Content based general steganography analysis method of spatial-domain image

Abstract

The invention provides a content based general steganography analysis method of a spatial-domain image, belongs to the technical field of information safety, and aims at describing the content complexity of the image reasonably, improving features of steganography detection effectively, and improving the accuracy of spatial-domain steganography analysis substantially on the premise that low operand is ensured. The method comprises the steps that 1) a characteristic vector of image content classification is obtained; 2) a K-means clustering algorithm is used divide a training sample into non-intersected classes; 3) a second-order joint probability density matrix in the corresponding vertical direction is calculated and serves as a characteristic vector of steganography detection; 4) classifier models of different types are obtained; 5) an image to be detected is given, the classification characteristic vector is calculated according to the step 1), the image to-be-detected is attributed to a corresponding class according to Euclidean distances from the image to different clustering centers, and characteristics of steganography detection are extracted according to the step 3); and 6) whether the image includes hidden information is determined. The steganography analysis method is mainly applied to image processing.

Description

technical field [0001] The invention belongs to the technical field of information security, and in particular relates to a content-based universal steganalysis method for air domain images. Background technique [0002] The purpose of steganalysis is to detect the existence of hidden information in multimedia data. Digital image has become one of the main carriers of steganography due to its high redundancy, easy access and convenient storage, and the corresponding steganalysis has also become a research hotspot. Compared with dedicated steganalysis, general steganalysis does not require prior knowledge of steganographic algorithm details, and with the improvement of the effectiveness of extracted features and the performance of classifiers, its detection accuracy gradually improves, even for some unknown Steganographic algorithms can also achieve better detection results, so general steganalysis is more in line with actual requirements, has a wider range of applications, ...

AI Technical Summary

Problems solved by technology

[0005] In order to overcome the deficiencies of the prior art, the present invention aims to propose a content-based general steganalysis of spatial domain images for the traditional steganalysis method that does not consider the statistical characteristics of the image itself, the feature dimension is high, and the amount of calculation is large. method, through the reasonable description of image content complexity and the effective improvement of steganographic detection features, the accuracy of spatial steganalysis is significantly improved under the premise of ensuring low computation

Images