Study on Image Steganalysis and Its Evaluation Method

Steganography as an important technology in the information security , with the vigorous development of the Internet and communication technologies caused the majority of scholars and relevant departments attach great importance , for steganography means of covert communication , anti- steganography technology - steganography analysis of surgery has also become a hot research topic in recent years, the field of information security , proposed steganalysis algorithm different characteristics , however , difficult to select and trade-offs in practical applications , so the effective evaluation of steganalysis algorithm also become an urgent problem to be solved . The basis of the analysis of algorithms written in -depth study of implicit theories and related technologies , already at this stage the majority of universal steganalysis analytical methods to determine the correct rate is not high shortcomings , propose a new JPEG image universal steganalysis method . The DCT coefficients of JPEG image decomposition to the wavelet domain , a series of high - level statistical characteristics of each sub-band wavelet coefficients calculated , and the introduction of the feature scaling technology , more sensitive to the subtle differences in the characteristics scaled , and finally with support vector machines to complete a reliable detection of the image . The experimental results show that this method improve the detection accuracy rate of the algorithm is an effective steganalysis method . Written analysis and evaluation methods for this stage of the majority implicit computational complexity , weight distribution is not quite reasonable phenomenon , this article by Grey Relevance Theory and Entropy combined weighted gray correlation analysis indicators calculated to be evaluated vector with optimal weighted gray correlation degree between the indicator vector steganalysis algorithm to solve the reliability evaluation . Instance simulation proved that this method can reasonably evaluate the pros and cons of various types of steganalysis algorithm is a viable evaluation method . The use of weighted gray relational analysis , wavelet higher order statistics for improved steganalysis algorithm to evaluate the results of the evaluation again proved the superiority of the improved algorithm .