Abstract
This paper introduces a novel solver, namely cross entropy (CE), into the MRF theory for medical image segmentation. The solver, which is based on the theory of rare event simulation, is general and stochastic. Unlike some popular optimization methods such as belief propagation and graph cuts, CE makes no assumption on the form of objective functions and thus can be applied to any type of MRF models. Furthermore, it achieves higher performance of finding more global optima because of its stochastic property. In addition, it is more efficient than other stochastic methods like simulated annealing. We tested the new solver in 4 series of segmentation experiments on synthetic and clinical, vascular and cerebral images. The experiments show that CE can give more accurate segmentation results.
Chapter PDF
Similar content being viewed by others
Keywords
- Belief Propagation
- Cross Entropy
- Markov Random Field Modeling
- Medical Image Segmentation
- Cross Entropy Method
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.
References
Villain, N., Goussard, Y., Idier, J., Allain, M.: Three-dimensional edge-preserving image enhancement for computed tomography. IEEE TMI 22, 1275–1287 (2003)
Chen, Y., Amini, A.: A map framework for tag line detection in spamm data using markov random fields on the b-spline solid. IEEE TMI 21, 1110–1122 (2002)
Svensen, M., Kruggel, F., Cramon, D.: Probabilistic modeling of single-trial fmri data. IEEE TMI 19, 25–35 (2000)
Chung, A.C.S., Noble, J.A., Summers, P.: Vascular segmentation of pcmra based on statistical mixture modeling and local phase coherence. IEEE TMI 23, 1490–1507 (2004)
Boykov, Y., Veksler, O., Zabin, R.: Fast approximate energy minimization via graph cuts. IEEE on PAMI 23, 1222–1239 (2001)
Kirkpatrick, S., Gellatt, D., Vecchi, M.: Optimization by simulated annealing. Science 220, 671–680 (1983)
Besag, J.: On the statistical analysis of dirty pictures. J. of the Roy. Stat. Soc. Series B 48, 259–302 (1986)
Li, S.Z. (ed.): Markov Random Field Modeling in Image Analysis. Springer, Heidelberg (2001)
Yedidia, J.S., Freeman, W.T., Weiss, Y.: Understanding belief propagation and its generalizations. Technical Report, Mitsubishi Electric Research TR-2001-22 (2001)
Zhu, S., Wu, Y., Mumford, D.: Frame: filters, random fields, and minimax entropy towards a unified theory for texture modeling. IEEE CVPR 6, 686–693 (1996)
Kolmogorov, V., Zabih, R.: What energy functions can be minimized via graph cuts? IEEE Trans. PAMI 26, 147–159 (2004)
Rubinstein, R.: Optimization of computer simulation models with rare events. European Journal of Operations Research 99, 89–112 (1997)
Tappen, M., Freeman, W.: Comparison of graph cuts with belief propagation for stereo, using identical mrf parameters. In: IEEE ICCV, pp. 900–906 (2003)
Rubinstein, R., Kroese, D. (eds.): The Cross-Entropy Method: A Unified Approach to Cominatorial Optimization. Monte-Carlo Simulation and Machine Learning. Springer, New York (2004)
Held, K., et al.: Markov random field segmentation of brain mr images. IEEE TMI 16, 878–886 (1997)
Ruan, S., et al.: Brain tissue classification of magnetic resonance images using partial volume modeling. IEEE TMI 19, 1179–1187 (2000)
BrainWeb, http://www.bic.mni.mcgill.ca/brainweb/
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2005 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Wu, J., Chung, A.C.S. (2005). Cross Entropy: A New Solver for Markov Random Field Modeling and Applications to Medical Image Segmentation. In: Duncan, J.S., Gerig, G. (eds) Medical Image Computing and Computer-Assisted Intervention – MICCAI 2005. MICCAI 2005. Lecture Notes in Computer Science, vol 3749. Springer, Berlin, Heidelberg. https://doi.org/10.1007/11566465_29
Download citation
DOI: https://doi.org/10.1007/11566465_29
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-29327-9
Online ISBN: 978-3-540-32094-4
eBook Packages: Computer ScienceComputer Science (R0)