Portable Network Graphics Approach to the Authentication of Halftone Images with Henon Map Encryption

Abstract

ABSTRACT A new method of blind image authentication based on secret sharing and data repair capability via the use of Portable Network Graphics (PNG) is proposed. The method manipulates the properties of alpha plane to embed the authentication signals as well as the data required for repairing. Care must be taken to maintain the transparency property. The image is halftoned and the authentication signals are generated from the halftoned image. Shamir secret sharing technique is used for the creation of data for repairing. The parameters are selected such that as many shares as possible should be generated for good quality reconstruction. The new scheme guarantees, no matter what type of attack, at least some shares must survive it and hence the actual data can be reconstructed. The authentication is done at the block level, but repairing is done at the pixel level. The embedded authentication signals are extracted from the alpha plane and compared with that computed from the received image. If there is a mismatch, then that block is marked as tampered. Then, repairing is done to the tampered blocks by reverse Shamir algorithm. To enhance the security of the grayscale document images, an encryption based on chaos theory is incorporated.