A feature-based classification technique for blind image steganalysis

摘要: In contrast to steganography, steganalysis is focused on detecting (the main goal of this research), tracking, extracting, and modifying secret messages transmitted through a covert channel. paper, feature classification technique, based the analysis two statistical properties in spatial DCT domains, proposed blindly (i.e., without knowledge steganographic schemes) determine existence hidden an image. To be effective class separation, nonlinear neural classifier was adopted. For evaluation, database composed 2088 plain stego images (generated by using six different embedding established. Based database, extensive experiments were conducted prove feasibility diversity our system. It found that system consists of: 1) 90%/sup +/ positive-detection rate; 2) not limited detection particular scheme; 3) capable with rate as low 0.01 bpp; 4) considering test incurred low-pass filtering, sharpening, JPEG compression.