Steganographic image analysis method based on multi-characteristic combination

Abstract

The invention provides a steganographic image analysis method based on multi-characteristic combination. The analysis method comprises the steps as follows: a user inputs a target image (steganographic image); multiple characteristics are extracted from the input target image, and a characteristic set is established; the extracted characteristic set is subjected to complementation, fusion and redundancy processing; then, dimension reduction process is performed; finally, decoding analysis is performed, and an analyzed result is processed. The method has high universality and practicality, and finally, an experiment carried out in accordance with an algorithm of the embodiment shows that the method has high accuracy and anti-interference capacity, and one effective method is provided for accurate detection and analysis of the steganographic image.

Description

technical field [0001] The invention relates to the field of computer technology, in particular to a steganographic image analysis method based on multi-feature union. Background technique [0002] With the rapid development of digital image processing technology, image steganography technology based on digital images has become a research hotspot in the field of information security, and how to accurately detect and analyze steganographic images has also become the focus of extensive attention of researchers. Commonly used steganographic image analysis methods mainly include SRM (Spatial Rich Model) and PSRM (projections spatial rich model). However, these two types of algorithms are low-dimensional general-purpose detection algorithms, which cannot realize the analysis of high-dimensional features. In order to improve the analysis of the high-dimensional features of the steganographic image, the maxSRM (max Spatial Rich Model) method is used to extract the embedded probab...

AI Technical Summary

Problems solved by technology

However, these two types of algorithms are low-dimensional general-purpose detection algorithms, which cannot realize the analysis of high-dimensional features.

In order to improve the analysis of the high-dimensional features of the steganographic image, the maxSRM (max Spatial Rich Model) method is used to extract the embedded probability of the image co-occurrence matrix, and the maximum value of the probability is used to replace the unit dimension in the original matrix, which effectively improves the high-dimensional features. The analysis accuracy is high, but the noise sensitivity of the extraction of the probability extreme value of this method is too strong; the authors Zhou Jie and Zhang Minqing wrote in the literature "Image Steganalysis Based on Dual Feature Selection for Dimensionality Reduction" (Computer Engineering and Design, 2016, 37 (11) : 2917-2922) adopted the method of image sub-block segmentation and proposed a dual feature analysis method, which effectively overcomes the problem of image noise sensitivity by taking the result of weighted fusion. However, this method is mainly for steganography The change of image statistical characteristics is used to extract and analyze features. In the process of feature extraction, a synchronous method is used, which ignores the statistical properties of the image itself. As a result, the detection performance of this type of algorithm is heavily dependent on the image used for steganography and the implementation of steganography. method

Images