NOVEL CHARACTERISTICS OFEVEN-STACKED COSINE-MODULATED FILTER BANKS:NON-REDUNDANT DIRECTIONAL TRANSFORM AND SHIFT-INVARIANCE

摘要

In this paper, we present non-redundant directional transform andshift-invariant transform based on even-stacked cosine-modulatedfilter banks (ECFBs). M-band dual-tree wavelet transform(MDTWT) recently proposed has rich directional selectivity andnear shift-invariance. However, it has two problems. First, it isnot critically-sampled transform. Second, it cannot achieve shiftinvarianceat all decomposition levels. For these problems, we showa solution based on ECFBs. Critically-sampled ECFBs providenon-redundant transform with rich directional selectivity. Furthermore,oversampled ECFBs with oversampling ratio of 2 can assurethe shift-invariance at any decomposition levels. Moreover, ECFBscan be designed from only one prototype filter and guarantee linearphase. In this paper, we verify these facts by theoretical analysisand simulations.