Robust segmentation of overlapping cells in histopathology specimens using parallel seed detection and repulsive level set

Xin Qi, Fuyong Xing, David J. Foran, Lin Yang

Research output: Contribution to journalArticle

151 Citations (Scopus)

Abstract

Automated image analysis of histopathology specimens could potentially provide support for early detection and improved characterization of breast cancer. Automated segmentation of the cells comprising imaged tissue microarrays (TMAs) is a prerequisite for any subsequent quantitative analysis. Unfortunately, crowding and overlapping of cells present significant challenges for most traditional segmentation algorithms. In this paper, we propose a novel algorithm that can reliably separate touching cells in hematoxylin-stained breast TMA specimens that have been acquired using a standard RGB camera. The algorithm is composed of two steps. It begins with a fast, reliable object center localization approach that utilizes single-path voting followed by mean-shift clustering. Next, the contour of each cell is obtained using a level set algorithm based on an interactive model. We compared the experimental results with those reported in the most current literature. Finally, performance was evaluated by comparing the pixel-wise accuracy provided by human experts with that produced by the new automated segmentation algorithm. The method was systematically tested on $234$ image patches exhibiting dense overlap and containing more than $2200$ cells. It was also tested on whole slide images including blood smears and TMAs containing thousands of cells. Since the voting step of the seed detection algorithm is well suited for parallelization, a parallel version of the algorithm was implemented using graphic processing units (GPU) that resulted in significant speedup over the C/C implementation.

Original languageEnglish (US)
Article number6099601
Pages (from-to)754-765
Number of pages12
JournalIEEE Transactions on Biomedical Engineering
Volume59
Issue number3
DOIs
StatePublished - Mar 1 2012

Fingerprint

Seed
Microarrays
Tissue
Image analysis
Blood
Pixels
Cameras
Chemical analysis

All Science Journal Classification (ASJC) codes

  • Biomedical Engineering

Keywords

  • Level set
  • mean shift
  • parallel computing
  • seed detection
  • segmentation

Cite this

@article{fa06e4142b614512b6641c0ef62bc4a6,
title = "Robust segmentation of overlapping cells in histopathology specimens using parallel seed detection and repulsive level set",
abstract = "Automated image analysis of histopathology specimens could potentially provide support for early detection and improved characterization of breast cancer. Automated segmentation of the cells comprising imaged tissue microarrays (TMAs) is a prerequisite for any subsequent quantitative analysis. Unfortunately, crowding and overlapping of cells present significant challenges for most traditional segmentation algorithms. In this paper, we propose a novel algorithm that can reliably separate touching cells in hematoxylin-stained breast TMA specimens that have been acquired using a standard RGB camera. The algorithm is composed of two steps. It begins with a fast, reliable object center localization approach that utilizes single-path voting followed by mean-shift clustering. Next, the contour of each cell is obtained using a level set algorithm based on an interactive model. We compared the experimental results with those reported in the most current literature. Finally, performance was evaluated by comparing the pixel-wise accuracy provided by human experts with that produced by the new automated segmentation algorithm. The method was systematically tested on $234$ image patches exhibiting dense overlap and containing more than $2200$ cells. It was also tested on whole slide images including blood smears and TMAs containing thousands of cells. Since the voting step of the seed detection algorithm is well suited for parallelization, a parallel version of the algorithm was implemented using graphic processing units (GPU) that resulted in significant speedup over the C/C implementation.",
keywords = "Level set, mean shift, parallel computing, seed detection, segmentation",
author = "Xin Qi and Fuyong Xing and Foran, {David J.} and Lin Yang",
year = "2012",
month = "3",
day = "1",
doi = "10.1109/TBME.2011.2179298",
language = "English (US)",
volume = "59",
pages = "754--765",
journal = "IEEE Transactions on Biomedical Engineering",
issn = "0018-9294",
publisher = "IEEE Computer Society",
number = "3",

}

Robust segmentation of overlapping cells in histopathology specimens using parallel seed detection and repulsive level set. / Qi, Xin; Xing, Fuyong; Foran, David J.; Yang, Lin.

In: IEEE Transactions on Biomedical Engineering, Vol. 59, No. 3, 6099601, 01.03.2012, p. 754-765.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Robust segmentation of overlapping cells in histopathology specimens using parallel seed detection and repulsive level set

AU - Qi, Xin

AU - Xing, Fuyong

AU - Foran, David J.

AU - Yang, Lin

PY - 2012/3/1

Y1 - 2012/3/1

N2 - Automated image analysis of histopathology specimens could potentially provide support for early detection and improved characterization of breast cancer. Automated segmentation of the cells comprising imaged tissue microarrays (TMAs) is a prerequisite for any subsequent quantitative analysis. Unfortunately, crowding and overlapping of cells present significant challenges for most traditional segmentation algorithms. In this paper, we propose a novel algorithm that can reliably separate touching cells in hematoxylin-stained breast TMA specimens that have been acquired using a standard RGB camera. The algorithm is composed of two steps. It begins with a fast, reliable object center localization approach that utilizes single-path voting followed by mean-shift clustering. Next, the contour of each cell is obtained using a level set algorithm based on an interactive model. We compared the experimental results with those reported in the most current literature. Finally, performance was evaluated by comparing the pixel-wise accuracy provided by human experts with that produced by the new automated segmentation algorithm. The method was systematically tested on $234$ image patches exhibiting dense overlap and containing more than $2200$ cells. It was also tested on whole slide images including blood smears and TMAs containing thousands of cells. Since the voting step of the seed detection algorithm is well suited for parallelization, a parallel version of the algorithm was implemented using graphic processing units (GPU) that resulted in significant speedup over the C/C implementation.

AB - Automated image analysis of histopathology specimens could potentially provide support for early detection and improved characterization of breast cancer. Automated segmentation of the cells comprising imaged tissue microarrays (TMAs) is a prerequisite for any subsequent quantitative analysis. Unfortunately, crowding and overlapping of cells present significant challenges for most traditional segmentation algorithms. In this paper, we propose a novel algorithm that can reliably separate touching cells in hematoxylin-stained breast TMA specimens that have been acquired using a standard RGB camera. The algorithm is composed of two steps. It begins with a fast, reliable object center localization approach that utilizes single-path voting followed by mean-shift clustering. Next, the contour of each cell is obtained using a level set algorithm based on an interactive model. We compared the experimental results with those reported in the most current literature. Finally, performance was evaluated by comparing the pixel-wise accuracy provided by human experts with that produced by the new automated segmentation algorithm. The method was systematically tested on $234$ image patches exhibiting dense overlap and containing more than $2200$ cells. It was also tested on whole slide images including blood smears and TMAs containing thousands of cells. Since the voting step of the seed detection algorithm is well suited for parallelization, a parallel version of the algorithm was implemented using graphic processing units (GPU) that resulted in significant speedup over the C/C implementation.

KW - Level set

KW - mean shift

KW - parallel computing

KW - seed detection

KW - segmentation

UR - http://www.scopus.com/inward/record.url?scp=84863170133&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84863170133&partnerID=8YFLogxK

U2 - 10.1109/TBME.2011.2179298

DO - 10.1109/TBME.2011.2179298

M3 - Article

C2 - 22167559

AN - SCOPUS:84863170133

VL - 59

SP - 754

EP - 765

JO - IEEE Transactions on Biomedical Engineering

JF - IEEE Transactions on Biomedical Engineering

SN - 0018-9294

IS - 3

M1 - 6099601

ER -