next up previous
Next: A Wavelet System Up: Wavelet-based Image Compression Previous: Wavelet-based Image Compression

A System for Wavelet-Based Image Compression

Abdou S. Youssef, ACMD

This task involves the design and development of a Matlab system for wavelet-based image compression. The system allows users to specify:

  1. their own wavelet;
  2. their own decomposition tree which specifies how, where, and how many times to apply the wavelet transform;
  3. one of many different quantizers presented in a menu, along with the quantizer parameters;
  4. one of many different entropy (lossless) encoders presented in menu, along with encoder parameters
  5. and the desired compression ratio or bit rate.

The system also provides the user with the ability to reconstruct compressed images, and evaluate their quality/fidelity using any of several quality metrics that will be presented in a pop-up menu.

Because of the bewildering diversity of choices and decisions at the three main stages of compression ( Transform, Quantization, and Entropy Coding), and because of the complex interactions between the three stages, it is critical to have a system that enables the users to painlessly make their own choices in a way that best suits their purposes. The system being developed in this project provides those capabilities. Also, the system enables researchers and developers in the area of data compression to investigate a wide variety of compression approaches, and optimize a large number of tradeoffs. The benefits are expected to be high, especially since image and video compression is already a large industry, and is expected to become considerably larger in various areas such as video on demand and telemedicine. The staggering amounts of data involved in those applications, and the need for fast (and even real-time) transmission of those data on relatively low bandwidth communication links, make data compression imperative.

Accomplishments thus far: The system is fully designed and partially developed. At this point, the user can specify their own wavelet and their own decomposition tree, both for the encoder and the decoder. The remaining features outlined above are still under development.

Future direction of the project: Once the system is complete and operational, much research will be conducted on it to determine optimal wavelets, optimal decomposition trees, optimal quantization techniques, and optimal entropy encoders, as part of any one single compression scheme. Several publications are expected from the research being conducted in this project.

next up previous
Next: A Wavelet System Up: Wavelet-based Image Compression Previous: Wavelet-based Image Compression

Generated by on Mon Aug 19 10:08:42 EDT 1996