Parallel data compression

Develop algorithms for compressing still images with experimental results

Research opportunity

Abstract: Parallel omni dimensional networks can improve compression of high-resolution still images when compared with the traditional Ziv Lempel Welch (LZW) codes. Compression can be either “lossless” according to the Shannon information theory, in which each bit is precisely reproduced, or “visually lossless” according to the Autosophy information theory, in which resolution is determined by the perception limits of the receiver. The purpose of this research would be to implement the algorithms in software for testing the algorithms and to gather performance data through experimentation. The algorithms may be improved by new innovation if possible. This research should include the following steps:

Algorithm development: Develop software algorithms for: pyramidal addressing, library growing, image encoding, and image retrieval. Gather a collection of test images, such as the eight standard images of the Waterloo Repertoire Color-set available at http://links.uwaterloo.ca/colorset.base.html.

Library generation: Grow a hyperspace library from the input images using pyramidal addressing and a parallel hyperspace library formats. Use a bubble algorithm to select the most common image patterns. Use the library saturation curve to determine the optimal practical library size (64k nodes?). Generate a standard hyperspace library for experimentation.

Performance testing: Test the compression performance on various images loaded via the Internet. Compare the results with other compression methods such as the lossless gif and tif standard and the lossy JPEG standard. Improve the algorithms if possible. Discuss the advantages and disadvantages of each image compression method.

Applications: The new image compression method could be licensed to commercial software companies to replace the gif, tif, pdf, and JPEG standards on the Internet. The algorithms could be implemented in cellular telephones or credit card sized digital cameras.

Keywords: Autosophy, Lossless image compression.

Available downloadable documents:

Publication 1996 – Tutorial – Webpage htm

Publication 1995 – Tutorial – Webpage htm

Publication 1993 – Conference – Webpage htm

Publication 1993 – Image compression – Webpage htm

Patent 1992 – Pyramidal addressing – Webpage htm

Publication 1991 – Convention – Webpage htm

Patent 1991 – Image compression – Webpage htm

Publication 1990 – IEEE Magazine – Webpage htm