Abstract
AbstractThis paper proposes a method of optimizing the partition of a two‐dimensional frequency band so that high‐quality reconstructed images are obtained by subband image coding. This partitioning is derived by minimizing the weighted quantization noise power with human visual sensitivity. To simplify the coding system, the structure of the filter bank is analyzed. The band partition parameters which makes the filter bank more efficient are selected, based on experimental results for actual images. The quality of images reconstructed by the proposed method is evaluated by using the relationship between the weighted quantized noise power and their entropy. The proposed method is subjectively evaluated by comparing it with a nonweighted method and the DCT method. © 2003 Wiley Periodicals, Inc. Syst Comp Jpn, 34(3): 96–110, 2003; Published online in Wiley InterScience (www.interscience.wiley.com). DOI 10.1002/scj.1198
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
