Stacked Predictive Sparse Decomposition for Classification of Histology Sections

Hang Chang, Yin Zhou, Alexander D Borowsky, Kenneth Barner, Paul Spellman, Bahram Parvin

Research output: Contribution to journalArticle

11 Citations (Scopus)

Abstract

Image-based classification of histology sections, in terms of distinct components (e.g., tumor, stroma, normal), provides a series of indices for histology composition (e.g., the percentage of each distinct components in histology sections), and enables the study of nuclear properties within each component. Furthermore, the study of these indices, constructed from each whole slide image in a large cohort, has the potential to provide predictive models of clinical outcome. For example, correlations can be established between the constructed indices and the patients’ survival information at cohort level, which is a fundamental step towards personalized medicine. However, performance of the existing techniques is hindered as a result of large technical variations (e.g., variations of color/textures in tissue images due to non-standard experimental protocols) and biological heterogeneities (e.g., cell type, cell state) that are always present in a large cohort. We propose a system that automatically learns a series of dictionary elements for representing the underlying spatial distribution using stacked predictive sparse decomposition. The learned representation is then fed into the spatial pyramid matching framework with a linear support vector machine classifier. The system has been evaluated for classification of distinct histological components for two cohorts of tumor types. Throughput has been increased by using of graphical processing unit (GPU), and evaluation indicates a superior performance results, compared with previous research.

Original languageEnglish (US)
Pages (from-to)3-18
Number of pages16
JournalInternational Journal of Computer Vision
Volume113
Issue number1
DOIs
StatePublished - May 1 2015

Fingerprint

Histology
Decomposition
Tumors
Nuclear properties
Glossaries
Spatial distribution
Medicine
Support vector machines
Classifiers
Textures
Throughput
Tissue
Color
Processing
Chemical analysis

Keywords

  • Classification
  • Sparse coding
  • Tissue histology
  • Unsupervised feature learning

ASJC Scopus subject areas

  • Software
  • Artificial Intelligence
  • Computer Vision and Pattern Recognition

Cite this

Stacked Predictive Sparse Decomposition for Classification of Histology Sections. / Chang, Hang; Zhou, Yin; Borowsky, Alexander D; Barner, Kenneth; Spellman, Paul; Parvin, Bahram.

In: International Journal of Computer Vision, Vol. 113, No. 1, 01.05.2015, p. 3-18.

Research output: Contribution to journalArticle

Chang, Hang ; Zhou, Yin ; Borowsky, Alexander D ; Barner, Kenneth ; Spellman, Paul ; Parvin, Bahram. / Stacked Predictive Sparse Decomposition for Classification of Histology Sections. In: International Journal of Computer Vision. 2015 ; Vol. 113, No. 1. pp. 3-18.
@article{f770f5c021a743b7912cf9978ed4ce22,
title = "Stacked Predictive Sparse Decomposition for Classification of Histology Sections",
abstract = "Image-based classification of histology sections, in terms of distinct components (e.g., tumor, stroma, normal), provides a series of indices for histology composition (e.g., the percentage of each distinct components in histology sections), and enables the study of nuclear properties within each component. Furthermore, the study of these indices, constructed from each whole slide image in a large cohort, has the potential to provide predictive models of clinical outcome. For example, correlations can be established between the constructed indices and the patients’ survival information at cohort level, which is a fundamental step towards personalized medicine. However, performance of the existing techniques is hindered as a result of large technical variations (e.g., variations of color/textures in tissue images due to non-standard experimental protocols) and biological heterogeneities (e.g., cell type, cell state) that are always present in a large cohort. We propose a system that automatically learns a series of dictionary elements for representing the underlying spatial distribution using stacked predictive sparse decomposition. The learned representation is then fed into the spatial pyramid matching framework with a linear support vector machine classifier. The system has been evaluated for classification of distinct histological components for two cohorts of tumor types. Throughput has been increased by using of graphical processing unit (GPU), and evaluation indicates a superior performance results, compared with previous research.",
keywords = "Classification, Sparse coding, Tissue histology, Unsupervised feature learning",
author = "Hang Chang and Yin Zhou and Borowsky, {Alexander D} and Kenneth Barner and Paul Spellman and Bahram Parvin",
year = "2015",
month = "5",
day = "1",
doi = "10.1007/s11263-014-0790-9",
language = "English (US)",
volume = "113",
pages = "3--18",
journal = "International Journal of Computer Vision",
issn = "0920-5691",
publisher = "Springer Netherlands",
number = "1",

}

TY - JOUR

T1 - Stacked Predictive Sparse Decomposition for Classification of Histology Sections

AU - Chang, Hang

AU - Zhou, Yin

AU - Borowsky, Alexander D

AU - Barner, Kenneth

AU - Spellman, Paul

AU - Parvin, Bahram

PY - 2015/5/1

Y1 - 2015/5/1

N2 - Image-based classification of histology sections, in terms of distinct components (e.g., tumor, stroma, normal), provides a series of indices for histology composition (e.g., the percentage of each distinct components in histology sections), and enables the study of nuclear properties within each component. Furthermore, the study of these indices, constructed from each whole slide image in a large cohort, has the potential to provide predictive models of clinical outcome. For example, correlations can be established between the constructed indices and the patients’ survival information at cohort level, which is a fundamental step towards personalized medicine. However, performance of the existing techniques is hindered as a result of large technical variations (e.g., variations of color/textures in tissue images due to non-standard experimental protocols) and biological heterogeneities (e.g., cell type, cell state) that are always present in a large cohort. We propose a system that automatically learns a series of dictionary elements for representing the underlying spatial distribution using stacked predictive sparse decomposition. The learned representation is then fed into the spatial pyramid matching framework with a linear support vector machine classifier. The system has been evaluated for classification of distinct histological components for two cohorts of tumor types. Throughput has been increased by using of graphical processing unit (GPU), and evaluation indicates a superior performance results, compared with previous research.

AB - Image-based classification of histology sections, in terms of distinct components (e.g., tumor, stroma, normal), provides a series of indices for histology composition (e.g., the percentage of each distinct components in histology sections), and enables the study of nuclear properties within each component. Furthermore, the study of these indices, constructed from each whole slide image in a large cohort, has the potential to provide predictive models of clinical outcome. For example, correlations can be established between the constructed indices and the patients’ survival information at cohort level, which is a fundamental step towards personalized medicine. However, performance of the existing techniques is hindered as a result of large technical variations (e.g., variations of color/textures in tissue images due to non-standard experimental protocols) and biological heterogeneities (e.g., cell type, cell state) that are always present in a large cohort. We propose a system that automatically learns a series of dictionary elements for representing the underlying spatial distribution using stacked predictive sparse decomposition. The learned representation is then fed into the spatial pyramid matching framework with a linear support vector machine classifier. The system has been evaluated for classification of distinct histological components for two cohorts of tumor types. Throughput has been increased by using of graphical processing unit (GPU), and evaluation indicates a superior performance results, compared with previous research.

KW - Classification

KW - Sparse coding

KW - Tissue histology

KW - Unsupervised feature learning

UR - http://www.scopus.com/inward/record.url?scp=84939942022&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84939942022&partnerID=8YFLogxK

U2 - 10.1007/s11263-014-0790-9

DO - 10.1007/s11263-014-0790-9

M3 - Article

AN - SCOPUS:84939942022

VL - 113

SP - 3

EP - 18

JO - International Journal of Computer Vision

JF - International Journal of Computer Vision

SN - 0920-5691

IS - 1

ER -