A Burrows-Wheeler transform-based compression algorithm for portable network graphics files

In this thesis we propose a different compression algorithm for use in Portable Network Graphics image files. The proposed algorithm is based on the Burrows-Wheeler Transform, with additional stages of a move-to-front coding step, two run-length encoding steps, and an arithmetic coding step.; We briefly explain data compression in general and the importance of good data models. Then, we review the basics of the involved algorithms, pointing out specific implementation details—for Portable Network Graphics and for our proposed algorithm. We experimentally study compression ratios by implementing a prototype and comparing compressed image file sizes. Our results indicate significant improvements for computer-generated graphics images, whereas the improvement is only marginal for photographic images. Finally, we conclude with a brief discussion of a possible application of the proposed algorithm in the Portable Network Graphics file format.