Image segmentation with implicit color standardization using cascaded EM

Detection of myelodysplastic syndromes

James Monaco, Phil Raess, Ronak Chawla, Adam Bagg, Mitchell Weiss, John Choi, Anant Madabhushi

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Color nonstandardness - the propensity for similar objects to exhibit different color properties across images - poses a significant problem in the computerized analysis of histopathology. Though many papers propose means for improving color constancy, the vast majority assume image formation via reflective light instead of light transmission as in microscopy, and thus are inappropriate for histological analysis. In this work, we present a novel Bayesian color segmentation algorithm for histological images that is highly robust to color nonstandardness; this algorithm employs a unique instantiation of the expectation maximization (EM) algorithm to dynamically estimate - for each individual image - the probability density functions (mixtures of gamma and von Mises distributions) that describe the colors of salient objects. To validate our segmentation scheme, we employ it as part of a computerized system to detect myelodysplastic syndromes (MDS) on bone marrow specimens. Qualitative anecdotal evidence suggests that biopsies of MDS exhibit abnormalities in the arrangement of erythroid precursors (immature red blood cells). Herein, we confirm and quantify this phenomenon, using it to discriminate MDS from normal tissue: over a dataset of 53 representative regions selected from 18 patients, our classification system correctly discriminates MDS from normal tissue with an accuracy of 85% and an area under the receiver operator characteristic curve of 0.8803.

Original languageEnglish (US)
Title of host publication2012 9th IEEE International Symposium on Biomedical Imaging
Subtitle of host publicationFrom Nano to Macro, ISBI 2012 - Proceedings
Pages740-743
Number of pages4
DOIs
StatePublished - Aug 15 2012
Externally publishedYes
Event2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012 - Barcelona, Spain
Duration: May 2 2012May 5 2012

Publication series

NameProceedings - International Symposium on Biomedical Imaging
ISSN (Print)1945-7928
ISSN (Electronic)1945-8452

Other

Other2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012
CountrySpain
CityBarcelona
Period5/2/125/5/12

Fingerprint

Myelodysplastic Syndromes
Image segmentation
Standardization
Color
Tissue
Light
Biopsy
Light transmission
Probability density function
Microscopy
Microscopic examination
Bone
Image processing
Blood
Erythrocytes
Bone Marrow
Cells

ASJC Scopus subject areas

  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging

Cite this

Monaco, J., Raess, P., Chawla, R., Bagg, A., Weiss, M., Choi, J., & Madabhushi, A. (2012). Image segmentation with implicit color standardization using cascaded EM: Detection of myelodysplastic syndromes. In 2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012 - Proceedings (pp. 740-743). [6235654] (Proceedings - International Symposium on Biomedical Imaging). https://doi.org/10.1109/ISBI.2012.6235654

Image segmentation with implicit color standardization using cascaded EM : Detection of myelodysplastic syndromes. / Monaco, James; Raess, Phil; Chawla, Ronak; Bagg, Adam; Weiss, Mitchell; Choi, John; Madabhushi, Anant.

2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012 - Proceedings. 2012. p. 740-743 6235654 (Proceedings - International Symposium on Biomedical Imaging).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Monaco, J, Raess, P, Chawla, R, Bagg, A, Weiss, M, Choi, J & Madabhushi, A 2012, Image segmentation with implicit color standardization using cascaded EM: Detection of myelodysplastic syndromes. in 2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012 - Proceedings., 6235654, Proceedings - International Symposium on Biomedical Imaging, pp. 740-743, 2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012, Barcelona, Spain, 5/2/12. https://doi.org/10.1109/ISBI.2012.6235654
Monaco J, Raess P, Chawla R, Bagg A, Weiss M, Choi J et al. Image segmentation with implicit color standardization using cascaded EM: Detection of myelodysplastic syndromes. In 2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012 - Proceedings. 2012. p. 740-743. 6235654. (Proceedings - International Symposium on Biomedical Imaging). https://doi.org/10.1109/ISBI.2012.6235654
Monaco, James ; Raess, Phil ; Chawla, Ronak ; Bagg, Adam ; Weiss, Mitchell ; Choi, John ; Madabhushi, Anant. / Image segmentation with implicit color standardization using cascaded EM : Detection of myelodysplastic syndromes. 2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012 - Proceedings. 2012. pp. 740-743 (Proceedings - International Symposium on Biomedical Imaging).
@inproceedings{7d797d6183e845399117dee74576b6d6,
title = "Image segmentation with implicit color standardization using cascaded EM: Detection of myelodysplastic syndromes",
abstract = "Color nonstandardness - the propensity for similar objects to exhibit different color properties across images - poses a significant problem in the computerized analysis of histopathology. Though many papers propose means for improving color constancy, the vast majority assume image formation via reflective light instead of light transmission as in microscopy, and thus are inappropriate for histological analysis. In this work, we present a novel Bayesian color segmentation algorithm for histological images that is highly robust to color nonstandardness; this algorithm employs a unique instantiation of the expectation maximization (EM) algorithm to dynamically estimate - for each individual image - the probability density functions (mixtures of gamma and von Mises distributions) that describe the colors of salient objects. To validate our segmentation scheme, we employ it as part of a computerized system to detect myelodysplastic syndromes (MDS) on bone marrow specimens. Qualitative anecdotal evidence suggests that biopsies of MDS exhibit abnormalities in the arrangement of erythroid precursors (immature red blood cells). Herein, we confirm and quantify this phenomenon, using it to discriminate MDS from normal tissue: over a dataset of 53 representative regions selected from 18 patients, our classification system correctly discriminates MDS from normal tissue with an accuracy of 85{\%} and an area under the receiver operator characteristic curve of 0.8803.",
author = "James Monaco and Phil Raess and Ronak Chawla and Adam Bagg and Mitchell Weiss and John Choi and Anant Madabhushi",
year = "2012",
month = "8",
day = "15",
doi = "10.1109/ISBI.2012.6235654",
language = "English (US)",
isbn = "9781457718588",
series = "Proceedings - International Symposium on Biomedical Imaging",
pages = "740--743",
booktitle = "2012 9th IEEE International Symposium on Biomedical Imaging",

