Methods for segmentation and classification of digital microscopy tissue images

Quoc Dang Vu, Simon Graham, Tahsin Kurc, Minh Nguyen Nhat To, Muhammad Shaban, Talha Qaiser, Navid Alemi Koohbanani, Syed Ali Khurram, Jayashree Kalpathy-Cramer, Tianhao Zhao, Rajarsi Gupta, Jin Tae Kwak, Nasir Rajpoot, Joel Saltz, Keyvan Farahani

Research output: Contribution to journalArticle

4 Citations (Scopus)

Abstract

High-resolution microscopy images of tissue specimens provide detailed information about the morphology of normal and diseased tissue. Image analysis of tissue morphology can help cancer researchers develop a better understanding of cancer biology. Segmentation of nuclei and classification of tissue images are two common tasks in tissue image analysis. Development of accurate and efficient algorithms for these tasks is a challenging problem because of the complexity of tissue morphology and tumor heterogeneity. In this paper we present two computer algorithms; one designed for segmentation of nuclei and the other for classification of whole slide tissue images. The segmentation algorithm implements a multiscale deep residual aggregation network to accurately segment nuclear material and then separate clumped nuclei into individual nuclei. The classification algorithm initially carries out patch-level classification via a deep learning method, then patch-level statistical and morphological features are used as input to a random forest regression model for whole slide image classification. The segmentation and classification algorithms were evaluated in the MICCAI 2017 Digital Pathology challenge. The segmentation algorithm achieved an accuracy score of 0.78. The classification algorithm achieved an accuracy score of 0.81. These scores were the highest in the challenge.

Original languageEnglish (US)
Article number53
JournalFrontiers in Bioengineering and Biotechnology
Volume7
Issue numberAPR
DOIs
StatePublished - Jan 1 2019
Externally publishedYes

Fingerprint

Microscopy
Microscopic examination
Tissue
Image analysis
Neoplasms
Image classification
Pathology
Tumors
Agglomeration
Research Personnel
Learning

Keywords

  • Classification
  • Digital pathology
  • Image analysis
  • Segmentation
  • Tissue images

ASJC Scopus subject areas

  • Biotechnology
  • Bioengineering
  • Histology
  • Biomedical Engineering

Cite this

Vu, Q. D., Graham, S., Kurc, T., To, M. N. N., Shaban, M., Qaiser, T., ... Farahani, K. (2019). Methods for segmentation and classification of digital microscopy tissue images. Frontiers in Bioengineering and Biotechnology, 7(APR), [53]. https://doi.org/10.3389/fbioe.2019.00053

Methods for segmentation and classification of digital microscopy tissue images. / Vu, Quoc Dang; Graham, Simon; Kurc, Tahsin; To, Minh Nguyen Nhat; Shaban, Muhammad; Qaiser, Talha; Koohbanani, Navid Alemi; Khurram, Syed Ali; Kalpathy-Cramer, Jayashree; Zhao, Tianhao; Gupta, Rajarsi; Kwak, Jin Tae; Rajpoot, Nasir; Saltz, Joel; Farahani, Keyvan.

In: Frontiers in Bioengineering and Biotechnology, Vol. 7, No. APR, 53, 01.01.2019.

Research output: Contribution to journalArticle

Vu, QD, Graham, S, Kurc, T, To, MNN, Shaban, M, Qaiser, T, Koohbanani, NA, Khurram, SA, Kalpathy-Cramer, J, Zhao, T, Gupta, R, Kwak, JT, Rajpoot, N, Saltz, J & Farahani, K 2019, 'Methods for segmentation and classification of digital microscopy tissue images', Frontiers in Bioengineering and Biotechnology, vol. 7, no. APR, 53. https://doi.org/10.3389/fbioe.2019.00053
Vu, Quoc Dang ; Graham, Simon ; Kurc, Tahsin ; To, Minh Nguyen Nhat ; Shaban, Muhammad ; Qaiser, Talha ; Koohbanani, Navid Alemi ; Khurram, Syed Ali ; Kalpathy-Cramer, Jayashree ; Zhao, Tianhao ; Gupta, Rajarsi ; Kwak, Jin Tae ; Rajpoot, Nasir ; Saltz, Joel ; Farahani, Keyvan. / Methods for segmentation and classification of digital microscopy tissue images. In: Frontiers in Bioengineering and Biotechnology. 2019 ; Vol. 7, No. APR.
@article{85e217cf6f4c4889ad792020687e4388,
title = "Methods for segmentation and classification of digital microscopy tissue images",
abstract = "High-resolution microscopy images of tissue specimens provide detailed information about the morphology of normal and diseased tissue. Image analysis of tissue morphology can help cancer researchers develop a better understanding of cancer biology. Segmentation of nuclei and classification of tissue images are two common tasks in tissue image analysis. Development of accurate and efficient algorithms for these tasks is a challenging problem because of the complexity of tissue morphology and tumor heterogeneity. In this paper we present two computer algorithms; one designed for segmentation of nuclei and the other for classification of whole slide tissue images. The segmentation algorithm implements a multiscale deep residual aggregation network to accurately segment nuclear material and then separate clumped nuclei into individual nuclei. The classification algorithm initially carries out patch-level classification via a deep learning method, then patch-level statistical and morphological features are used as input to a random forest regression model for whole slide image classification. The segmentation and classification algorithms were evaluated in the MICCAI 2017 Digital Pathology challenge. The segmentation algorithm achieved an accuracy score of 0.78. The classification algorithm achieved an accuracy score of 0.81. These scores were the highest in the challenge.",
keywords = "Classification, Digital pathology, Image analysis, Segmentation, Tissue images",
author = "Vu, {Quoc Dang} and Simon Graham and Tahsin Kurc and To, {Minh Nguyen Nhat} and Muhammad Shaban and Talha Qaiser and Koohbanani, {Navid Alemi} and Khurram, {Syed Ali} and Jayashree Kalpathy-Cramer and Tianhao Zhao and Rajarsi Gupta and Kwak, {Jin Tae} and Nasir Rajpoot and Joel Saltz and Keyvan Farahani",
year = "2019",
month = "1",
day = "1",
doi = "10.3389/fbioe.2019.00053",
language = "English (US)",
volume = "7",
journal = "Frontiers in Bioengineering and Biotechnology",
issn = "2296-4185",
publisher = "Frontiers Media S. A.",
number = "APR",

