Title :
Finite Prolate Spheroidal Sequences and Their Applications I: Generation and Properties
Author_Institution :
Department of Computer Science, University of Warwick, Coventry CV4 7AL, England.
Abstract :
The finite prolate spheroidal sequences are those finite sequences which have extremal energy concentrations both in space and spatial frequency. The first part of the paper is devoted to a study of the eigenvalue problem defining the sequences. This reveals those fundamental properties of the sequences which are relevant in image processing applications and shows how they can be generated efficiently. A new version of the sampling theorem is also demonstrated.
Keywords :
Discrete Fourier transforms; Eigenvalues and eigenfunctions; Frequency; Image processing; Image sampling; Sampling methods; Signal processing; Signal sampling; Uncertainty; Vectors; Eigenvalue analysis; feature description; spatial frequency filtering; uncertainty principle;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
DOI :
10.1109/TPAMI.1987.4767985