Image authentication and tamper localization based on relative difference between DCT coefficient and its estimated value

摘要

Digital images are increasingly transmitted over non-secure channels such as Internet, therefore image authentication techniques have recently gained great attention due to their importance for a large number of multimedia applications. To protect the authenticity of images, several approaches have been proposed. These approaches include conventional cryptography, semi-fragile watermarking and digital signatures. In this paper, we propose two techniques of the same type based on what we call characteristic data digest. Both techniques can blindly detect and localize malicious tampering, while maintaining reasonable tolerance to conventional content-preserving manipulations. The characteristic data is derived from the relative difference between each pair of selected DCT coefficient, AC for one technique and DC for another technique, in a central block and its counterpart estimated by the center block and its adjacent blocks. In order to maintain the relative difference relationship when the image undergoes legitimate processing, we make a pre-compensation for the coefficients. Experimental results show that our techniques are significantly superior to semi-fragile techniques under the condition of the same image fidelity, especially in tolerance range of legitimate processing, and/or the ability to detect and localize the tampered area. Due to the simplicity of the algorithms, our techniques can be used in video frame authentication, and even other digital media. In addition, this kind of proposed techniques can be extended to use other characteristic data, such as high-level moment, statistical data of images, and so on.