6.2.1 Evaluations of Edge Detectors, Multiple Algorithms

Chapter Contents (Back)
Evaluation, Edges.

Fram, J.R., and Deutsch, E.S.,
On the Quantitative Evaluation of Edge Detection Schemes and Their Comparisons with Human Performance,
TC(24), No. 6, June 1975, pp. 616-627. Ground truth, comparison using vertical step edge. BibRef 7506

Deutsch, E.S., and Fram, J.R.,
A Quantitative Study of the Orientational Bias of Some Edge Detector Schemes,
TC(27), No. 3, March 1978, pp. 205-213. And comments below. BibRef 7803

MacLeod, I.D.G.,
Comments on 'A Quantitative Study of the Orientation Bias of Some Edge Detector Schemes',
PAMI(1), No. 4, October 1979, 408-409. Comment on above paper. BibRef 7910

Abdou, I.E., and Pratt, W.K.,
Qualitative Design and Evaluation of Enhancement/Thresholding Edge Detector,
PIEEE(67), No. 5, May 1979, pp. 753-763. Design of edge evaluation experiments. Evaluation for horizontal, vertical and diagonal step edges. BibRef 7905

Abdou, I.E.,
Quantitative Methods of Edge Detection,
Ph.D.July 1978. BibRef 7807 USC BibRef

Panda, D.P., and Dubitzki, T.,
Statistical Analysis of Some Edge Operators,
CGIP(11), No. 4, December 1979, pp. 313-348.
WWW Link. BibRef 7912

Peli, T., and Malah, D.,
A Study of Edge Detection Algorithms,
CGIP(20), No. 1, September 1982, pp. 1-21.
WWW Link. BibRef 8209

Basseville, M., Benveniste, A.,
Design and Comparative Study of Some Sequential Jump Detection Algorithms for Digital Signals,
ASSP(31), 1983, pp. 521-535. BibRef 8300

Bernsen, J.A.C.,
An Objective and Subjective Evaluation of Edge Detection Methods in Images,
Phillips J. Res.(46), 1991, pp. 57-94. BibRef 9100

Fleck, M.M.,
Some Defects in Finite-Difference Edge Finders,
PAMI(14), No. 3, March 1992, pp. 337-345.
IEEE DOI Edges, Evaluation. Analysis of See also Computational Approach to Edge Detection, A. See also Two Dimensional Optimal Edge Recognition Using Matched and Weiner Filters for Machine Vision. and See also Theory of Edge Detection. with regard to gaps, deformations, and spurious boundaries. The results are similar, all have major problems due to design, etc. Interesting closing question that the unpublished details are more important the the shape of the filters. BibRef 9203

Zhou, Y.T., Venkateswar, V., and Chellappa, R.,
Edge Detection and Linear Feature Extraction Using a 2-D Random Field Model,
PAMI(11), No. 1, January 1989, pp. 84-95.
IEEE DOI Comparison of the results with See also Linear Feature Extraction and Description. See also Facet Model for Image Data, A. See also Computational Approach to Edge Detection, A. and See also Theory of Edge Detection. Find edges and then find long straight lines. BibRef 8901

Kanungo, T., Jaisimha, M.Y., Palmer, J., Haralick, R.M.,
A Methodology for Quantitative Performance Evaluation of Detection Algorithms,
IP(4), No. 12, December 1995, pp. 1667-1674.
IEEE DOI BibRef 9512
Earlier:
A Quantitative Methodology for Analyzing the Performance of Detection Algorithms,
ICCV93(247-252).
IEEE DOI Vertical edge with added noise. BibRef

Ramesh, V., and Haralick, R.M.,
Performance Characterization of Edge Operators,
DARPA93(1071-1079). BibRef 9300
And:
Performance Characterization of Edge Detectors,
SPIE(1708), April 1992, pp. 252-266. A ramp edge embedded in noise. BibRef

Haralick, R.M.[Robert M.], and Ramesh, V.[Visvanathan],
An Integrated Gradient Edge Detector -- Theory and Performance Evaluation,
ARPA94(I:689-702). BibRef 9400

Ramesh, V., Haralick, R.M., Zhang, X., Nadadur, D.C., Thornton, K.B.,
Automatic Selection of Tuning Parameters for Feature Extraction Sequences,
CVPR94(672-677).
IEEE DOI BibRef 9400

Cho, K., Meer, P., Cabrera, J.,
Performance Assessment Through Bootstrap,
PAMI(19), No. 11, November 1997, pp. 1185-1198.
IEEE DOI 9712
BibRef
And: Correction: PAMI(20), No. 1, January 1998, pp. 94.
IEEE DOI BibRef
Earlier:
Quantitative Evaluation of Performance Through Bootstrapping: Edge Detection,
SCV95(491-496).
IEEE DOI Rutgers University. Evaluation of detectors using the same edge model. Bootstrap is a resampling technique. BibRef

