Abstract
In this paper, a novel detection and recovery system is proposed for image authentication based on two compression schemes: block truncation coding and vector quantization. In order to reduce the overhead in the communication, the sender first compresses an image and then refines the characteristic values from the compressed image. The characteristic values are encrypted to digital signatures using the RSA scheme (proposed by Rivest, Shamir and Adleman). The signatures are embedded into the compressed image and thus the image integrity can be authenticated after transmission. Upon receipt of the compressed image, the receiver extracts the signatures and then decrypts the characteristic values using the sender's public key. Another characteristic value can also be decoded by the image compression scheme. By comparing a few characteristic values, the system determines whether or not the decompressed image is tampered and then outputs a correct characteristic value to rebuild the tampered regions. Furthermore, the proposed scheme authenticates the integrity of the image to the receiver and effectively recovers the tampered image.