Not Logged In

Segmenting Brain Tumor with Conditional Random Fields and Support Vector Machines

Full Text: FINAL.pdf PDF

Markov Random Fields (MRFs) are a popular and well motivated model for many medical image processing tasks such as segmentation. Discriminative Random Fields (DRFs), a discriminative alternative to the traditionally generative MRFs, allow tractable computation with less restrictive simplifying assumptions, and achieve better performance in many tasks. In this paper, we investigate the tumor segmentation performance of a recent variant of DRF models that takes advantage of the powerful Support Vector Machine (SVM) classification method. Combined with a powerful Magnetic Resonance (MR) preprocessing pipeline and a set of alignment-based features, we evaluate theuse of SVMs, MRFs, and two types of DRFs as classifiers for threes egmentation tasks related to radiation therapy target planning for brain tumors, two of which do not rely on contrast agent enhancement. Our results indicate that the SVM-based DRFs offer a significant advantageover the other approaches.

Citation

C. Lee, M. Schmidt, A. Murtha, A. Bistritz, J. Sander, R. Greiner. "Segmenting Brain Tumor with Conditional Random Fields and Support Vector Machines". Computer Vision for Biomedical Image Applications: Current Techniques and Future Trends, Springer, (ed: Liu, Yanxi; Jiang, Tianzi; Zhang, Changshui), pp 469-478, October 2005.

Keywords: brain segmentation, brain tumor, SVM, MRF, Random Field, machine learning
Category: In Workshop
Web Links: Springer Link

BibTeX

@misc{Lee+al:ICCV-CVBIA05,
  author = {Chi-Hoon Lee and Mark Schmidt and Albert Murtha and Aalo Bistritz
    and Joerg Sander and Russ Greiner},
  title = {Segmenting Brain Tumor with Conditional Random Fields and Support
    Vector Machines},
  Booktitle = {Computer Vision for Biomedical Image Applications First
    International Workshop...Proceedings},
  Publisher = "Springer",
  Editor = {Liu, Yanxi; Jiang, Tianzi; Zhang, Changshui},
  Pages = {469-478},
  booktitle = {Computer Vision for Biomedical Image Applications:  Current
    Techniques and Future Trends},
  year = 2005,
}

Last Updated: November 21, 2019
Submitted by Sabina P

University of Alberta Logo AICML Logo