Computer-Aided Diagnosis

Semiautomatic tumor segmentation with multimodal images in a conditional random field framework

[+] Author Affiliations
Yu-chi Hu

Memorial Sloan Kettering Cancer Center, Department of Medical Physics, 1275 York Avenue, New York, New York 10065, United States

City College of New York, Department of Computer Science, 160 Convent Avenue, New York, New York 10031, United States

Michael Grossberg

City College of New York, Department of Computer Science, 160 Convent Avenue, New York, New York 10031, United States

Gikas Mageras

Memorial Sloan Kettering Cancer Center, Department of Medical Physics, 1275 York Avenue, New York, New York 10065, United States

J. Med. Imag. 3(2), 024503 (Jun 28, 2016). doi:10.1117/1.JMI.3.2.024503
History: Received November 11, 2015; Accepted June 2, 2016
Text Size: A A A

Abstract.  Volumetric medical images of a single subject can be acquired using different imaging modalities, such as computed tomography, magnetic resonance imaging (MRI), and positron emission tomography. In this work, we present a semiautomatic segmentation algorithm that can leverage the synergies between different image modalities while integrating interactive human guidance. The algorithm provides a statistical segmentation framework partly automating the segmentation task while still maintaining critical human oversight. The statistical models presented are trained interactively using simple brush strokes to indicate tumor and nontumor tissues and using intermediate results within a patient’s image study. To accomplish the segmentation, we construct the energy function in the conditional random field (CRF) framework. For each slice, the energy function is set using the estimated probabilities from both user brush stroke data and prior approved segmented slices within a patient study. The progressive segmentation is obtained using a graph-cut-based minimization. Although no similar semiautomated algorithm is currently available, we evaluated our method with an MRI data set from Medical Image Computing and Computer Assisted Intervention Society multimodal brain segmentation challenge (BRATS 2012 and 2013) against a similar fully automatic method based on CRF and a semiautomatic method based on grow-cut, and our method shows superior performance.

Figures in this Article
© 2016 Society of Photo-Optical Instrumentation Engineers

Citation

Yu-chi Hu ; Michael Grossberg and Gikas Mageras
"Semiautomatic tumor segmentation with multimodal images in a conditional random field framework", J. Med. Imag. 3(2), 024503 (Jun 28, 2016). ; http://dx.doi.org/10.1117/1.JMI.3.2.024503


Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).

Some tools below are only available to our subscribers or users with an online account.

Related Content

Customize your page view by dragging & repositioning the boxes below.

Related Book Chapters

Topic Collections

PubMed Articles
Advertisement
  • Don't have an account?
  • Subscribe to the SPIE Digital Library
  • Create a FREE account to sign up for Digital Library content alerts and gain access to institutional subscriptions remotely.
Access This Article
Sign in or Create a personal account to Buy this article ($20 for members, $25 for non-members).
Access This Proceeding
Sign in or Create a personal account to Buy this article ($15 for members, $18 for non-members).
Access This Chapter

Access to SPIE eBooks is limited to subscribing institutions and is not available as part of a personal subscription. Print or electronic versions of individual SPIE books may be purchased via SPIE.org.