Noise tolerant image authentication with error localization and correction

Images transmitted over insecure communication channels are protected using a message authentication code (MAC) for authenticity and integrity. Conventional MAC algorithms are in particular not suitable for image authentication over a noisy channel. Even a single bit error will render the image unauthentic and therefore being rejected. Retransmissions might not be possible over a uni-directional channel or in case of real-time transmissions where accepting a part of image might be better than having no image at all. In this paper, algorithms for image authentication, error localization in the image and their correction are proposed. Once the errors in the image are localized to a smaller area, it is shown how the errors can be corrected and if there are still some minor errors left, the soft authentication algorithm will help in partial image acceptance. Simulation results show how the algorithm reacts to bit and block errors and how the error correction and soft authentication has little or no visual impact on the resultant image quality.