Heath, M.D.[Mike D.], Sarkar, S.[Sudeep], Sanocki, T.A.[Thomas A.], and Bowyer, K.W.[Kevin W.],
A Robust Visual Method for Assessing the Relative Performance of Edge Detection Algorithms,
PAMI(19), No. 12, December 1997, pp. 1338-1359.
IEEE DOI 9712
Table of 21 recent algorithms with what they compared to and how many examples. They did not show many results or make many comparisons. Also lists 12 comparison papers. Compares: Canny ( See also Computational Approach to Edge Detection, A. ), Nalwa-Binford ( See also On Detecting Edges. ), Iverson-Zucker ( See also Logical/Linear Operators for Image Curves. ), Bergholm ( See also Edge Focusing. ), and Rothwell. ( See also Driving Vision by Topology. ). Visual (human) comparison of results on real images with a variety of input parameters for each. BibRef

Sanocki, T.A.[Thomas A.], Bowyer, K.W.[Kevin W.], Heath, M.D.[Mike D.], Sarkar, S.[Sudeep],
Are Edges Sufficient for Object Recognition,
JEP:HPP(24), No. 1, 1998, pp. 340-349. Edges as line drawings are not reality, edges as outputs of local edge extractors are. These are not good enough (50% performance in a recognition task). BibRef 9800

Heath, M.D.[Mike D.], Sarkar, S.[Sudeep], Sanocki, T.A.[Thomas A.], Bowyer, K.W.[Kevin W.],
Comparison of Edge Detectors,
CVIU(69), No. 1, January 1998, pp. 38-54.
DOI Link Earlier version of See also Robust Visual Method for Assessing the Relative Performance of Edge Detection Algorithms, A. with fewer images and edge detectors and less rigorous sampling of parameter space. BibRef 9801

Heath, M.D.[Mike D.], Sarkar, S.[Sudeep], Sanocki, T.A.[Thomas A.], and Bowyer, K.W.[Kevin W.],
Comparison of Edge Detectors: A Methodology and Initial Study,
CVPR96(143-148).
IEEE DOI Use humans to rate the image algorithms. Pairwise comparisons. Canny ( See also Computational Approach to Edge Detection, A. ), Nalwa-Binford ( See also On Detecting Edges. ), Sarkar-Boyer ( See also On Optimal Infinite Impulse Response Edge Detection Filters. ), Sobel ( See also Visual Perception by a Computer. ). The article above extends these results with a better set of edge detectors. BibRef 9600

Dougherty, S.[Sean], and Bowyer, K.W.[Kevin W.],
Objective Evaluation of Edge Detectors Using a Formally Defined Framework,
EEMTV98(xx). BibRef 9800
And: EEMCV98(xx). Uses Real Images, hand-specified ground truth, ROC analysis of true positive and false positives. 6 edge detectors. General conclusion is that the reputation of Canny ( See also Computational Approach to Edge Detection, A. ) is deserved. Heitger ( See also Feature Detection using Suppression and Enhancement. ) gives good results. Also included Bergholm ( See also Edge Focusing. ), Rothwell ( See also Driving Vision by Topology. ). Sobel ( See also Visual Perception by a Computer. ), Sarkar-Boyer ( See also On Optimal Infinite Impulse Response Edge Detection Filters. ). BibRef

Bowyer, K.W.[Kevin W.], Kranenburg, C.[Christine], Dougherty, S.[Sean],
Edge Detector Evaluation Using Empirical ROC Curves,
CVIU(84), No. 1, October 2001, pp. 77-103.
DOI Link 0203
BibRef
Earlier: CVPR99(I: 354-359).
IEEE DOI Explains the method of evaluation, used in the next one. Canny ( See also Computational Approach to Edge Detection, A. ), Heitger ( See also Feature Detection using Suppression and Enhancement. ) give the best results. Also included Bergholm ( See also Edge Focusing. ), Rothwell ( See also Driving Vision by Topology. ). Black ( See also Robust Anisotropic Diffusion. ), Sobel ( See also Visual Perception by a Computer. ), Susan ( See also Susan: A New Approach to Low-Level Image-Processing. ). BibRef

Dougherty, S., Bowyer, K.W., Kranenburg, C.,
ROC curve evaluation of edge detector performance,
ICIP98(II: 525-529).
IEEE DOI 9810
BibRef

Shin, M.C.[Min C.], Goldgof, D.B.[Dmitry B.], Bowyer, K.W.[Kevin W.],
Comparison of Edge Detector Performance through Use in an Object Recognition Task,
CVIU(84), No. 1, October 2001, pp. 160-178.
DOI Link 0203
BibRef
Earlier:
Comparison of Edge Detectors Using an Object Recognition Task,
CVPR99(I: 360-365).
IEEE DOI Cross refs for CVPR version, I assume they are the same papers. Use the recognition method of Huttenlocher, et al. ( See also Comparing Images Using the Hausdorff Distance. ). Using edge detectors from: Bergholm ( See also Edge Focusing. ), Canny ( See also Computational Approach to Edge Detection, A. ), Heitger ( See also Feature Detection using Suppression and Enhancement. ), Sobel ( See also Visual Perception by a Computer. ), Susan ( See also Susan: A New Approach to Low-Level Image-Processing. ). Training is essential. Conclusions are different from the other studies. BibRef

