No abstract

No abstract

No abstract

In this paper, we introduce and study vector-valued multiresolution analysis with multiplicity r (VMRA) and m-band orthogonal vector-valued multiwavelets which have potential to form a convenient tool for analyzing vector-valued signals. Necessary conditions for orthonormality of vector-valued multiwavelets are presented in terms of filter banks. The existence of m-band vector-valued orthonormal multiwavelets is proved by means of bi-infinite matrix. The relationship between vector-valued multiwavelets and traditional multiwavelets are considered, and it is found that multiwavelets can be derived from row vector of vector-valued multiwavelets. The construction of vector-valued multiwavelets from several scalar-valued wavelets is proposed. Furthermore, we show how to construct vector-valued multiwavelets by using paraunitary multifilter bank, in particular, we give formulations of highpass filters when its corresponding lowpass filters satisfy certain conditions and m = 2. An example is provided to illustrate this algorithm. At last, we present fast vector-valued multiwavelets transform in form of bi-infinite vector.

No abstract