HIGHER-DENSITY DUAL TREE DISCRETE WAVELET TRANSFORM

摘要

This paper introduces the higher-density dual-tree (HDDT) discrete wavelet transform (DWT). A new MRA is introduced to describe higher-density DWT and used to obtain the sufficient condition for the HDDT Hilbert transform pair. In designing HDDT filters we use the extended common-factor method which not only includes the common-factor method but also provides exact linear phase bi-frame filters. Both HDDT tight frame and bi-frame (anti)symmetric filter design methods are given. At last, the results of denoising experiments by our newly designed HDDT filters in this paper prove the effectiveness of higher-density dual-tree DWT in the application of image processing.