Shin, M.C.[Min C.], Goldgof, D.B.[Dmitry B.], Bowyer, K.W.[Kevin W.], Nikiforou, S.,
Comparison of Edge Detection Algorithms Using a Structure from Motion Task,
SMC-B(31), No. 4, August 2001, pp. 589-601.
IEEE Top Reference. 0109
Use the structure from motion accuracy to test edge detectors. Results are that these results are well correlated wht the results on pixel-level metrics. Canny ( See also Computational Approach to Edge Detection, A. )and Heitger ( See also Feature Detection using Suppression and Enhancement. )detector offer the best performance. BibRef

Shin, M.C.[Min Chul], Goldgof, D.B.[Dmitry B.], Bowyer, K.W.[Kevin W.],
An Objective Comparison Methodology of Edge Detection Algorithms Using a Structure from Motion Task,
EEMCV98(xx). BibRef 9800
Earlier: CVPR98(190-195).
IEEE DOI Bergholm ( See also Edge Focusing. ), Canny ( See also Computational Approach to Edge Detection, A. ), Rothwell. ( See also Driving Vision by Topology. ) and Sarkar ( See also On Optimal Infinite Impulse Response Edge Detection Filters. ). BibRef

Shin, M.C.[Min Chul], Goldgof, D.B.[Dmitry B.], and Bowyer, K.W.[Kevin W.],
Evaluation of Edge Detection Algorithms Using a Structure from Motion Task,
EEMCV98(xx). BibRef 9800
And:
An Objective Comparison Methodology of Edge Detection Algorithms Using a Structure from Motion Task,
EEMTV98(xx) BibRef

Ziou, D., Koukam, A.,
Knowledge Based Assistant for the Selection of Edge Detectors,
PR(31), No. 5, May 1998, pp. 587-596.
WWW Link. 9805
See also influence of edge direction on the estimation of edge contrast and orientation, The. BibRef

Nguyen, T.B., Ziou, D.,
Contextual and non-contextual performance evaluation of edge detectors,
PRL(21), No. 9, August 2000, pp. 805-816. 0008
BibRef

Medina Carnicer, R., Madrid Cuevas, F.J., Fernández García, N.L., Carmona Poyato, A.,
Evaluation of global thresholding techniques in non-contextual edge detection,
PRL(26), No. 10, 15 July 2005, pp. 1423-1434.
WWW Link. 0506
BibRef

Medina-Carnicer, R., Madrid-Cuevas, F.J., Munoz-Salinas, R., Carmona-Poyato, A.,
Solving the process of hysteresis without determining the optimal thresholds,
PR(43), No. 4, April 2010, pp. 1224-1232.
Elsevier DOI 1002
Hysteresis; Thresholding; Edge detection BibRef

Chabrier, S.[Sébastien], Laurent, H.[Hélène], Rosenberger, C.[Christophe], Emile, B.[Bruno],
Comparative Study of Contour Detection Evaluation Criteria Based on Dissimilarity Measures,
JIVP(2008), No. 2008, pp. xx-yy.
DOI Link 0804
BibRef

Benezeth, Y.[Yannick], Hemery, B.[Baptiste], Laurent, H.[Hélène], Emile, B.[Bruno], Rosenberger, C.[Christophe],
Evaluation of Human Detection Algorithms in Image Sequences,
ACIVS10(II: 121-130).
Springer DOI 1012
BibRef

Benezeth, Y., Jodoin, P.M., Emile, B., Laurent, H., Rosenberger, C.,
Human Detection with a Multi-sensors Stereovision System,
ICISP10(228-235).
Springer DOI 1006
BibRef

Benezeth, Y.[Yannick], Emile, B.[Bruno], Laurent, H.[Hélène], Rosenberger, C.[Christophe],
A Real Time Human Detection System Based on Far Infrared Vision,
ICISP08(76-84).
Springer DOI 0807
BibRef

Hemery, B.[Baptiste], Laurent, H.[Hélène], Rosenberger, C.[Christophe], and Emile, B.[Bruno],
Evaluation Protocol for Localization Metrics: Application to a Comparative Study,
ICISP08(273-280).
Springer DOI 0807
BibRef


Qiang, S.[Song], Liu, L.X.[Ling-Xia],
Compare between Several Linear Image Edge Detection Algorithm,
ICMV09(259-263).
IEEE DOI 0912
BibRef

Mazumdar, M., Sinha, B.K., and Li, C.C.,
A Comparison of Several Estimators of Edge Point in Noisy Digital Data Across a Step Edge,
CVPR85(27-33). (Univ. of Pittsburgh) Compare: moments, maximum likelihood, and Bayes for edge detection in 1-D data. BibRef 8500

Chapter on Edge Detection and Analysis, Lines, Segments, Curves, Corners, Hough Transform continues in
Boundaries - More Than Simple Edge Points, Linking .


Last update:Nov 11, 2017 at 13:31:57