Abstract: This paper attempts to propose a novel technique of blind authentication based on the method of secret in addition to data repair capability for grayscale document images through the use of the Portable Network Graphics (PNG) image. For every block of a grayscale document image, an authentication signal is generated, which, along with the block content in binary, is transformed into numerous shares using the Shamir secret sharing scheme. The parameters involved are carefully selected so that as many shares as possible can be generated and embedded into an alpha channel plane. After this, the alpha channel plane is combined with the original grayscale image to yield a PNG image. During this process, the computed share values are recorded as a range of alpha channel values near their maximum value of 255 to return a transparent stego-image with a disguised effect. In the image authentication process, marking of an image block is done as tampered, if the authentication signal computed from the current block content does not match the one extracted from the shares embedded in the alpha channel plane. Each tampered block is then subjected to data repairing by a reverse Shamir scheme after collecting two shares from unmarked blocks. Procedures to protect the safety of the data that lies concealed in the alpha channel have been proposed. Decent experimental results demonstrate the efficiency of the proposed method.

Keywords: Data hiding, data repair, grayscale document image, image authentication, Portable Network Graphics (PNG) image, secret sharing.