Trends of Tree-Based, Set-Partitioning Compression Techniques in Still and Moving Image Systems
William A. Pearlman · 2001
In addition to high compression efficiency, future still and moving image coding systems will require many other features. They include fidelity and resolution scalability, region of interest enhancement, random access decoding, resilience to errors due to channel noise or packet loss, fast encoding and/or decoding speed, and low computational and hardware complexity. Moreover, there are emerging new venues for the application of image compression techniques to data associated with points of a two or three-dimensional grid that will demand these features and perhaps others. We shall discuss these new venues along with the usual ones and show how tree-based, set-partitioning wavelet coding methods, such as SPIHT (Set Partitioning in Hierarchical Trees) and SPECK (Set Partitioning Embedded bloCK) will fulfill most of the demands of current and future applications. We shall also discuss the emerging JPEG-2000 in this framework.