We present a wavelet packet image coding algorithm combined with the actual coding costs. It is composed of five parts: (1) wavelet packet decomposition and best basis selection based on a new cost function, which is designed to be combined with the actual coding costs so as to select a sequential coder-friendly wavelet packet basis; (2) uniform quantization for all the wavelet packet coefficients; (3) a quadtree classifying procedure, used to code the significance map as well as to classify the wavelet packet coefficients; (4) the universal trellis coded quantization (UTCQ), used to code the classified coefficients; and (5) the entropy coder, used to code the indices of the universal trellis coded quantizer's output, the significance map coding output, and the wavelet packet decomposition structure. The image coding results, calculated from actual file sizes and images reconstructed by the decoding algorithm are either comparable to or surpass previous results obtained through much more sophisticated and computationally complex methods for texture-rich images, such as "Barbara" and fingerprints.

Date Published: 1 June 2003
PDF: 9 pages
Opt. Eng. 42(6) doi: 10.1117/1.1572497
Published in: Optical Engineering Volume 42, Issue 6