As the pervasive computing age is forthcoming, information intercommunication between persons or groups becomes more and more convenient with the rapid development of network, communication and multimedia technologies. Image, the effective information carrier, has been transmitted over network broadly. According to different applications, private information of users, and important issue of groups may be contained in those images or videos transmitted. How to protect those image contents efficiently, and guarantee their security (privacy and integrity) when they are transferred from one place to another over network frequently? This is a question that becomes one of the most popular research interests in last several years. In this thesis, we try to cope with this question with researching on image content scrambling, biometric features protecting key, and image content authentication, etc. Digital image secure transmission will be handled by focusing on following aspects: Digital image content scrambling comes from TV image applied techniques. With some kinds of mathematic transform, pixels will be distributed randomly in that image. Scrambled image seems chaos, and loses the meaning of original image. Without correct transform, illegal effractor could not get anything from the scrambled image. This is also means that original image content has been protected. We use various knight-tours on the chessboard as the basic tool for image scrambling, and scramble the wavelet coefficients of the transmitted image instead of the image itself. Experimental results show that our algorithm can achieve better scrambling effect. Some useful conclusions are obtained through the analysis and explanation of the experimental data. which lay a solid foundation for further research. However we also find that scrambled image has low ability for compression, and then induces inefficient transmission. We present compression and encryption combined method, involving EZW and AES. It is shown that image content privacy has been obtained as well as high compression ratio. Key management is a decisive factor for security of encryption system. [t is also true for image secure transmission. Biometric features may be a promising mean used to manage key at network environment. There are many successful access control applications in physical situation with biometrics. But as there is inherent contradiction between precise mathematic equality and statistical equality of pattern, it is a challenge for biometric features to be used at network environment. We introduce two schemes in detail, in which key is protected by fingerprint. These two schemes are tested by real fingerprint image. We also get some useful experience from analysis results. The goal of image content authentication is to guarantee the integrity of image received by customer. It needs to distinguish acceptable manipulations (e.g. compression) from ma
修改评论