A tale of two classifiers: SNoW vs. SVM in visual recognition

Ming Hsuan Yang; Dan Roth; Narendra Ahuja

doi:10.1007/3-540-47979-1_46

A tale of two classifiers: SNoW vs. SVM in visual recognition

Ming Hsuan Yang, Dan Roth, Narendra Ahuja

College of Computing

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

9 Citations (Scopus)

Abstract

Numerous statistical learning methods have been developed for visual recognition tasks. Few attempts, however, have been made to address theoretical issues, and in particular, study the suitability of different learning algorithms for visual recognition. Large margin classifiers, such as SNoW and SVM, have recently demonstrated their success in object detection and recognition. In this paper, we present a theoretical account of these two learning approaches, and their suitability to visual recognition. Using tools from computational learning theory, we show that the main difference between the generalization bounds of SVM and SNoW depends on the properties of the data. We argue that learning problems in the visual domain have sparseness characteristics and exhibit them by analyzing data taken from face detection experiments. Experimental results exhibit good generalization and robustness properties of the SNoW-based method, and conform to the theoretical analysis.

Original language	English
Title of host publication	Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings
Editors	Anders Heyden, Gunnar Sparr, Mads Nielsen, Peter Johansen
Publisher	Springer Verlag
Pages	685-699
Number of pages	15
ISBN (Electronic)	9783540437482
DOIs	https://doi.org/10.1007/3-540-47979-1_46
Publication status	Published - 2002
Event	7th European Conference on Computer Vision, ECCV 2002 - Copenhagen, Denmark Duration: 2002 May 28 → 2002 May 31

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	2353
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	7th European Conference on Computer Vision, ECCV 2002
Country/Territory	Denmark
City	Copenhagen
Period	02/5/28 → 02/5/31

Bibliographical note

Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 2002.

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
Computer Science(all)

Access to Document

10.1007/3-540-47979-1_46

Cite this

Yang, M. H., Roth, D., & Ahuja, N. (2002). A tale of two classifiers: SNoW vs. SVM in visual recognition. In A. Heyden, G. Sparr, M. Nielsen, & P. Johansen (Eds.), Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings (pp. 685-699). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2353). Springer Verlag. https://doi.org/10.1007/3-540-47979-1_46

Yang, Ming Hsuan ; Roth, Dan ; Ahuja, Narendra. / A tale of two classifiers : SNoW vs. SVM in visual recognition. Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings. editor / Anders Heyden ; Gunnar Sparr ; Mads Nielsen ; Peter Johansen. Springer Verlag, 2002. pp. 685-699 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{a402aadb7bce4531a602ba740b22c008,

title = "A tale of two classifiers: SNoW vs. SVM in visual recognition",

abstract = "Numerous statistical learning methods have been developed for visual recognition tasks. Few attempts, however, have been made to address theoretical issues, and in particular, study the suitability of different learning algorithms for visual recognition. Large margin classifiers, such as SNoW and SVM, have recently demonstrated their success in object detection and recognition. In this paper, we present a theoretical account of these two learning approaches, and their suitability to visual recognition. Using tools from computational learning theory, we show that the main difference between the generalization bounds of SVM and SNoW depends on the properties of the data. We argue that learning problems in the visual domain have sparseness characteristics and exhibit them by analyzing data taken from face detection experiments. Experimental results exhibit good generalization and robustness properties of the SNoW-based method, and conform to the theoretical analysis.",

author = "Yang, {Ming Hsuan} and Dan Roth and Narendra Ahuja",

note = "Publisher Copyright: {\textcopyright} Springer-Verlag Berlin Heidelberg 2002.; 7th European Conference on Computer Vision, ECCV 2002 ; Conference date: 28-05-2002 Through 31-05-2002",

year = "2002",

doi = "10.1007/3-540-47979-1_46",

language = "English",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Verlag",

pages = "685--699",

editor = "Anders Heyden and Gunnar Sparr and Mads Nielsen and Peter Johansen",

booktitle = "Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings",

address = "Germany",

}

Yang, MH, Roth, D & Ahuja, N 2002, A tale of two classifiers: SNoW vs. SVM in visual recognition. in A Heyden, G Sparr, M Nielsen & P Johansen (eds), Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 2353, Springer Verlag, pp. 685-699, 7th European Conference on Computer Vision, ECCV 2002, Copenhagen, Denmark, 02/5/28. https://doi.org/10.1007/3-540-47979-1_46

A tale of two classifiers : SNoW vs. SVM in visual recognition. / Yang, Ming Hsuan; Roth, Dan; Ahuja, Narendra.

Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings. ed. / Anders Heyden; Gunnar Sparr; Mads Nielsen; Peter Johansen. Springer Verlag, 2002. p. 685-699 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 2353).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - A tale of two classifiers

T2 - 7th European Conference on Computer Vision, ECCV 2002

AU - Yang, Ming Hsuan

AU - Roth, Dan

AU - Ahuja, Narendra

PY - 2002

Y1 - 2002

N2 - Numerous statistical learning methods have been developed for visual recognition tasks. Few attempts, however, have been made to address theoretical issues, and in particular, study the suitability of different learning algorithms for visual recognition. Large margin classifiers, such as SNoW and SVM, have recently demonstrated their success in object detection and recognition. In this paper, we present a theoretical account of these two learning approaches, and their suitability to visual recognition. Using tools from computational learning theory, we show that the main difference between the generalization bounds of SVM and SNoW depends on the properties of the data. We argue that learning problems in the visual domain have sparseness characteristics and exhibit them by analyzing data taken from face detection experiments. Experimental results exhibit good generalization and robustness properties of the SNoW-based method, and conform to the theoretical analysis.

AB - Numerous statistical learning methods have been developed for visual recognition tasks. Few attempts, however, have been made to address theoretical issues, and in particular, study the suitability of different learning algorithms for visual recognition. Large margin classifiers, such as SNoW and SVM, have recently demonstrated their success in object detection and recognition. In this paper, we present a theoretical account of these two learning approaches, and their suitability to visual recognition. Using tools from computational learning theory, we show that the main difference between the generalization bounds of SVM and SNoW depends on the properties of the data. We argue that learning problems in the visual domain have sparseness characteristics and exhibit them by analyzing data taken from face detection experiments. Experimental results exhibit good generalization and robustness properties of the SNoW-based method, and conform to the theoretical analysis.

UR - http://www.scopus.com/inward/record.url?scp=84937539670&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84937539670&partnerID=8YFLogxK

U2 - 10.1007/3-540-47979-1_46

DO - 10.1007/3-540-47979-1_46

M3 - Conference contribution

AN - SCOPUS:84937539670

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 685

EP - 699

BT - Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings

A2 - Heyden, Anders

A2 - Sparr, Gunnar

A2 - Nielsen, Mads

A2 - Johansen, Peter

PB - Springer Verlag

Y2 - 28 May 2002 through 31 May 2002

ER -

Yang MH, Roth D, Ahuja N. A tale of two classifiers: SNoW vs. SVM in visual recognition. In Heyden A, Sparr G, Nielsen M, Johansen P, editors, Computer Vision - ECCV 2002 - 7th European Conference on Computer Vision, Proceedings. Springer Verlag. 2002. p. 685-699. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/3-540-47979-1_46

A tale of two classifiers: SNoW vs. SVM in visual recognition

Abstract

Publication series

Conference

Bibliographical note

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this