Object recognition based on critical nodes

Arda Boluk, M. Fatih Demirci

Research output: Contribution to journalArticle

Abstract

In recent decades, the need for efficient and effective image search from large databases has increased. In this paper, we present a novel shape matching framework based on structures common to similar shapes. After representing shapes as medial axis graphs, in which nodes show skeleton points and edges connect nearby points, we determine the critical nodes connecting or representing a shape’s different parts. By using the shortest path distance from each skeleton (node) to each of the critical nodes, we effectively retrieve shapes similar to a given query through a transportation-based distance function. To improve the effectiveness of the proposed approach, we employ a unified framework that takes advantage of the feature representation of the proposed algorithm and the classification capability of a supervised machine learning algorithm. A set of shape retrieval experiments including a comparison with several well-known approaches demonstrate the proposed algorithm’s efficacy and perturbation experiments show its robustness.

Original languageEnglish
Pages (from-to)147-163
Number of pages17
JournalPattern Analysis and Applications
Volume22
Issue number1
DOIs
Publication statusPublished - Feb 5 2019

Fingerprint

Object recognition
Learning algorithms
Learning systems
Experiments

Keywords

  • Earth mover’s distance
  • Medial axis graph
  • Shape matching
  • Shape retrieval

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Artificial Intelligence

Cite this

Object recognition based on critical nodes. / Boluk, Arda; Demirci, M. Fatih.

In: Pattern Analysis and Applications, Vol. 22, No. 1, 05.02.2019, p. 147-163.

Research output: Contribution to journalArticle

@article{7472c0a83175410bb82c77c49ab2239c,
title = "Object recognition based on critical nodes",
abstract = "In recent decades, the need for efficient and effective image search from large databases has increased. In this paper, we present a novel shape matching framework based on structures common to similar shapes. After representing shapes as medial axis graphs, in which nodes show skeleton points and edges connect nearby points, we determine the critical nodes connecting or representing a shape’s different parts. By using the shortest path distance from each skeleton (node) to each of the critical nodes, we effectively retrieve shapes similar to a given query through a transportation-based distance function. To improve the effectiveness of the proposed approach, we employ a unified framework that takes advantage of the feature representation of the proposed algorithm and the classification capability of a supervised machine learning algorithm. A set of shape retrieval experiments including a comparison with several well-known approaches demonstrate the proposed algorithm’s efficacy and perturbation experiments show its robustness.",
keywords = "Earth mover’s distance, Medial axis graph, Shape matching, Shape retrieval",
author = "Arda Boluk and Demirci, {M. Fatih}",
year = "2019",
month = "2",
day = "5",
doi = "10.1007/s10044-018-00777-w",
language = "English",
volume = "22",
pages = "147--163",
journal = "Pattern Analysis and Applications",
issn = "1433-7541",
publisher = "Springer London",
number = "1",

}

TY - JOUR

T1 - Object recognition based on critical nodes

AU - Boluk, Arda

AU - Demirci, M. Fatih

PY - 2019/2/5

Y1 - 2019/2/5

N2 - In recent decades, the need for efficient and effective image search from large databases has increased. In this paper, we present a novel shape matching framework based on structures common to similar shapes. After representing shapes as medial axis graphs, in which nodes show skeleton points and edges connect nearby points, we determine the critical nodes connecting or representing a shape’s different parts. By using the shortest path distance from each skeleton (node) to each of the critical nodes, we effectively retrieve shapes similar to a given query through a transportation-based distance function. To improve the effectiveness of the proposed approach, we employ a unified framework that takes advantage of the feature representation of the proposed algorithm and the classification capability of a supervised machine learning algorithm. A set of shape retrieval experiments including a comparison with several well-known approaches demonstrate the proposed algorithm’s efficacy and perturbation experiments show its robustness.

AB - In recent decades, the need for efficient and effective image search from large databases has increased. In this paper, we present a novel shape matching framework based on structures common to similar shapes. After representing shapes as medial axis graphs, in which nodes show skeleton points and edges connect nearby points, we determine the critical nodes connecting or representing a shape’s different parts. By using the shortest path distance from each skeleton (node) to each of the critical nodes, we effectively retrieve shapes similar to a given query through a transportation-based distance function. To improve the effectiveness of the proposed approach, we employ a unified framework that takes advantage of the feature representation of the proposed algorithm and the classification capability of a supervised machine learning algorithm. A set of shape retrieval experiments including a comparison with several well-known approaches demonstrate the proposed algorithm’s efficacy and perturbation experiments show its robustness.

KW - Earth mover’s distance

KW - Medial axis graph

KW - Shape matching

KW - Shape retrieval

UR - http://www.scopus.com/inward/record.url?scp=85059856037&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85059856037&partnerID=8YFLogxK

U2 - 10.1007/s10044-018-00777-w

DO - 10.1007/s10044-018-00777-w

M3 - Article

AN - SCOPUS:85059856037

VL - 22

SP - 147

EP - 163

JO - Pattern Analysis and Applications

JF - Pattern Analysis and Applications

SN - 1433-7541

IS - 1

ER -