Component evolution analysis in descriptor graphs for descriptor ranking

Levente Kovács, Anita Keszler, T. Szirányi

Research output: Contribution to journalArticle

1 Citation (Scopus)

Abstract

This paper presents a method based on graph behaviour analysis for the evaluation of descriptor graphs (applied to image/video datasets) for descriptor performance analysis and ranking. Starting from the Erdos-Rényi model on uniform random graphs, the paper presents results of investigating random geometric graph behaviour in relation with the appearance of the giant component as a basis for ranking descriptors based on their clustering properties. We analyse the phase transition and the evolution of components in such graphs, and based on their behaviour, the corresponding descriptors are compared, ranked, and validated in retrieval tests. The goal is to build an evaluation framework where descriptors can be analysed for automatic feature selection.

Original languageEnglish
Pages (from-to)1-12
Number of pages12
JournalDigital Signal Processing: A Review Journal
Volume31
DOIs
Publication statusPublished - 2014

Fingerprint

Feature extraction
Phase transitions

Keywords

  • Descriptor evaluation
  • Feature extraction
  • Feature selection
  • Graph components
  • Graph representation

ASJC Scopus subject areas

  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Component evolution analysis in descriptor graphs for descriptor ranking. / Kovács, Levente; Keszler, Anita; Szirányi, T.

In: Digital Signal Processing: A Review Journal, Vol. 31, 2014, p. 1-12.

Research output: Contribution to journalArticle

@article{44b6ea6c513243818a68c51067a82b7c,
title = "Component evolution analysis in descriptor graphs for descriptor ranking",
abstract = "This paper presents a method based on graph behaviour analysis for the evaluation of descriptor graphs (applied to image/video datasets) for descriptor performance analysis and ranking. Starting from the Erdos-R{\'e}nyi model on uniform random graphs, the paper presents results of investigating random geometric graph behaviour in relation with the appearance of the giant component as a basis for ranking descriptors based on their clustering properties. We analyse the phase transition and the evolution of components in such graphs, and based on their behaviour, the corresponding descriptors are compared, ranked, and validated in retrieval tests. The goal is to build an evaluation framework where descriptors can be analysed for automatic feature selection.",
keywords = "Descriptor evaluation, Feature extraction, Feature selection, Graph components, Graph representation",
author = "Levente Kov{\'a}cs and Anita Keszler and T. Szir{\'a}nyi",
year = "2014",
doi = "10.1016/j.dsp.2014.04.013",
language = "English",
volume = "31",
pages = "1--12",
journal = "Digital Signal Processing: A Review Journal",
issn = "1051-2004",
publisher = "Elsevier Inc.",

}

TY - JOUR

T1 - Component evolution analysis in descriptor graphs for descriptor ranking

AU - Kovács, Levente

AU - Keszler, Anita

AU - Szirányi, T.

PY - 2014

Y1 - 2014

N2 - This paper presents a method based on graph behaviour analysis for the evaluation of descriptor graphs (applied to image/video datasets) for descriptor performance analysis and ranking. Starting from the Erdos-Rényi model on uniform random graphs, the paper presents results of investigating random geometric graph behaviour in relation with the appearance of the giant component as a basis for ranking descriptors based on their clustering properties. We analyse the phase transition and the evolution of components in such graphs, and based on their behaviour, the corresponding descriptors are compared, ranked, and validated in retrieval tests. The goal is to build an evaluation framework where descriptors can be analysed for automatic feature selection.

AB - This paper presents a method based on graph behaviour analysis for the evaluation of descriptor graphs (applied to image/video datasets) for descriptor performance analysis and ranking. Starting from the Erdos-Rényi model on uniform random graphs, the paper presents results of investigating random geometric graph behaviour in relation with the appearance of the giant component as a basis for ranking descriptors based on their clustering properties. We analyse the phase transition and the evolution of components in such graphs, and based on their behaviour, the corresponding descriptors are compared, ranked, and validated in retrieval tests. The goal is to build an evaluation framework where descriptors can be analysed for automatic feature selection.

KW - Descriptor evaluation

KW - Feature extraction

KW - Feature selection

KW - Graph components

KW - Graph representation

UR - http://www.scopus.com/inward/record.url?scp=84901818757&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84901818757&partnerID=8YFLogxK

U2 - 10.1016/j.dsp.2014.04.013

DO - 10.1016/j.dsp.2014.04.013

M3 - Article

AN - SCOPUS:84901818757

VL - 31

SP - 1

EP - 12

JO - Digital Signal Processing: A Review Journal

JF - Digital Signal Processing: A Review Journal

SN - 1051-2004

ER -