QUAD TREE-STRUCTURES FOR IMAGE COMPRESSION APPLICATIONS

Traditionally, lossy compression schemes have focused on compressing data at fixed bit rates to either communicate information over limited bandwidth communication channels, or to store information in a fixed-size storage media. In this paper we describe a class of lossy algorithms that is capable of compressing image data over a wide range of rates so that quick browsing of large amounts of information as well as detailed examination of high resolution areas can be achieved by the same compression system. To accomplish this we use a quad tree structure to decompose an image into variable size blocks which are subsequently quantized using a Tree-Structured Vector Quantizer (TSVQ). The developed algorithms utilize variable-size image blocks encoded within quad tree data structures to efficiently encode image areas with different information content. These algorithms are also capable of compressing images so that the loss of information complies with user defined distortion requirements. In this paper we describe the use of quad tree structures in image compression type applications, and we analyze their advantages over the classic vector quantization schemes. Finally, we describe their progressive compression capabilities and we demonstrate that they achieve higher compression/distortion performance compared to the classic TSVQ algorithm.