Signal representation using wavelet transform modulus maxima

Author

Liew, A. ; Nguyen, D.T.

Author_Institution

Dept. of Electr. & Electron. Eng., Tasmania Univ., Hobart, Tas., Australia

Volume

1

fYear

1994

fDate

14-18 Nov 1994

Firstpage

195

Abstract

We examine the representation of signals by the wavelet transform modulus maxima together with the alternating projection reconstruction algorithm proposed by Mallat. We show that the projection algorithm of Mallat (see IEEE Trans. Patt. Anal. Mach. Intell., vol.14, p.710-732, July 1992, and IEEE Trans. Inform. Theory, vol.38, p.617-643, March 1992) is not minimum distance and not necessarily nonexpansive. Convergence is therefore not guaranteed. We describe an alternative projection operator which is minimum distance and nonexpansive and show that it is monotonically converging

Keywords

convergence of numerical methods; signal reconstruction; signal representation; wavelet transforms; alternating projection reconstruction algorithm; convergence; minimum distance; projection operator; signal representation; wavelet transform modulus maxima; Discrete wavelet transforms; Image edge detection; Image reconstruction; Modular construction; Reconstruction algorithms; Signal processing; Signal processing algorithms; Signal representations; Wavelet analysis; Wavelet transforms;

fLanguage

English

Publisher

ieee

Conference_Titel

Singapore ICCS '94. Conference Proceedings.

Print_ISBN

0-7803-2046-8

Type

conf

DOI

10.1109/ICCS.1994.474079

Filename

474079