Classifying compact convex sets with frames

Amina Chebira, Matthew Fickus, Jelena Kovacevic

Research output: Contribution to journalArticle

Abstract

Classification is a fundamental image processing task. Recent empirical evidence suggests that classification algorithms which make use of redundant linear transforms will regularly outperform their nonredundant counterparts. We provide a rigorous explanation of this phenomenon in the single-class case. We begin by developing a measure-theoretic analysis of the set of points at which a given decision rule will have an intolerable chance of making a classification error. We then apply this general theory to the special case where the class is compact and convex, showing that such a class may be arbitrarily well approximated by frame sets, namely, preimages of hyperrectangles under frame analysis operators. This leads to a frame-based classification scheme in which frame coefficients are regarded as features. We show that, indeed, the accuracy of such a classification scheme approaches perfect accuracy as the redundancy of the frame grows large.

Original languageEnglish (US)
Pages (from-to)73-86
Number of pages14
JournalApplied and Computational Harmonic Analysis
Volume27
Issue number1
DOIs
StatePublished - Jul 1 2009

Fingerprint

Compact Convex Set
Decision Rules
Classification Algorithm
Set of points
Redundancy
Image Processing
Image processing
Transform
Coefficient
Operator
Class

Keywords

  • Classification
  • Convex sets
  • Decision rule
  • Frames

ASJC Scopus subject areas

  • Applied Mathematics

Cite this

Classifying compact convex sets with frames. / Chebira, Amina; Fickus, Matthew; Kovacevic, Jelena.

In: Applied and Computational Harmonic Analysis, Vol. 27, No. 1, 01.07.2009, p. 73-86.

Research output: Contribution to journalArticle

Chebira, Amina ; Fickus, Matthew ; Kovacevic, Jelena. / Classifying compact convex sets with frames. In: Applied and Computational Harmonic Analysis. 2009 ; Vol. 27, No. 1. pp. 73-86.
@article{050d969888644c6e80f0288b7994af14,
title = "Classifying compact convex sets with frames",
abstract = "Classification is a fundamental image processing task. Recent empirical evidence suggests that classification algorithms which make use of redundant linear transforms will regularly outperform their nonredundant counterparts. We provide a rigorous explanation of this phenomenon in the single-class case. We begin by developing a measure-theoretic analysis of the set of points at which a given decision rule will have an intolerable chance of making a classification error. We then apply this general theory to the special case where the class is compact and convex, showing that such a class may be arbitrarily well approximated by frame sets, namely, preimages of hyperrectangles under frame analysis operators. This leads to a frame-based classification scheme in which frame coefficients are regarded as features. We show that, indeed, the accuracy of such a classification scheme approaches perfect accuracy as the redundancy of the frame grows large.",
keywords = "Classification, Convex sets, Decision rule, Frames",
author = "Amina Chebira and Matthew Fickus and Jelena Kovacevic",
year = "2009",
month = "7",
day = "1",
doi = "10.1016/j.acha.2008.11.002",
language = "English (US)",
volume = "27",
pages = "73--86",
journal = "Applied and Computational Harmonic Analysis",
issn = "1063-5203",
publisher = "Academic Press Inc.",
number = "1",

}

TY - JOUR

T1 - Classifying compact convex sets with frames

AU - Chebira, Amina

AU - Fickus, Matthew

AU - Kovacevic, Jelena

PY - 2009/7/1

Y1 - 2009/7/1

N2 - Classification is a fundamental image processing task. Recent empirical evidence suggests that classification algorithms which make use of redundant linear transforms will regularly outperform their nonredundant counterparts. We provide a rigorous explanation of this phenomenon in the single-class case. We begin by developing a measure-theoretic analysis of the set of points at which a given decision rule will have an intolerable chance of making a classification error. We then apply this general theory to the special case where the class is compact and convex, showing that such a class may be arbitrarily well approximated by frame sets, namely, preimages of hyperrectangles under frame analysis operators. This leads to a frame-based classification scheme in which frame coefficients are regarded as features. We show that, indeed, the accuracy of such a classification scheme approaches perfect accuracy as the redundancy of the frame grows large.

AB - Classification is a fundamental image processing task. Recent empirical evidence suggests that classification algorithms which make use of redundant linear transforms will regularly outperform their nonredundant counterparts. We provide a rigorous explanation of this phenomenon in the single-class case. We begin by developing a measure-theoretic analysis of the set of points at which a given decision rule will have an intolerable chance of making a classification error. We then apply this general theory to the special case where the class is compact and convex, showing that such a class may be arbitrarily well approximated by frame sets, namely, preimages of hyperrectangles under frame analysis operators. This leads to a frame-based classification scheme in which frame coefficients are regarded as features. We show that, indeed, the accuracy of such a classification scheme approaches perfect accuracy as the redundancy of the frame grows large.

KW - Classification

KW - Convex sets

KW - Decision rule

KW - Frames

UR - http://www.scopus.com/inward/record.url?scp=67349259635&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=67349259635&partnerID=8YFLogxK

U2 - 10.1016/j.acha.2008.11.002

DO - 10.1016/j.acha.2008.11.002

M3 - Article

AN - SCOPUS:67349259635

VL - 27

SP - 73

EP - 86

JO - Applied and Computational Harmonic Analysis

JF - Applied and Computational Harmonic Analysis

SN - 1063-5203

IS - 1

ER -