Share Email Print
cover

Proceedings Paper

Unsupervised MRI segmentation with spatial connectivity
Format Member Price Non-Member Price
PDF $14.40 $18.00

Paper Abstract

Magnetic Resonance Imaging (MRI) offers a wealth of information for medical examination. Fast, accurate and reproducible segmentation of MRI is desirable in many applications. We have developed a new unsupervised MRI segmentation method based on k-means and fuzzy c-means (FCM) algorithms, which uses spatial constraints. Spatial constraints are included by the use of a Markov Random Field model. The result of segmentation with a four-neighbor Markov Random Field model applied to multi-spectral MRI (5 images including one T1-weighted image, one Proton Density image and three T2-weighted images) in different noise levels is compared to the segmentation results of standard k-means and FCM algorithms. This comparison shows that the proposed method outperforms previous methods.

Paper Details

Date Published: 9 May 2002
PDF: 9 pages
Proc. SPIE 4684, Medical Imaging 2002: Image Processing, (9 May 2002); doi: 10.1117/12.467147
Show Author Affiliations
Mohammad Mehdi Khalighi, Univ. of Tehran, Institute for Studies of Theoretical Physics and Mathematics and Henry Fo (United States)
Hamid Soltanian-Zadeh, Univ. of Tehran, Institute for Studies of Theoretical Physics and Mathematics and Henry Fo (United States)
Caro Lucas, Univ. of Tehran and Institute for Studies of Theoretical Physics and Mathematics (Iran)


Published in SPIE Proceedings Vol. 4684:
Medical Imaging 2002: Image Processing
Milan Sonka; J. Michael Fitzpatrick, Editor(s)

© SPIE. Terms of Use
Back to Top