Research on imperceptibility index of image Information Hiding

The deficiency of imperceptibility indexes of information hiding algorithm has been analyzed. According to human vision characteristics, a new imperceptibility evaluation index with the original carrier image for comparing. This index comprehensively considers the background illumination of carrier image, the complexity of background texture, the spatial position and the directivity influence on vision masking effect. Experiments prove that this index is better than Power Signal-to-Noise Ratio (PSNR) in measuring the imperceptibility of hidden information. It can evaluate gray image and binary image, this solves the problem that gray image and binary image can not use the same index. Besides, a special situation of this index is just PSNR.