}

TY - GEN

T1 - Image segmentation with implicit color standardization using cascaded EM

T2 - Detection of myelodysplastic syndromes

AU - Monaco, James

AU - Raess, Phil

AU - Chawla, Ronak

AU - Bagg, Adam

AU - Weiss, Mitchell

AU - Choi, John

AU - Madabhushi, Anant

PY - 2012/8/15

Y1 - 2012/8/15

N2 - Color nonstandardness - the propensity for similar objects to exhibit different color properties across images - poses a significant problem in the computerized analysis of histopathology. Though many papers propose means for improving color constancy, the vast majority assume image formation via reflective light instead of light transmission as in microscopy, and thus are inappropriate for histological analysis. In this work, we present a novel Bayesian color segmentation algorithm for histological images that is highly robust to color nonstandardness; this algorithm employs a unique instantiation of the expectation maximization (EM) algorithm to dynamically estimate - for each individual image - the probability density functions (mixtures of gamma and von Mises distributions) that describe the colors of salient objects. To validate our segmentation scheme, we employ it as part of a computerized system to detect myelodysplastic syndromes (MDS) on bone marrow specimens. Qualitative anecdotal evidence suggests that biopsies of MDS exhibit abnormalities in the arrangement of erythroid precursors (immature red blood cells). Herein, we confirm and quantify this phenomenon, using it to discriminate MDS from normal tissue: over a dataset of 53 representative regions selected from 18 patients, our classification system correctly discriminates MDS from normal tissue with an accuracy of 85% and an area under the receiver operator characteristic curve of 0.8803.

AB - Color nonstandardness - the propensity for similar objects to exhibit different color properties across images - poses a significant problem in the computerized analysis of histopathology. Though many papers propose means for improving color constancy, the vast majority assume image formation via reflective light instead of light transmission as in microscopy, and thus are inappropriate for histological analysis. In this work, we present a novel Bayesian color segmentation algorithm for histological images that is highly robust to color nonstandardness; this algorithm employs a unique instantiation of the expectation maximization (EM) algorithm to dynamically estimate - for each individual image - the probability density functions (mixtures of gamma and von Mises distributions) that describe the colors of salient objects. To validate our segmentation scheme, we employ it as part of a computerized system to detect myelodysplastic syndromes (MDS) on bone marrow specimens. Qualitative anecdotal evidence suggests that biopsies of MDS exhibit abnormalities in the arrangement of erythroid precursors (immature red blood cells). Herein, we confirm and quantify this phenomenon, using it to discriminate MDS from normal tissue: over a dataset of 53 representative regions selected from 18 patients, our classification system correctly discriminates MDS from normal tissue with an accuracy of 85% and an area under the receiver operator characteristic curve of 0.8803.

UR - http://www.scopus.com/inward/record.url?scp=84864833472&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84864833472&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2012.6235654

DO - 10.1109/ISBI.2012.6235654

M3 - Conference contribution

SN - 9781457718588

T3 - Proceedings - International Symposium on Biomedical Imaging

SP - 740

EP - 743

BT - 2012 9th IEEE International Symposium on Biomedical Imaging

ER -