AN IMPROVEMENT OF DIAMOND ENCODING FOR PROTECTING STEGANALYSIS OF SPAM DATA MINING

摘要

A large amount of information is transmitted on the Internet, some don't want to be seen by others, how is data-hiding in an image as not to be seen by the eye or pass through steganalyzer tools, for example, Subtractive Pixel Adjacency Matrix (SPAM) detect possible? This has become an important problem. Chao et al. proposed A Novel Image Data Hiding Scheme with Diamond Encoding way of data hiding, shortened as DE, DE has high capacity and whether information is hidden is not easily seen by the eye. As steganalysis's development matures, DE is unable to pass the steganalysis technology, SPAM. As a result, DE has lost its data hiding use, so in this research, we discuss how to improve DE's camouflage ability. The way DE embeds is from left to right, from up to down, putting two elements in one group and changing these elements to embed information, this is also easier to embed information into continuous areas, and thus, easily detected by SPAM. As a result, this research proposed a way to disperse information randomly and hiding the information in a dispersed way in the image to strengthen the ability of camouflage. The results of this experiment show that by using what we proposed, it can effectively lower the possibility of being detected.