Global and local multi-valued dissimilarity-based classification: Application to computer-aided detection of tuberculosis

Yulia Arzhaeva; Laurens Hogeweg; Pim A de Jong; Max A Viergever; Bram van Ginneken

doi:10.1007/978-3-642-04271-3_88

Global and local multi-valued dissimilarity-based classification: Application to computer-aided detection of tuberculosis

Yulia Arzhaeva^*
, Laurens Hogeweg
, Pim A de Jong
, Max A Viergever
, Bram van Ginneken

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

5 Citations (Scopus)

Abstract

In many applications of computer-aided detection (CAD) it is not possible to precisely localize lesions or affected areas in images that are known to be abnormal. In this paper a novel approach to computer-aided detection is presented that can deal effectively with such weakly labeled data. Our approach is based on multi-valued dissimilarity measures that retain more information about underlying local image features than single-valued dissimilarities. We show how this approach can be extended by applying it locally as well as globally, and by merging the local and global classification results into an overall opinion about the image to be classified. The framework is applied to the detection of tuberculosis (TB) in chest radiographs. This is the first study to apply a CAD system to a large database of digital chest radiographs obtained from a TB screening program, including normal cases, suspect cases and cases with proven TB. The global dissimilarity approach achieved an area under the ROC curve of 0.81. The combination of local and global classifications increased this value to 0.83.

Original language	English
Title of host publication	Medical Image Computing and Computer-Assisted Intervention - MICCAI2009 - 12th International Conference, Proceedings
Editors	Guang-Zhong Yang, David Hawkes, Daniel Rueckert, Alison Noble, Chris Taylor
Pages	724-731
Number of pages	8
Edition	PART 2
ISBN (Electronic)	978-3-642-04271-3
DOIs	https://doi.org/10.1007/978-3-642-04271-3_88
Publication status	Published - 1 Dec 2009
Event	12th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2009 - London, United Kingdom Duration: 20 Sept 2009 → 24 Sept 2009

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 2
Volume	5762 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Conference

Conference	12th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2009
Country/Territory	United Kingdom
City	London
Period	20/09/09 → 24/09/09

Keywords

Algorithms
Humans
Lung
Pattern Recognition, Automated
Radiographic Image Enhancement
Radiographic Image Interpretation, Computer-Assisted
Radiography, Thoracic
Reproducibility of Results
Sensitivity and Specificity
Tuberculosis
Journal Article

Access to Document

10.1007/978-3-642-04271-3_88

Cite this

Arzhaeva, Y., Hogeweg, L., de Jong, P. A., Viergever, M. A., & van Ginneken, B. (2009). Global and local multi-valued dissimilarity-based classification: Application to computer-aided detection of tuberculosis. In G.-Z. Yang, D. Hawkes, D. Rueckert, A. Noble, & C. Taylor (Eds.), Medical Image Computing and Computer-Assisted Intervention - MICCAI2009 - 12th International Conference, Proceedings (PART 2 ed., pp. 724-731). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5762 LNCS, No. PART 2). https://doi.org/10.1007/978-3-642-04271-3_88

Arzhaeva, Yulia ; Hogeweg, Laurens ; de Jong, Pim A et al. / Global and local multi-valued dissimilarity-based classification : Application to computer-aided detection of tuberculosis. Medical Image Computing and Computer-Assisted Intervention - MICCAI2009 - 12th International Conference, Proceedings. editor / Guang-Zhong Yang ; David Hawkes ; Daniel Rueckert ; Alison Noble ; Chris Taylor. PART 2. ed. 2009. pp. 724-731 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{db9c5e7ef7cf44338227df1ccc695b90,

title = "Global and local multi-valued dissimilarity-based classification: Application to computer-aided detection of tuberculosis",

abstract = "In many applications of computer-aided detection (CAD) it is not possible to precisely localize lesions or affected areas in images that are known to be abnormal. In this paper a novel approach to computer-aided detection is presented that can deal effectively with such weakly labeled data. Our approach is based on multi-valued dissimilarity measures that retain more information about underlying local image features than single-valued dissimilarities. We show how this approach can be extended by applying it locally as well as globally, and by merging the local and global classification results into an overall opinion about the image to be classified. The framework is applied to the detection of tuberculosis (TB) in chest radiographs. This is the first study to apply a CAD system to a large database of digital chest radiographs obtained from a TB screening program, including normal cases, suspect cases and cases with proven TB. The global dissimilarity approach achieved an area under the ROC curve of 0.81. The combination of local and global classifications increased this value to 0.83.",

keywords = "Algorithms, Humans, Lung, Pattern Recognition, Automated, Radiographic Image Enhancement, Radiographic Image Interpretation, Computer-Assisted, Radiography, Thoracic, Reproducibility of Results, Sensitivity and Specificity, Tuberculosis, Journal Article",

author = "Yulia Arzhaeva and Laurens Hogeweg and \{de Jong\}, \{Pim A\} and Viergever, \{Max A\} and \{van Ginneken\}, Bram",

year = "2009",

month = dec,

day = "1",

doi = "10.1007/978-3-642-04271-3\_88",

language = "English",

isbn = "3642042708",

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

number = "PART 2",

pages = "724--731",

editor = "Guang-Zhong Yang and Hawkes, \{David \} and Rueckert, \{Daniel \} and Noble, \{Alison \} and Taylor, \{Chris \}",

booktitle = "Medical Image Computing and Computer-Assisted Intervention - MICCAI2009 - 12th International Conference, Proceedings",

edition = "PART 2",

note = "12th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2009 ; Conference date: 20-09-2009 Through 24-09-2009",

