You are here

Technical report | Compression of Discrete Probability Maps

Abstract

The first project report for the Efficient Generation and Evolution of Probability Density Maps project is reproduced here as a Defence Science and Technology Group technical report. Here we consider the application and suitability of the Joint Photographic Experts Group (JPEG) compression method and a simpler approach to Discrete Probability Density Maps (DPDMs). The report has been written in such way to be more pedagogical rather than complete.

Executive Summary

The importance of compression techniques for minimising the amount of data to be transmitted and stored in many applications is broadly acknowledged in engineering and computer science. This technical report is focussed on the application and suitability of two widely known image compression techniques to discrete probability density maps, namely Joint Photographic Experts Group (JPEG) compression which utilises several encoding techniques and another which solely implements run-length encoding.

The report contains a brief introduction to redundancy and compression, then summarises the most used portion of the JPEG compression standard. The approach taken to introduce this baseline JPEG method has been pedagogical because of its complexity. Numerical results from the application of JPEG to a very sparse DPDM provided by the DST Group is then presented and discussed. We then consider the application of a simpler method utilising run-length encoding only to the same DPDM. We then conclude and discuss future research directions.

In the appendices we have provided a few additional titbits of information. MATLAB code is stored in Govdex and is available upon request. In particular, a simple one component (black and white) implementation of the most used portion of the JPEG standard is included. Simple MATLAB scripts have also been provided which illustrate various ideas presented throughout the report and how to use the provided JPEG implementation. All figures and values found in tables presented in this report can be obtained using these scripts.

Key information

Author

Daniel L. Whittenbury, Anthony W. Thomas, Ayse Kizilersu and Samuel P. Drake

Publication number

DST-Group-TR-3514

Publication type

Technical report

Publish Date

July 2018

Classification

Unclassified - public release

Keywords

JPEG compression, compression techniques, Discrete Probability Maps