Degree Name

Doctor of Philosophy


Department of Computer Science - Faculty of Informatics


Advanced digital technologies have made multimedia data widely available. As multimedia applications become common in practice, security of multimedia data has become main concern. Digital images are widely used in various applications, that include military, legal and medical systems and these applications need to control access to images and provide the means to verify integrity of images. Image encryption algorithms protect data against unauthorized access. In almost all cases image data is compressed before it is stored or transmitted because of the enormity of multimedia data and their high level of redundancy. Compressing plaintext before applying the encryption algorithm effectively increases security of the overall system. However direct application of encryption algorithms to image data i) requires high computational power and ii) introduces delay in real-time communication. If a data compression algorithm can be made to also provide security, less processing overhead could be expected as a single algorithm achieves two goals. Image authentication provides the means to verify the genuineness of images. Authentication codes provide a method of ensuring integrity of data. The challenge in image authentication is that in many cases images need to be compressed and so the authentication algorithms need to be compression tolerant. Cryptographic authentication systems are sensitive to bit changes and so are not suitable for image authentication. In this thesis, we study existing image encryption and authentication systems and demonstrate various attacks against these systems. We propose a JPEG encryption system that encrypts only part of the data, and a JPEG2000 encryption system that uses a simple operation, i.e. permutation, and show methods to minimize the computation cost for encryption. We also propose an image authentication system that remains tolerant to changes due to JPEG lossy compression.

02Whole.pdf (3342 kB)