Degree Name

Master of Science (Hons.)


Department of Computer Science


Image processing is of growing interest in computer applications. Since the size of image files are generally large and image data is usually highly redundant, using compression algorithms provide effective way of increasing communication and storages efficiency. JPEG, which is one of the most widely used image compression standards, uses Huffman or arithmetic coding for its lossless compression part. With rapid growth of the Internet, controlling access to data is of increasing importance and hence encryption is of much wider use. Adding security to the algorithm is an attractive proposal as it could reduce the overall processing cost of providing secure compressed data. A number of methods for combining encryption and compression had been proposed and various attacks on these proposals were published. This thesis reviews the known proposals for arithmetic coding encryption schemes and examines proposed attacks on such systems. We extend the attacks, introduce new ones, and finally propose efficient methods of enhancing security of these systems.