Non-parametric bootstrap ensembles for detection of tumor lesions

Mireya Diaz, Jonnagadda S Rao

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

We propose and assess a set of non-parametric ensembles, including bagging and boosting schemes, to recognize tumors in digital mammograms. Different approaches were examined as candidates for the two major components of the bagging ensembles, three spatial resampling schemes (residuals, centers and standardized centers), and four combination criteria (at least one, majority vote, top 25% models, and false discovery rate). A conversion to a classification problem prior to aggregation was employed for the boosting ensemble. The ensembles were compared at the lesion level against a single expert, and to a set of Markov Random Field (MRF) models in real images using three different criteria. The performance of the ensembles depended on its components, particularly the combination, with at least one and top 25% models offering a greater detection power independently of the type of lesion, and of the booststrapping scheme in a lesser degree. The ensembles were comparable in performance to MRFs in the unsupervised recognition of patterns exhibiting spatial structure.

Original languageEnglish
Pages (from-to)2273-2283
Number of pages11
JournalPattern Recognition Letters
Volume28
Issue number16
DOIs
StatePublished - Dec 1 2007
Externally publishedYes

Fingerprint

Tumors
Agglomeration

Keywords

  • Ensembles
  • Image analysis
  • Markov Random Fields
  • Spatial correlation
  • Statistical pattern recognition
  • Unsupervised training

ASJC Scopus subject areas

  • Computer Vision and Pattern Recognition
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Non-parametric bootstrap ensembles for detection of tumor lesions. / Diaz, Mireya; Rao, Jonnagadda S.

In: Pattern Recognition Letters, Vol. 28, No. 16, 01.12.2007, p. 2273-2283.

Research output: Contribution to journalArticle

@article{4ab32ee47f7e4fc8a2f7e595164322dc,
title = "Non-parametric bootstrap ensembles for detection of tumor lesions",
abstract = "We propose and assess a set of non-parametric ensembles, including bagging and boosting schemes, to recognize tumors in digital mammograms. Different approaches were examined as candidates for the two major components of the bagging ensembles, three spatial resampling schemes (residuals, centers and standardized centers), and four combination criteria (at least one, majority vote, top 25{\%} models, and false discovery rate). A conversion to a classification problem prior to aggregation was employed for the boosting ensemble. The ensembles were compared at the lesion level against a single expert, and to a set of Markov Random Field (MRF) models in real images using three different criteria. The performance of the ensembles depended on its components, particularly the combination, with at least one and top 25{\%} models offering a greater detection power independently of the type of lesion, and of the booststrapping scheme in a lesser degree. The ensembles were comparable in performance to MRFs in the unsupervised recognition of patterns exhibiting spatial structure.",
keywords = "Ensembles, Image analysis, Markov Random Fields, Spatial correlation, Statistical pattern recognition, Unsupervised training",
author = "Mireya Diaz and Rao, {Jonnagadda S}",
year = "2007",
month = "12",
day = "1",
doi = "10.1016/j.patrec.2007.07.009",
language = "English",
volume = "28",
pages = "2273--2283",
journal = "Pattern Recognition Letters",
issn = "0167-8655",
publisher = "Elsevier",
number = "16",

}

TY - JOUR

T1 - Non-parametric bootstrap ensembles for detection of tumor lesions

AU - Diaz, Mireya

AU - Rao, Jonnagadda S

PY - 2007/12/1

Y1 - 2007/12/1

N2 - We propose and assess a set of non-parametric ensembles, including bagging and boosting schemes, to recognize tumors in digital mammograms. Different approaches were examined as candidates for the two major components of the bagging ensembles, three spatial resampling schemes (residuals, centers and standardized centers), and four combination criteria (at least one, majority vote, top 25% models, and false discovery rate). A conversion to a classification problem prior to aggregation was employed for the boosting ensemble. The ensembles were compared at the lesion level against a single expert, and to a set of Markov Random Field (MRF) models in real images using three different criteria. The performance of the ensembles depended on its components, particularly the combination, with at least one and top 25% models offering a greater detection power independently of the type of lesion, and of the booststrapping scheme in a lesser degree. The ensembles were comparable in performance to MRFs in the unsupervised recognition of patterns exhibiting spatial structure.

AB - We propose and assess a set of non-parametric ensembles, including bagging and boosting schemes, to recognize tumors in digital mammograms. Different approaches were examined as candidates for the two major components of the bagging ensembles, three spatial resampling schemes (residuals, centers and standardized centers), and four combination criteria (at least one, majority vote, top 25% models, and false discovery rate). A conversion to a classification problem prior to aggregation was employed for the boosting ensemble. The ensembles were compared at the lesion level against a single expert, and to a set of Markov Random Field (MRF) models in real images using three different criteria. The performance of the ensembles depended on its components, particularly the combination, with at least one and top 25% models offering a greater detection power independently of the type of lesion, and of the booststrapping scheme in a lesser degree. The ensembles were comparable in performance to MRFs in the unsupervised recognition of patterns exhibiting spatial structure.

KW - Ensembles

KW - Image analysis

KW - Markov Random Fields

KW - Spatial correlation

KW - Statistical pattern recognition

KW - Unsupervised training

UR - http://www.scopus.com/inward/record.url?scp=35148814875&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=35148814875&partnerID=8YFLogxK

U2 - 10.1016/j.patrec.2007.07.009

DO - 10.1016/j.patrec.2007.07.009

M3 - Article

AN - SCOPUS:35148814875

VL - 28

SP - 2273

EP - 2283

JO - Pattern Recognition Letters

JF - Pattern Recognition Letters

SN - 0167-8655

IS - 16

ER -