}

TY - JOUR

T1 - Methods for segmentation and classification of digital microscopy tissue images

AU - Vu, Quoc Dang

AU - Graham, Simon

AU - Kurc, Tahsin

AU - To, Minh Nguyen Nhat

AU - Shaban, Muhammad

AU - Qaiser, Talha

AU - Koohbanani, Navid Alemi

AU - Khurram, Syed Ali

AU - Kalpathy-Cramer, Jayashree

AU - Zhao, Tianhao

AU - Gupta, Rajarsi

AU - Kwak, Jin Tae

AU - Rajpoot, Nasir

AU - Saltz, Joel

AU - Farahani, Keyvan

PY - 2019/1/1

Y1 - 2019/1/1

N2 - High-resolution microscopy images of tissue specimens provide detailed information about the morphology of normal and diseased tissue. Image analysis of tissue morphology can help cancer researchers develop a better understanding of cancer biology. Segmentation of nuclei and classification of tissue images are two common tasks in tissue image analysis. Development of accurate and efficient algorithms for these tasks is a challenging problem because of the complexity of tissue morphology and tumor heterogeneity. In this paper we present two computer algorithms; one designed for segmentation of nuclei and the other for classification of whole slide tissue images. The segmentation algorithm implements a multiscale deep residual aggregation network to accurately segment nuclear material and then separate clumped nuclei into individual nuclei. The classification algorithm initially carries out patch-level classification via a deep learning method, then patch-level statistical and morphological features are used as input to a random forest regression model for whole slide image classification. The segmentation and classification algorithms were evaluated in the MICCAI 2017 Digital Pathology challenge. The segmentation algorithm achieved an accuracy score of 0.78. The classification algorithm achieved an accuracy score of 0.81. These scores were the highest in the challenge.

AB - High-resolution microscopy images of tissue specimens provide detailed information about the morphology of normal and diseased tissue. Image analysis of tissue morphology can help cancer researchers develop a better understanding of cancer biology. Segmentation of nuclei and classification of tissue images are two common tasks in tissue image analysis. Development of accurate and efficient algorithms for these tasks is a challenging problem because of the complexity of tissue morphology and tumor heterogeneity. In this paper we present two computer algorithms; one designed for segmentation of nuclei and the other for classification of whole slide tissue images. The segmentation algorithm implements a multiscale deep residual aggregation network to accurately segment nuclear material and then separate clumped nuclei into individual nuclei. The classification algorithm initially carries out patch-level classification via a deep learning method, then patch-level statistical and morphological features are used as input to a random forest regression model for whole slide image classification. The segmentation and classification algorithms were evaluated in the MICCAI 2017 Digital Pathology challenge. The segmentation algorithm achieved an accuracy score of 0.78. The classification algorithm achieved an accuracy score of 0.81. These scores were the highest in the challenge.

KW - Classification

KW - Digital pathology

KW - Image analysis

KW - Segmentation

KW - Tissue images

UR - http://www.scopus.com/inward/record.url?scp=85064655258&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85064655258&partnerID=8YFLogxK

U2 - 10.3389/fbioe.2019.00053

DO - 10.3389/fbioe.2019.00053

M3 - Article

AN - SCOPUS:85064655258

VL - 7

JO - Frontiers in Bioengineering and Biotechnology

JF - Frontiers in Bioengineering and Biotechnology

SN - 2296-4185

IS - APR

M1 - 53

ER -