Researches On Self-Located And Recoverable Authentication Methods For Digital Images

With the popularization of the Internet and the widely application of multimedia data, the authenticity and integrity of information media are more and more concerned. Therefore content-based multimedia data authentication technique has become a research focus, within which image authentication is important and crucial.The basic objective of image authentication is to verify the authenticity of image information, to accept processes of fidelity preservation for content and reject irregular manipulations of damage for image content. Digital signature and watermarking are two main methods of image authentication. While traditional methods ignore the redundancy of image data, which cause its authenticity requirement so strict that image can't have any change. On the other hand, most of the existing digital signature and digital watermarking methods are not able to locate accurately for the data damage or provide with recovery function, so they are difficult to meet the need of rapid development of information society.Based on the research of existing image authentication methods, we propose a block-pair independent robust authentication signature (BIAS) and an authentication watermarking with self-location and recovery (ALR) aimed at digital signature and watermarking respectively. The two methods remedy a good many defects of traditional digital signature and watermarking methods, accurate at the location of damaged pixels and the recovery of manipulated images, furthermore, they improve the reliability and usability of image authentication greatly. What's more, we also study some key techniques and realization difficulties of image authentication in-depth, including referent block to help tamper localizing, sign bit adjustment for binary encoding, mean quantization improvement, diagonally extension to help tamper localizing, and discard isolated error point to tolerate incidental distortion. These techniques can be used for good reference to the development of image authentication technique. Finally, we build an experiment prototype system in allusion to the two image authentication methods presented. The system can validate the performance process of BIAS and ALR methods and can evaluate their reliability, usability and performance. Plentiful experiments show that our experiment prototype system for image authentication not only rejects illegal manipulations, while accepts legal content-based operations, but also can exactly detect the location of altered image areas and approximately recover the discarded data modified by some malicious manipulations. In conclusion, our digital image authentication system is valuable and worth popularizing...