Distributed high dimensional information theoretical image registration via random projections

Zoltán Szabó, A. Lőrincz

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

Information theoretical measures, such as entropy, mutual information, and various divergences, exhibit robust characteristics in image registration applications. However, the estimation of these quantities is computationally intensive in high dimensions. On the other hand, consistent estimation from pairwise distances of the sample points is possible, which suits random projection (RP) based low dimensional embeddings. We adapt the RP technique to this task by means of a simple ensemble method. To the best of our knowledge, this is the first distributed, RP based information theoretical image registration approach. The efficiency of the method is demonstrated through numerical examples.

Original languageEnglish
Pages (from-to)894-902
Number of pages9
JournalDigital Signal Processing: A Review Journal
Volume22
Issue number6
DOIs
Publication statusPublished - Dec 2012

Fingerprint

Image registration
Entropy

Keywords

  • Distributed solution
  • High dimensional features
  • Information theoretical image registration
  • Random projection

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Distributed high dimensional information theoretical image registration via random projections. / Szabó, Zoltán; Lőrincz, A.

In: Digital Signal Processing: A Review Journal, Vol. 22, No. 6, 12.2012, p. 894-902.

Research output: Contribution to journalArticle

@article{53ec453b35384441822176b5d5c5524f,
title = "Distributed high dimensional information theoretical image registration via random projections",
abstract = "Information theoretical measures, such as entropy, mutual information, and various divergences, exhibit robust characteristics in image registration applications. However, the estimation of these quantities is computationally intensive in high dimensions. On the other hand, consistent estimation from pairwise distances of the sample points is possible, which suits random projection (RP) based low dimensional embeddings. We adapt the RP technique to this task by means of a simple ensemble method. To the best of our knowledge, this is the first distributed, RP based information theoretical image registration approach. The efficiency of the method is demonstrated through numerical examples.",
keywords = "Distributed solution, High dimensional features, Information theoretical image registration, Random projection",
author = "Zolt{\'a}n Szab{\'o} and A. Lőrincz",
year = "2012",
month = "12",
doi = "10.1016/j.dsp.2012.04.018",
language = "English",
volume = "22",
pages = "894--902",
journal = "Digital Signal Processing: A Review Journal",
issn = "1051-2004",
publisher = "Elsevier Inc.",
number = "6",

}

TY - JOUR

T1 - Distributed high dimensional information theoretical image registration via random projections

AU - Szabó, Zoltán

AU - Lőrincz, A.

PY - 2012/12

Y1 - 2012/12

N2 - Information theoretical measures, such as entropy, mutual information, and various divergences, exhibit robust characteristics in image registration applications. However, the estimation of these quantities is computationally intensive in high dimensions. On the other hand, consistent estimation from pairwise distances of the sample points is possible, which suits random projection (RP) based low dimensional embeddings. We adapt the RP technique to this task by means of a simple ensemble method. To the best of our knowledge, this is the first distributed, RP based information theoretical image registration approach. The efficiency of the method is demonstrated through numerical examples.

AB - Information theoretical measures, such as entropy, mutual information, and various divergences, exhibit robust characteristics in image registration applications. However, the estimation of these quantities is computationally intensive in high dimensions. On the other hand, consistent estimation from pairwise distances of the sample points is possible, which suits random projection (RP) based low dimensional embeddings. We adapt the RP technique to this task by means of a simple ensemble method. To the best of our knowledge, this is the first distributed, RP based information theoretical image registration approach. The efficiency of the method is demonstrated through numerical examples.

KW - Distributed solution

KW - High dimensional features

KW - Information theoretical image registration

KW - Random projection

UR - http://www.scopus.com/inward/record.url?scp=84865993680&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84865993680&partnerID=8YFLogxK

U2 - 10.1016/j.dsp.2012.04.018

DO - 10.1016/j.dsp.2012.04.018

M3 - Article

AN - SCOPUS:84865993680

VL - 22

SP - 894

EP - 902

JO - Digital Signal Processing: A Review Journal

JF - Digital Signal Processing: A Review Journal

SN - 1051-2004

IS - 6

ER -