A performance evaluation of statistical tests for edge
detection in textured images

J 2014

A performance evaluation of statistical tests for edge detection in textured images

WILLIAMS, Ian, Nicholas BOWRING and David SVOBODA

Basic information

Original name

A performance evaluation of statistical tests for edge detection in textured images

Authors

WILLIAMS, Ian (826 United Kingdom of Great Britain and Northern Ireland), Nicholas BOWRING (826 United Kingdom of Great Britain and Northern Ireland) and David SVOBODA (203 Czech Republic, guarantor, belonging to the institution)

Edition

Computer Vision and Image Understanding, Elsevier, 2014, 1077-3142

Other information

Language

English

Type of outcome

Článek v odborném periodiku

Field of Study

20200 2.2 Electrical engineering, Electronic engineering, Information engineering

Country of publisher

Netherlands

Confidentiality degree

není předmětem státního či obchodního tajemství

Impact factor

Impact factor: 1.540

RIV identification code

RIV/00216224:14330/14:00075212

Organization unit

Faculty of Informatics

DOI

http://dx.doi.org/10.1016/j.cviu.2014.02.009

UT WoS

000334394900011

Keywords in English

Edge detection; Statistical tests; Textured images; Histological images; Performance measures

Abstract

V originále

This work presents an objective performance analysis of statistical tests for edge detection which are suitable for textured or cluttered images. The tests are subdivided into two-sample parametric and non-parametric tests and are applied using a dual-region based edge detector which analyses local image texture difference. Through a series of experimental tests objective results are presented across a comprehensive dataset of images using a Pixel Correspondence Metric (PCM). The results show that statistical tests can in many cases, outperform the Canny edge detection method giving robust edge detection, accurate edge localisation and improved edge connectivity throughout. A visual comparison of the tests is also presented using representative images taken from typical textured histological data sets. The results conclude that the non-parametric Chi Square and Kolmogorov Smirnov statistical tests are the most robust edge detection tests where image statistical properties cannot be assumed a priori or where intensity changes in the image are nonuniform and that the parametric Difference of Boxes test and the Student’s t-test are the most suitable for intensity based edges. Conclusions and recommendations are finally presented contrasting the tests and giving guidelines for their practical use while finally confirming which situations improved edge detection can be expected.

Citovat

WILLIAMS, Ian, Nicholas BOWRING and David SVOBODA. A performance evaluation of statistical tests for edge detection in textured images. Computer Vision and Image Understanding. Elsevier, 2014, vol. 122, May 2014, p. 115-130. ISSN 1077-3142. Available from: https://dx.doi.org/10.1016/j.cviu.2014.02.009.

@article{1173216,
   author = {Williams, Ian and Bowring, Nicholas and Svoboda, David},
   article_number = {May 2014},
   doi = {http://dx.doi.org/10.1016/j.cviu.2014.02.009},
   keywords = {Edge detection; Statistical tests; Textured images; Histological images; Performance measures},
   language = {eng},
   issn = {1077-3142},
   journal = {Computer Vision and Image Understanding},
   title = {A performance evaluation of statistical tests for edge detection in textured images},
   volume = {122},
   year = {2014}
}

TY  - JOUR
ID  - 1173216
AU  - Williams, Ian - Bowring, Nicholas - Svoboda, David
PY  - 2014
TI  - A performance evaluation of statistical tests for edge detection in textured images
JF  - Computer Vision and Image Understanding
VL  - 122
IS  - May 2014
SP  - 115-130
EP  - 115-130
PB  - Elsevier
SN  - 10773142
KW  - Edge detection
KW  - Statistical tests
KW  - Textured images
KW  - Histological images
KW  - Performance measures
N2  - This work presents an objective performance analysis of statistical tests for edge detection which are suitable for textured or cluttered images. The tests are subdivided into two-sample parametric and non-parametric tests and are applied using a dual-region based edge detector which analyses local image texture difference. Through a series of experimental tests objective results are presented across a comprehensive dataset of images using a Pixel Correspondence Metric (PCM). The results show that statistical tests can in many cases, outperform the Canny edge detection method giving robust edge detection, accurate edge localisation and improved edge connectivity throughout. A visual comparison of the tests is also presented using representative images taken from typical textured histological data sets. The results conclude that the non-parametric Chi Square and Kolmogorov Smirnov statistical tests are the most robust edge detection tests where image statistical properties cannot be assumed a priori or where intensity changes in the image are nonuniform and that the parametric Difference of Boxes test and the Student’s t-test are the most suitable for intensity based edges. Conclusions and recommendations are finally presented contrasting the tests and giving guidelines for their practical use while finally confirming which situations improved edge detection can be expected.
ER  -

WILLIAMS, Ian, Nicholas BOWRING and David SVOBODA. A performance evaluation of statistical tests for edge detection in textured images. \textit{Computer Vision and Image Understanding}. Elsevier, 2014, vol.~122, May 2014, p.~115-130. ISSN~1077-3142. Available from: https://dx.doi.org/10.1016/j.cviu.2014.02.009.

Detailed Information on Publication Record