Application Of Visual Characteristics Described By Information Theory In Information Hiding

Recently, various kinds of visual models appear, including the visual masking effect,visual saliency,visually significant. It is of great scientific significance for us to research these.In this thesis, the image which is a special information system apply the concept of information entropy of information theory. Introduced the concept of weighted entropy,using the visual contrast sensitivity function to determine weights in the definition of weighted entropy, put weighted entropy into practical application. Combining with character that the weighted entropy of image sub-block change with the size of image sub-block. Will the Saliency figure as a function of weighted entropy of image sub-block and size of image sub-block.We defined weighted entropy matrix basis on weighted entropy. The image similarity measurement method was determined by eigenvalue of weighted entropy matrix. We defined result as weighted entropy difference. Verify the feasibility of weighted entropy difference by contrasting with Watson distance and SSIM.Appling weighted entropy Saliency model in digital watermark, a new watermarking algorithm was proposed. We measured similarity of the original image and watermarking extraction by weighted entropy difference.