Abstract

In this paper, we describe a pixel based approach for automated segmentation of tumor components from MR images. Sparse coding with data-adapted dictionaries has been successfully employed in several image recovery and vision problems. Since it is trivial to obtain sparse codes for pixel values, we propose to consider their non-linear similarities to perform kernel sparse coding in a high dimensional feature space. We develop the kernel K-lines clustering procedure for inferring kernel dictionaries and use the kernel sparse codes to determine if a pixel belongs to a tumorous region. By incorporating spatial locality information of the pixels, contiguous tumor regions can be efficiently identified. A low complexity segmentation approach, which allows the user to initialize the tumor region, is also presented. Results show that both of the proposed approaches lead to accurate tumor identification with a low false positive rate, when compared to manual segmentation by an expert.

Original languageEnglish (US)
Title of host publicationIEEE 12th International Conference on BioInformatics and BioEngineering, BIBE 2012
Pages401-406
Number of pages6
DOIs
StatePublished - 2012
Event12th IEEE International Conference on BioInformatics and BioEngineering, BIBE 2012 - Larnaca, Cyprus
Duration: Nov 11 2012Nov 13 2012

Other

Other12th IEEE International Conference on BioInformatics and BioEngineering, BIBE 2012
CountryCyprus
CityLarnaca
Period11/11/1211/13/12

Fingerprint

Tumors
Pixels
Glossaries
Recovery

Keywords

  • kernel methods
  • MRI
  • sparse representations
  • tumor segmentation

ASJC Scopus subject areas

  • Bioengineering
  • Biomedical Engineering

Cite this

Thiagarajan, J. J., Rajan, D., Ramamurthy, K. N., Frakes, D., & Spanias, A. (2012). Automated tumor segmentation using kernel sparse representations. In IEEE 12th International Conference on BioInformatics and BioEngineering, BIBE 2012 (pp. 401-406). [6399658] https://doi.org/10.1109/BIBE.2012.6399658

Automated tumor segmentation using kernel sparse representations. / Thiagarajan, Jayaraman J.; Rajan, Deepta; Ramamurthy, Karthikeyan Natesan; Frakes, David; Spanias, Andreas.

IEEE 12th International Conference on BioInformatics and BioEngineering, BIBE 2012. 2012. p. 401-406 6399658.

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

Thiagarajan, JJ, Rajan, D, Ramamurthy, KN, Frakes, D & Spanias, A 2012, Automated tumor segmentation using kernel sparse representations. in IEEE 12th International Conference on BioInformatics and BioEngineering, BIBE 2012., 6399658, pp. 401-406, 12th IEEE International Conference on BioInformatics and BioEngineering, BIBE 2012, Larnaca, Cyprus, 11/11/12. https://doi.org/10.1109/BIBE.2012.6399658
Thiagarajan JJ, Rajan D, Ramamurthy KN, Frakes D, Spanias A. Automated tumor segmentation using kernel sparse representations. In IEEE 12th International Conference on BioInformatics and BioEngineering, BIBE 2012. 2012. p. 401-406. 6399658 https://doi.org/10.1109/BIBE.2012.6399658
Thiagarajan, Jayaraman J. ; Rajan, Deepta ; Ramamurthy, Karthikeyan Natesan ; Frakes, David ; Spanias, Andreas. / Automated tumor segmentation using kernel sparse representations. IEEE 12th International Conference on BioInformatics and BioEngineering, BIBE 2012. 2012. pp. 401-406
@inproceedings{592af95e0b6a4e9896dae905593bc10b,
title = "Automated tumor segmentation using kernel sparse representations",
abstract = "In this paper, we describe a pixel based approach for automated segmentation of tumor components from MR images. Sparse coding with data-adapted dictionaries has been successfully employed in several image recovery and vision problems. Since it is trivial to obtain sparse codes for pixel values, we propose to consider their non-linear similarities to perform kernel sparse coding in a high dimensional feature space. We develop the kernel K-lines clustering procedure for inferring kernel dictionaries and use the kernel sparse codes to determine if a pixel belongs to a tumorous region. By incorporating spatial locality information of the pixels, contiguous tumor regions can be efficiently identified. A low complexity segmentation approach, which allows the user to initialize the tumor region, is also presented. Results show that both of the proposed approaches lead to accurate tumor identification with a low false positive rate, when compared to manual segmentation by an expert.",
keywords = "kernel methods, MRI, sparse representations, tumor segmentation",
author = "Thiagarajan, {Jayaraman J.} and Deepta Rajan and Ramamurthy, {Karthikeyan Natesan} and David Frakes and Andreas Spanias",
year = "2012",
doi = "10.1109/BIBE.2012.6399658",
language = "English (US)",
isbn = "9781467343589",
pages = "401--406",
booktitle = "IEEE 12th International Conference on BioInformatics and BioEngineering, BIBE 2012",

}

TY - GEN

T1 - Automated tumor segmentation using kernel sparse representations

AU - Thiagarajan, Jayaraman J.

AU - Rajan, Deepta

AU - Ramamurthy, Karthikeyan Natesan

AU - Frakes, David

AU - Spanias, Andreas

PY - 2012

Y1 - 2012

N2 - In this paper, we describe a pixel based approach for automated segmentation of tumor components from MR images. Sparse coding with data-adapted dictionaries has been successfully employed in several image recovery and vision problems. Since it is trivial to obtain sparse codes for pixel values, we propose to consider their non-linear similarities to perform kernel sparse coding in a high dimensional feature space. We develop the kernel K-lines clustering procedure for inferring kernel dictionaries and use the kernel sparse codes to determine if a pixel belongs to a tumorous region. By incorporating spatial locality information of the pixels, contiguous tumor regions can be efficiently identified. A low complexity segmentation approach, which allows the user to initialize the tumor region, is also presented. Results show that both of the proposed approaches lead to accurate tumor identification with a low false positive rate, when compared to manual segmentation by an expert.

AB - In this paper, we describe a pixel based approach for automated segmentation of tumor components from MR images. Sparse coding with data-adapted dictionaries has been successfully employed in several image recovery and vision problems. Since it is trivial to obtain sparse codes for pixel values, we propose to consider their non-linear similarities to perform kernel sparse coding in a high dimensional feature space. We develop the kernel K-lines clustering procedure for inferring kernel dictionaries and use the kernel sparse codes to determine if a pixel belongs to a tumorous region. By incorporating spatial locality information of the pixels, contiguous tumor regions can be efficiently identified. A low complexity segmentation approach, which allows the user to initialize the tumor region, is also presented. Results show that both of the proposed approaches lead to accurate tumor identification with a low false positive rate, when compared to manual segmentation by an expert.

KW - kernel methods

KW - MRI

KW - sparse representations

KW - tumor segmentation

UR - http://www.scopus.com/inward/record.url?scp=84872874128&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84872874128&partnerID=8YFLogxK

U2 - 10.1109/BIBE.2012.6399658

DO - 10.1109/BIBE.2012.6399658

M3 - Conference contribution

AN - SCOPUS:84872874128

SN - 9781467343589

SP - 401

EP - 406

BT - IEEE 12th International Conference on BioInformatics and BioEngineering, BIBE 2012

ER -