}

Arzhaeva, Y, Hogeweg, L, de Jong, PA , Viergever, MA & van Ginneken, B 2009, Global and local multi-valued dissimilarity-based classification: Application to computer-aided detection of tuberculosis. in G-Z Yang, D Hawkes, D Rueckert, A Noble & C Taylor (eds), Medical Image Computing and Computer-Assisted Intervention - MICCAI2009 - 12th International Conference, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 2, vol. 5762 LNCS, pp. 724-731, 12th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2009, London, United Kingdom, 20/09/09. https://doi.org/10.1007/978-3-642-04271-3_88

Global and local multi-valued dissimilarity-based classification: Application to computer-aided detection of tuberculosis. / Arzhaeva, Yulia; Hogeweg, Laurens; de Jong, Pim A et al.
Medical Image Computing and Computer-Assisted Intervention - MICCAI2009 - 12th International Conference, Proceedings. ed. / Guang-Zhong Yang; David Hawkes; Daniel Rueckert; Alison Noble; Chris Taylor. PART 2. ed. 2009. p. 724-731 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 5762 LNCS, No. PART 2).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution › Academic › peer-review

TY - GEN

T1 - Global and local multi-valued dissimilarity-based classification

T2 - 12th International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2009

AU - Arzhaeva, Yulia

AU - Hogeweg, Laurens

AU - de Jong, Pim A

AU - Viergever, Max A

AU - van Ginneken, Bram

PY - 2009/12/1

Y1 - 2009/12/1

N2 - In many applications of computer-aided detection (CAD) it is not possible to precisely localize lesions or affected areas in images that are known to be abnormal. In this paper a novel approach to computer-aided detection is presented that can deal effectively with such weakly labeled data. Our approach is based on multi-valued dissimilarity measures that retain more information about underlying local image features than single-valued dissimilarities. We show how this approach can be extended by applying it locally as well as globally, and by merging the local and global classification results into an overall opinion about the image to be classified. The framework is applied to the detection of tuberculosis (TB) in chest radiographs. This is the first study to apply a CAD system to a large database of digital chest radiographs obtained from a TB screening program, including normal cases, suspect cases and cases with proven TB. The global dissimilarity approach achieved an area under the ROC curve of 0.81. The combination of local and global classifications increased this value to 0.83.

AB - In many applications of computer-aided detection (CAD) it is not possible to precisely localize lesions or affected areas in images that are known to be abnormal. In this paper a novel approach to computer-aided detection is presented that can deal effectively with such weakly labeled data. Our approach is based on multi-valued dissimilarity measures that retain more information about underlying local image features than single-valued dissimilarities. We show how this approach can be extended by applying it locally as well as globally, and by merging the local and global classification results into an overall opinion about the image to be classified. The framework is applied to the detection of tuberculosis (TB) in chest radiographs. This is the first study to apply a CAD system to a large database of digital chest radiographs obtained from a TB screening program, including normal cases, suspect cases and cases with proven TB. The global dissimilarity approach achieved an area under the ROC curve of 0.81. The combination of local and global classifications increased this value to 0.83.

KW - Algorithms

KW - Humans

KW - Lung

KW - Pattern Recognition, Automated

KW - Radiographic Image Enhancement

KW - Radiographic Image Interpretation, Computer-Assisted

KW - Radiography, Thoracic

KW - Reproducibility of Results

KW - Sensitivity and Specificity

KW - Tuberculosis

KW - Journal Article

UR - https://www.scopus.com/pages/publications/84855398334

U2 - 10.1007/978-3-642-04271-3_88

DO - 10.1007/978-3-642-04271-3_88

M3 - Conference contribution

C2 - 20426176

SN - 3642042708

SN - 9783642042706

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

SP - 724

EP - 731

BT - Medical Image Computing and Computer-Assisted Intervention - MICCAI2009 - 12th International Conference, Proceedings

A2 - Yang, Guang-Zhong

A2 - Hawkes, David

A2 - Rueckert, Daniel

A2 - Noble, Alison

A2 - Taylor, Chris

Y2 - 20 September 2009 through 24 September 2009

ER -

Arzhaeva Y, Hogeweg L, de Jong PA , Viergever MA, van Ginneken B. Global and local multi-valued dissimilarity-based classification: Application to computer-aided detection of tuberculosis. In Yang GZ, Hawkes D, Rueckert D, Noble A, Taylor C, editors, Medical Image Computing and Computer-Assisted Intervention - MICCAI2009 - 12th International Conference, Proceedings. PART 2 ed. 2009. p. 724-731. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). doi: 10.1007/978-3-642-04271-3_88

Global and local multi-valued dissimilarity-based classification: Application to computer-aided detection of tuberculosis

Abstract

Publication series

Conference

Keywords

Access to Document

Other files and links

Fingerprint

Cite this