Contour-based corner detection and classification by using mean projection transform

Seyed Mostafa Mousavi Kahaki, Md. Jan Nordin, Amir Hossein Ashtari

Research output: Contribution to journalArticle

26 Citations (Scopus)

Abstract

Image corner detection is a fundamental task in computer vision. Many applications require reliable detectors to accurately detect corner points, commonly achieved by using image contour information. The curvature definition is sensitive to local variation and edge aliasing, and available smoothing methods are not sufficient to address these problems properly. Hence, we propose Mean Projection Transform (MPT) as a corner classifier and parabolic fit approximation to form a robust detector. The first step is to extract corner candidates using MPT based on the integral properties of the local contours in both the horizontal and vertical directions. Then, an approximation of the parabolic fit is calculated to localize the candidate corner points. The proposed method presents fewer false-positive (FP) and false-negative (FN) points compared with recent standard corner detection techniques, especially in comparison with curvature scale space (CSS) methods. Moreover, a new evaluation metric, called accuracy of repeatability (AR), is introduced. AR combines repeatability and the localization error () for finding the probability of correct detection in the target image. The output results exhibit better repeatability, localization, and AR for the detected points compared with the criteria in original and transformed images.

Original languageEnglish
Pages (from-to)4126-4143
Number of pages18
JournalSensors (Switzerland)
Volume14
Issue number3
DOIs
Publication statusPublished - 28 Feb 2014

Fingerprint

projection
Mathematical transformations
Detectors
Computer vision
Classifiers
curvature
detectors
computer vision
classifiers
approximation
smoothing
evaluation
output
Direction compound

Keywords

  • Contour-based corner detector
  • Corner detection
  • Mean projection transform
  • Polygonal approximation

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Atomic and Molecular Physics, and Optics
  • Analytical Chemistry
  • Biochemistry

Cite this

Contour-based corner detection and classification by using mean projection transform. / Kahaki, Seyed Mostafa Mousavi; Nordin, Md. Jan; Ashtari, Amir Hossein.

In: Sensors (Switzerland), Vol. 14, No. 3, 28.02.2014, p. 4126-4143.

Research output: Contribution to journalArticle

Kahaki, Seyed Mostafa Mousavi ; Nordin, Md. Jan ; Ashtari, Amir Hossein. / Contour-based corner detection and classification by using mean projection transform. In: Sensors (Switzerland). 2014 ; Vol. 14, No. 3. pp. 4126-4143.
@article{8d98f0009a2d4c48953c63fce3328026,
title = "Contour-based corner detection and classification by using mean projection transform",
abstract = "Image corner detection is a fundamental task in computer vision. Many applications require reliable detectors to accurately detect corner points, commonly achieved by using image contour information. The curvature definition is sensitive to local variation and edge aliasing, and available smoothing methods are not sufficient to address these problems properly. Hence, we propose Mean Projection Transform (MPT) as a corner classifier and parabolic fit approximation to form a robust detector. The first step is to extract corner candidates using MPT based on the integral properties of the local contours in both the horizontal and vertical directions. Then, an approximation of the parabolic fit is calculated to localize the candidate corner points. The proposed method presents fewer false-positive (FP) and false-negative (FN) points compared with recent standard corner detection techniques, especially in comparison with curvature scale space (CSS) methods. Moreover, a new evaluation metric, called accuracy of repeatability (AR), is introduced. AR combines repeatability and the localization error () for finding the probability of correct detection in the target image. The output results exhibit better repeatability, localization, and AR for the detected points compared with the criteria in original and transformed images.",
keywords = "Contour-based corner detector, Corner detection, Mean projection transform, Polygonal approximation",
author = "Kahaki, {Seyed Mostafa Mousavi} and Nordin, {Md. Jan} and Ashtari, {Amir Hossein}",
year = "2014",
month = "2",
day = "28",
doi = "10.3390/s140304126",
language = "English",
volume = "14",
pages = "4126--4143",
journal = "Sensors (Switzerland)",
issn = "1424-8220",
publisher = "Multidisciplinary Digital Publishing Institute (MDPI)",
number = "3",

}

TY - JOUR

T1 - Contour-based corner detection and classification by using mean projection transform

AU - Kahaki, Seyed Mostafa Mousavi

AU - Nordin, Md. Jan

AU - Ashtari, Amir Hossein

PY - 2014/2/28

Y1 - 2014/2/28

N2 - Image corner detection is a fundamental task in computer vision. Many applications require reliable detectors to accurately detect corner points, commonly achieved by using image contour information. The curvature definition is sensitive to local variation and edge aliasing, and available smoothing methods are not sufficient to address these problems properly. Hence, we propose Mean Projection Transform (MPT) as a corner classifier and parabolic fit approximation to form a robust detector. The first step is to extract corner candidates using MPT based on the integral properties of the local contours in both the horizontal and vertical directions. Then, an approximation of the parabolic fit is calculated to localize the candidate corner points. The proposed method presents fewer false-positive (FP) and false-negative (FN) points compared with recent standard corner detection techniques, especially in comparison with curvature scale space (CSS) methods. Moreover, a new evaluation metric, called accuracy of repeatability (AR), is introduced. AR combines repeatability and the localization error () for finding the probability of correct detection in the target image. The output results exhibit better repeatability, localization, and AR for the detected points compared with the criteria in original and transformed images.

AB - Image corner detection is a fundamental task in computer vision. Many applications require reliable detectors to accurately detect corner points, commonly achieved by using image contour information. The curvature definition is sensitive to local variation and edge aliasing, and available smoothing methods are not sufficient to address these problems properly. Hence, we propose Mean Projection Transform (MPT) as a corner classifier and parabolic fit approximation to form a robust detector. The first step is to extract corner candidates using MPT based on the integral properties of the local contours in both the horizontal and vertical directions. Then, an approximation of the parabolic fit is calculated to localize the candidate corner points. The proposed method presents fewer false-positive (FP) and false-negative (FN) points compared with recent standard corner detection techniques, especially in comparison with curvature scale space (CSS) methods. Moreover, a new evaluation metric, called accuracy of repeatability (AR), is introduced. AR combines repeatability and the localization error () for finding the probability of correct detection in the target image. The output results exhibit better repeatability, localization, and AR for the detected points compared with the criteria in original and transformed images.

KW - Contour-based corner detector

KW - Corner detection

KW - Mean projection transform

KW - Polygonal approximation

UR - http://www.scopus.com/inward/record.url?scp=84896892280&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84896892280&partnerID=8YFLogxK

U2 - 10.3390/s140304126

DO - 10.3390/s140304126

M3 - Article

AN - SCOPUS:84896892280

VL - 14

SP - 4126

EP - 4143

JO - Sensors (Switzerland)

JF - Sensors (Switzerland)

SN - 1424-8220

IS - 3

ER -