EZW and Wavelet-Based Embedded Image Coding Algorithms: SPIHT, SPECK, and CREW

This paper conducts a comprehensive comparison and discussion of embedded image coding algorithms such as EZW (Embedded Zerotree Wavelet), SPIHT (Set Partitioning in Hierarchical Trees), SPECK (Set Partitioning Embedded Block Coding), and CREW (Compression with Reversible Embedded Wavelets), analyzing their underlying principles and performance characteristics. From an implementation perspective, these algorithms employ progressive bit-plane coding techniques where coefficients are organized into significance maps using tree-based structures (EZW/SPIHT) or block-based partitions (SPECK). The study also introduces other relevant embedded image coding methods and provides detailed explanations of current research directions and future development trends in the field, including approaches for enhancing coding efficiency and image quality through optimized quantization strategies and adaptive entropy coding. Key algorithmic components discussed include zerotree prediction for exploiting spatial redundancy, significance testing mechanisms for bit-plane encoding, and reversible wavelet transforms enabling lossless compression. In summary, this paper offers a thorough exploration of embedded image coding algorithms and technologies, serving as a comprehensive reference for readers seeking deeper understanding of this domain.