Subband Image Compression

Aria Nosratinia, Geoffrey M. Davis, Zixiang Xiong, Rajesh Rajagopalan · Kluwer Academic Publishers eBooks · 2006

: This chapter presents an overview of subband/wavelet image compression. Shannon showed that optimality in compression can only be achieved with Vector Quantizers (VQ). There are practical difficulties associated with VQ, however, that motivate transform coding. In particular, reverse waterfilling arguments motivate subband coding. Using a simplified model of a subband coder, we explore key design issues. The role of smoothness and compact support of the basis elements in compression performance are addressed. We then look at the evolution of practical subband image coders. We present the rudiments of three generations of subband coders, which have introduced increasing degrees of sophistication and performance in image compression: The first generation attracted much attention and interest by introducing the zerotree concept. The second generation used adaptive space-frequency and rate-distortion optimized techniques. These first two generations focused largely on inter-band dependen...

Read the paper · More papers on PaperTik