A lossy image codec based on adaptively scanned 
     wavelet difference reduction

Abstract

This paper describes a new method of lossy still image compression,
called Adaptively Scanned Wavelet Difference Reduction (ASWDR). The
ASWDR method produces an embedded bit stream with region of interest
capability. It is a simple generalization of the compression method
developed by Tian and Wells, which they have dubbed Wavelet Difference
Reduction (WDR). While the WDR method employs a fixed ordering of 
the positions of wavelet coefficients, the ASWDR method employs a 
varying order which aims to adapt itself to specific image features.
This image adaptive approach enables ASWDR to outperform WDR in a 
rate-distortion sense, and to essentially match the rate-distortion
performance of the widely used codec, SPIHT, of Said and Pearlman. 
ASWDR compressed images exhibit better perceptual qualities, 
especially at low bit rates, than WDR and SPIHT compressed images. 
ASWDR retains all of the important features of WDR: low complexity, 
region of interest capability, embeddedness, and progressive SNR.