Quantitative consensus of supervised learners for diffuse lung parenchymal HRCT patterns

Sushravya Raghunath; Srinivasan Rajagopalan; Ronald A. Karwoski; Brian J. Bartholmai; Richard A. Robb

doi:10.1117/12.2008110

Quantitative consensus of supervised learners for diffuse lung parenchymal HRCT patterns

Sushravya Raghunath, Srinivasan Rajagopalan, Ronald A. Karwoski, Brian J. Bartholmai, Richard A. Robb

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

1 Scopus citations

Abstract

Automated lung parenchymal classification usually relies on supervised learning of expert chosen regions representative of the visually differentiable HRCT patterns specific to different pathologies (eg. emphysema, ground glass, honey combing, reticular and normal). Considering the elusiveness of a single most discriminating similarity measure, a plurality of weak learners can be combined to improve the machine learnability. Though a number of quantitative combination strategies exist, their efficacy is data and domain dependent. In this paper, we investigate multiple (N=12) quantitative consensus approaches to combine the clusters obtained with multiple (n=33) probability density-based similarity measures. Our study shows that hypergraph based meta-clustering and probabilistic clustering provides optimal expert-metric agreement.

Original language	English (US)
Title of host publication	Medical Imaging 2013
Subtitle of host publication	Computer-Aided Diagnosis
DOIs	https://doi.org/10.1117/12.2008110
State	Published - Jun 5 2013
Event	Medical Imaging 2013: Computer-Aided Diagnosis - Lake Buena Vista, FL, United States Duration: Feb 12 2013 → Feb 14 2013

Publication series

Name	Proceedings of SPIE - The International Society for Optical Engineering
Volume	8670
ISSN (Print)	0277-786X

Other

Other	Medical Imaging 2013: Computer-Aided Diagnosis
Country/Territory	United States
City	Lake Buena Vista, FL
Period	2/12/13 → 2/14/13

Keywords

Cluster ensemble
Meta clustering

ASJC Scopus subject areas

Electronic, Optical and Magnetic Materials
Condensed Matter Physics
Computer Science Applications
Applied Mathematics
Electrical and Electronic Engineering

Access to Document

10.1117/12.2008110

Cite this

Quantitative consensus of supervised learners for diffuse lung parenchymal HRCT patterns. / Raghunath, Sushravya; Rajagopalan, Srinivasan; Karwoski, Ronald A. et al.
Medical Imaging 2013: Computer-Aided Diagnosis. 2013. 867037 (Proceedings of SPIE - The International Society for Optical Engineering; Vol. 8670).

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

Raghunath, S, Rajagopalan, S, Karwoski, RA, Bartholmai, BJ & Robb, RA 2013, Quantitative consensus of supervised learners for diffuse lung parenchymal HRCT patterns. in Medical Imaging 2013: Computer-Aided Diagnosis., 867037, Proceedings of SPIE - The International Society for Optical Engineering, vol. 8670, Medical Imaging 2013: Computer-Aided Diagnosis, Lake Buena Vista, FL, United States, 2/12/13. https://doi.org/10.1117/12.2008110

@inproceedings{3011ad6bcf8f4345956ed516960c7486,

title = "Quantitative consensus of supervised learners for diffuse lung parenchymal HRCT patterns",

abstract = "Automated lung parenchymal classification usually relies on supervised learning of expert chosen regions representative of the visually differentiable HRCT patterns specific to different pathologies (eg. emphysema, ground glass, honey combing, reticular and normal). Considering the elusiveness of a single most discriminating similarity measure, a plurality of weak learners can be combined to improve the machine learnability. Though a number of quantitative combination strategies exist, their efficacy is data and domain dependent. In this paper, we investigate multiple (N=12) quantitative consensus approaches to combine the clusters obtained with multiple (n=33) probability density-based similarity measures. Our study shows that hypergraph based meta-clustering and probabilistic clustering provides optimal expert-metric agreement.",

keywords = "Cluster ensemble, Meta clustering",

author = "Sushravya Raghunath and Srinivasan Rajagopalan and Karwoski, {Ronald A.} and Bartholmai, {Brian J.} and Robb, {Richard A.}",

year = "2013",

month = jun,

day = "5",

doi = "10.1117/12.2008110",

language = "English (US)",

isbn = "9780819494443",

series = "Proceedings of SPIE - The International Society for Optical Engineering",

booktitle = "Medical Imaging 2013",

note = "Medical Imaging 2013: Computer-Aided Diagnosis ; Conference date: 12-02-2013 Through 14-02-2013",

}

TY - GEN

T1 - Quantitative consensus of supervised learners for diffuse lung parenchymal HRCT patterns

AU - Raghunath, Sushravya

AU - Rajagopalan, Srinivasan

AU - Karwoski, Ronald A.

AU - Bartholmai, Brian J.

AU - Robb, Richard A.

PY - 2013/6/5

Y1 - 2013/6/5

N2 - Automated lung parenchymal classification usually relies on supervised learning of expert chosen regions representative of the visually differentiable HRCT patterns specific to different pathologies (eg. emphysema, ground glass, honey combing, reticular and normal). Considering the elusiveness of a single most discriminating similarity measure, a plurality of weak learners can be combined to improve the machine learnability. Though a number of quantitative combination strategies exist, their efficacy is data and domain dependent. In this paper, we investigate multiple (N=12) quantitative consensus approaches to combine the clusters obtained with multiple (n=33) probability density-based similarity measures. Our study shows that hypergraph based meta-clustering and probabilistic clustering provides optimal expert-metric agreement.

AB - Automated lung parenchymal classification usually relies on supervised learning of expert chosen regions representative of the visually differentiable HRCT patterns specific to different pathologies (eg. emphysema, ground glass, honey combing, reticular and normal). Considering the elusiveness of a single most discriminating similarity measure, a plurality of weak learners can be combined to improve the machine learnability. Though a number of quantitative combination strategies exist, their efficacy is data and domain dependent. In this paper, we investigate multiple (N=12) quantitative consensus approaches to combine the clusters obtained with multiple (n=33) probability density-based similarity measures. Our study shows that hypergraph based meta-clustering and probabilistic clustering provides optimal expert-metric agreement.

KW - Cluster ensemble

KW - Meta clustering

UR - http://www.scopus.com/inward/record.url?scp=84878388409&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84878388409&partnerID=8YFLogxK

U2 - 10.1117/12.2008110

DO - 10.1117/12.2008110

M3 - Conference contribution

AN - SCOPUS:84878388409

SN - 9780819494443

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Medical Imaging 2013

T2 - Medical Imaging 2013: Computer-Aided Diagnosis

Y2 - 12 February 2013 through 14 February 2013

ER -

Quantitative consensus of supervised learners for diffuse lung parenchymal HRCT patterns

Abstract

Publication series

Other

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this