Research On Visual Hidding Algorithm Based On Image Texture Features

Human visual system is not insensitive to partial or subtle changes in image, and acertain degree of information redundancy is exist in the image itself. Based on that, thetechnology of image information hiding embeds the secret information in public stegoimages through some effective algorithm. Steganography, watermark and visualcryptography can achieve image information hiding effectively. In the existing researchon steganography and watermark, complex algorithm is required to decrypt, whichcause high computational complexity, so, the technology of steganography andwatermark exist limitation in some specialized applications. Information recovering invisual cryptography accomplishs the decryption just need superimposing the stegoimages simply without computer, but the visual quality of recovered image is low. Forthese problems, this thesis proprose hope propose a visual algorithm, which has theconvenience of recovering secret information in visual cryptography, and ensure thequality of recovered image while making the recovering process simple. The mainresearch as follows:1. By summarizing and analysing the existing technology of image informationhiding, two reciprocal images is used as stego images, which can mask the redundantinformation completely. To ensure the stego images dissemination, a research idea isproposed, which targeted embed the secret information by dividing strong and weaktexture region of the cover image through the image information entropy and variance,and improve the visual quality of stego image.2. Proposes a visual hiding algorithm for binary image based on image texturefeatures, which make a modifying rule of the pixel to be processed according to humanvisual system: assign different weights to horizontal, vertical, or diagonal processedpixel and unprocessed pixel in connected region, and restrict the changes to stego imagethrough setting a suitable modifying threshold, we can get the stego images with highvisual quality. Then the feasibility of this visual hiding algorithm is proved byexperiment result.3. Based on the grayscale image superposition features, we extend the algorithm tothe gray images, which lead to greatly improvements in visual quality of stego images and recovered images. Then the validity of this visual hiding algorithm is proved byalgorithm analysing and experiment.