Discrete Wavelet Filters of Fourier, Cosine and
Sine Moments
Dr. S.P. Yung
The University
of Hong Kong
Abstract
Since the discovery of compactly-supported wavelets by
I. Daubechies in 1988, wavelets have been studies and
applied vigorously. From a simplified stand-point, wavelets are basis functions
whose decomposition coefficients can be computed in a relatively few numerical
operations and they are localized both in time and frequency (i.e. a local
disturbance affects only finitely-many coefficients in a wavelet
decomposition). In this talk, we shall exhibit a new species of discrete
wavelet filters that possess Fourier, Cosine or Sine moments. We have found
that their decomposition coefficients decay very rapidly and they have similar
merits as those of the Daubechies’. Their
performances in signal and image compressions will also be discussed.