Robust image authentication scheme with self-repair capability for greyscale source document images via PNG format

Recently, Lee and Tsai proposed a new authentication method for greyscale document images using the portable network graphics format image; however, their method cannot resist the self-substitution attack, the same-position-substitution attack, or the cut-off attack. Furthermore, those attacks can be completed by the popular image editing software Adobe Photoshop. Therefore, the authors proposed a security-enhanced authentication scheme based on Lee and Tsai's method. The authors’ proposed scheme uses three random binary sequences to randomise the binary version of a given greyscale document image, and thus overcomes the security flaws mentioned above. The authors’ proposed scheme is capable of repairing the content of the given stego-image if attacked by the methods mentioned above. Moreover, the authors’ proposed scheme retains all of the strengths of Lee and Tsai's scheme. The authors improve the opacity of the alpha channel of stego-image by using Wang and Su's extended secret sharing, and enhance the data security by using Hash functions. The experimental results show that the authors’ proposed scheme is effective for application.