For many applications in signal processing and numerical analysis, it
is important to use periodic scaling functions and wavelets. The aim o
f this paper is a constructive approach to periodic spline wavelets an
d to related decomposition and reconstruction algorithms. We apply the
periodization of the semiorthogonal Chui-Wang wavelets and the well-k
nown Euler-Frobenius functions. Using a new approach to the decomposit
ion relations via two-scale symbol (2,2)-matrices, we obtain new and e
fficient decomposition and reconstruction algorithms which are mainly
based on the fast Fourier transform technique. The presented algorithm
s can be used for the decomposition and reconstruction of L(2)(R)-func
tions, too. Finally, our decomposition algorithm is used to analyze th
e local regularity of periodic functions. (C) 1995 Academic Press, Inc
.