|Articles and journals | Tariffs | Payments | Your profile|
Increasing the efficiency of steganoanalysis in the area of discrete wavelet image transformation by analyzing the parameters of the frequency domain of the image
Abstract.The object of the study are the methods of stegan analysis in the area of discrete wavelet transformation of an image. The author investigate the influence of the fact of embedding in the region of a concrete wavelet transformation on the values of the coefficients of the regions of discrete cosine transform and image in order to improve the efficiency of detecting the fact of embedding into the discrete wavelet transformation domain. The influence of the fact of embedding in the region of discrete wavelet transformation on certain coefficients of regions of discretely cosine transform and discrete sine transformation of the image is shown. The author proposed to use certain coefficients to improve the quality of training of the support vector machine. Method of research: to assess the effectiveness of the steganoanalysis method proposed in the article using the proposed coefficients, a comparison of the efficiency of image classification with other popular steganoanalysis methods for the wavelet decomposition region is performed. As a steganographic influence, the values of the least significant bits of the coefficients of the discrete wavelet transform are used. Main results of the study is the possibility of using certain coefficients of discrete cosine transform and discretely sinus transformation of the region with the purpose of steganoanalysis in the region of discrete wavelet transformation is shown. According to the results of the study, an original method of steganoanalysis is proposed, which makes it possible to increase the efficiency of steganoanalysis for the LH and HL regions of the discrete wavelet transformation of the image. The obtained results can be used in the development of steganoanalysis systems to provide an effective detection of the fact of embedding into the discrete wavelet transformation region of an image.
Keywords: support vector machine, machine learning, discrete sine transform, discrete cosine transform, discrete wavelet transform, frequency domain, steganalysis, steganography, binary classification, wavelet domain
Article was received:26-02-2018
This article written in Russian. You can find full text of article in Russian here .