University of Delaware - College of Engineering
ELECTRICAL & COMPUTER ENGINEERING

Research

Communication & Signal Processing

Lossless Data Compression

Charles Boncelet

A long term research interest is in lossless data compression, both theoretically and in specific applications, such as lossless image compression.

We have developed new methods for entropy compression (block arithmetic coding) and several new methods for image compression. In recent image compression work, we have invented the Med-Bzip2 and the BCTW methods.

Med-Bzip2 is meant as a simple, but general purpose, lossless image compressor. It is very easy to program, yet achieves compression ratios about 10% better than JPEG-LS (the best standard for lossless image compression).

BCTW is a bitplane by bitplane compression method that achieves compression ratios about 15% better than JPEG-LS (on mixed suites of images, but not necessarily on each image). This performance is state of the art. BCTW has recently been used as a core technology in our image stegnalaysis work.

Recent publications

"On the Use of Context Weighting in Lossless Bilevel Image Compression," S. Xiao and C. G. Boncelet Jr., IEEE Trans. on Image Processing, November 2006.

"Lossless Image Compression with BCTW," C. G. Boncelet Jr., Proceedings of the IEEE International Conference on Image Processing, October 2006.

"Simple, High Performance Lossless Image Compression," C. G. Boncelet Jr., Proc. IEEE ICIP Conference, Thessaloniki, Greece, October 2001.


Bookmark and Share