Learning invariant features of tumor signatures

Quoc V. Le, Ju Han, Joe Gray, Paul Spellman, Alexander Borowsky, Bahram Parvin

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

23 Citations (Scopus)

Abstract

We present a novel method for automated learning of features from unlabeled image patches for classification of tumor architecture. In contrast to previous manually-designed feature detectors (e.g., Gabor basis function), the proposed method utilizes inexpensive un-labeled data to construct features. The algorithm, also known as reconstruction independent subspace analysis, can be described as a two-layer network with non-linear responses, where the second layer represents subspace structures. The technique is applied to tissue sections for characterizing necrosis, apoptotic, and viable regions of Glioblastoma Multifrome (GBM) from TCGA dataset. Experimental results show that this method outperforms more complex expert-designed approaches. The fact that our approach learns features automatically from unlabeled data promises a wider application of self-learning strategies for tissue characterization.

Original languageEnglish (US)
Title of host publicationProceedings - International Symposium on Biomedical Imaging
Pages302-305
Number of pages4
DOIs
StatePublished - 2012
Event2012 9th IEEE International Symposium on Biomedical Imaging: From Nano to Macro, ISBI 2012 - Barcelona, Spain
Duration: May 2 2012May 5 2012

Other

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

Fingerprint

Tumors
Learning
Tissue
Network layers
Neoplasms
Glioblastoma
Detectors
Necrosis
Datasets

Keywords

  • apoptotic and necrotic signatures
  • subspace learning
  • tumor architecture

ASJC Scopus subject areas

  • Biomedical Engineering
  • Radiology Nuclear Medicine and imaging

Cite this

Le, Q. V., Han, J., Gray, J., Spellman, P., Borowsky, A., & Parvin, B. (2012). Learning invariant features of tumor signatures. In Proceedings - International Symposium on Biomedical Imaging (pp. 302-305). [6235544] https://doi.org/10.1109/ISBI.2012.6235544

Learning invariant features of tumor signatures. / Le, Quoc V.; Han, Ju; Gray, Joe; Spellman, Paul; Borowsky, Alexander; Parvin, Bahram.

Proceedings - International Symposium on Biomedical Imaging. 2012. p. 302-305 6235544.

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

Le, QV, Han, J, Gray, J, Spellman, P, Borowsky, A & Parvin, B 2012, Learning invariant features of tumor signatures. in Proceedings - International Symposium on Biomedical Imaging., 6235544, pp. 302-305, 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.2012.6235544
Le QV, Han J, Gray J, Spellman P, Borowsky A, Parvin B. Learning invariant features of tumor signatures. In Proceedings - International Symposium on Biomedical Imaging. 2012. p. 302-305. 6235544 https://doi.org/10.1109/ISBI.2012.6235544
Le, Quoc V. ; Han, Ju ; Gray, Joe ; Spellman, Paul ; Borowsky, Alexander ; Parvin, Bahram. / Learning invariant features of tumor signatures. Proceedings - International Symposium on Biomedical Imaging. 2012. pp. 302-305
@inproceedings{0fb1eba63a754854996043b113e81c24,
title = "Learning invariant features of tumor signatures",
abstract = "We present a novel method for automated learning of features from unlabeled image patches for classification of tumor architecture. In contrast to previous manually-designed feature detectors (e.g., Gabor basis function), the proposed method utilizes inexpensive un-labeled data to construct features. The algorithm, also known as reconstruction independent subspace analysis, can be described as a two-layer network with non-linear responses, where the second layer represents subspace structures. The technique is applied to tissue sections for characterizing necrosis, apoptotic, and viable regions of Glioblastoma Multifrome (GBM) from TCGA dataset. Experimental results show that this method outperforms more complex expert-designed approaches. The fact that our approach learns features automatically from unlabeled data promises a wider application of self-learning strategies for tissue characterization.",
keywords = "apoptotic and necrotic signatures, subspace learning, tumor architecture",
author = "Le, {Quoc V.} and Ju Han and Joe Gray and Paul Spellman and Alexander Borowsky and Bahram Parvin",
year = "2012",
doi = "10.1109/ISBI.2012.6235544",
language = "English (US)",
isbn = "9781457718588",
pages = "302--305",
booktitle = "Proceedings - International Symposium on Biomedical Imaging",

}

TY - GEN

T1 - Learning invariant features of tumor signatures

AU - Le, Quoc V.

AU - Han, Ju

AU - Gray, Joe

AU - Spellman, Paul

AU - Borowsky, Alexander

AU - Parvin, Bahram

PY - 2012

Y1 - 2012

N2 - We present a novel method for automated learning of features from unlabeled image patches for classification of tumor architecture. In contrast to previous manually-designed feature detectors (e.g., Gabor basis function), the proposed method utilizes inexpensive un-labeled data to construct features. The algorithm, also known as reconstruction independent subspace analysis, can be described as a two-layer network with non-linear responses, where the second layer represents subspace structures. The technique is applied to tissue sections for characterizing necrosis, apoptotic, and viable regions of Glioblastoma Multifrome (GBM) from TCGA dataset. Experimental results show that this method outperforms more complex expert-designed approaches. The fact that our approach learns features automatically from unlabeled data promises a wider application of self-learning strategies for tissue characterization.

AB - We present a novel method for automated learning of features from unlabeled image patches for classification of tumor architecture. In contrast to previous manually-designed feature detectors (e.g., Gabor basis function), the proposed method utilizes inexpensive un-labeled data to construct features. The algorithm, also known as reconstruction independent subspace analysis, can be described as a two-layer network with non-linear responses, where the second layer represents subspace structures. The technique is applied to tissue sections for characterizing necrosis, apoptotic, and viable regions of Glioblastoma Multifrome (GBM) from TCGA dataset. Experimental results show that this method outperforms more complex expert-designed approaches. The fact that our approach learns features automatically from unlabeled data promises a wider application of self-learning strategies for tissue characterization.

KW - apoptotic and necrotic signatures

KW - subspace learning

KW - tumor architecture

UR - http://www.scopus.com/inward/record.url?scp=84864859719&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84864859719&partnerID=8YFLogxK

U2 - 10.1109/ISBI.2012.6235544

DO - 10.1109/ISBI.2012.6235544

M3 - Conference contribution

AN - SCOPUS:84864859719

SN - 9781457718588

SP - 302

EP - 305

BT - Proceedings - International Symposium on Biomedical Imaging

ER -