era of digitalization. With the rapid development of computer and multimedia technology, people are increasingly inseparable from digital media. Digital media is the most important avenue in today's society. On the Internet, information is disseminated between individuals, between individuals and groups, and between groups through digital media. In addition to the information carried by themselves, digital media can also be used to hide secret information, thereby relying on the public channels to spread bad information, or to engage in covert communications. Because information hiding technology can be used by criminals as a...

AI Technical Summary

Problems solved by technology

According to the required conditions for use, there are currently two methods for detecting hidden information in images, blind detection and non-blind detection. The non-blind detection method uses the characteristics of a specific information hiding algorithm to detect a specific steganographic method. Therefore, it can be used for specific algorithms, but not for other steganographic methods other than this steganographic method

Images