Research And Design Of Information Hiding Methodson Binary Image

The binary image is widely used in the area of data fax, character recognition,barcode recognition and digital signatures because of its simple storage and structure.Especially in the fields like military and commercial which have a higher requirementof information security, in order to increase the security of information transmission,the information hiding methods achieved by embedding the secret into the binaryimages become the subjects which are currently focused on.In this paper, we study the binary information hiding technology and focus onthe block-based embedding strategy. Further, Select the design of flippable pixelselection method as the entry point and then build the information hiding algorithmwhich has a better capacity, Imperceptibility and anti-detection ability. The majorcontribution includes:1．Propose a flippable pixel selection method on binary image. According to thevisual characteristics of pixel arrangement in local area, the exclusion methods forthree types of class templates: the "isolated point","unique bridge" and "irregularflow" are proposed. Experiments show that the number of flippable pixels obtained bythis method can be guaranteed and the flipping of them will cause less visual impactof the local image.2．Propose a minimizing number of flip pixels information hiding algorithm onbinary image. By using the proposed flippable pixel selection method and the idea of“embeddable block”, the number of embeddable elements increases. At the same time,the use of Hamming Code reduces the number of flipped pixels. The experimentalresults show that the algorithm can effectively reduce the visual impact and has a strong anti-detection ability.3．Propose a reversible information hiding algorithm on binary image. By apre-statistic of the number of each template type, the algorithm get proper ones ofthem dynamically and forming a sequence to be embedded by compressinginformation and flippable pixel sequence by arithmetic code, in this way, theembedded information can be extract losslessly. Experimental results show that thealgorithm has a wide degree of trial and good imperceptibility.