Rich Model steganography detection feature selection method based on feature component correlation

Abstract

The invention provides a Rich Model steganography detection feature selection method based on feature component correlation. The method comprises the following steps of: 1, disassembling a high-dimensional Rich Model steganography detection feature into a plurality of Rich Model sub-models; 2, for each Rich Model sub-model, measuring the separability of each feature component of the Rich Model sub-model, and sorting the feature components in a descending order according to the measured value of the separability; 3, for each Rich Model sub-model, calculating the correlation between any two feature components of the Rich Model sub-model, and performing feature selection on the feature components according to the strength of the correlation; and 4, combining the Rich Model sub-models subjected to the feature selection so as to obtain a final steganography detection feature. When the method is applied to the Rich Model features of the frequency domain and the spatial domain, the dimension of the Rich Model features can be effectively reduced under the condition that the steganography detection accuracy is not affected, and the effect on the frequency domain is more remarkable.

Description

technical field [0001] The invention relates to the technical field of steganography detection, in particular to a feature selection method for Rich Model steganography detection based on feature component correlation. Background technique [0002] Digital steganography is a technology that embeds information in the redundancy of media such as digital images, audio, video, and text to achieve the purpose of covert communication. With the introduction of the HUGO steganography algorithm in 2010, adaptive steganography based on the framework of "distortion function design + STC coding" has become the mainstream of image steganography. Based on this framework, researchers have successively proposed a series of highly resistant Adaptive steganographic algorithms for detection performance. These algorithms make most of the traditional steganographic detection algorithms invalid. In 2012, Fridrich et al. proposed the Rich Model feature (reference 1 "Fridrich J, Kodovsky J.Rich M...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem that the traditional steganographic detection feature dimensionality reduction method has limited applicability and is difficult to effectively reduce the dimensionality of multiple feature components with strong correlation and redundancy, the present invention provides a RichModel hidden feature based on feature component correlation. The writing detection feature selection method, based on the correlation between feature components and the structure of Rich Model steganographic detection features, can effectively reduce the dimensionality of multiple feature components with strong correlation and redundancy

Images