Abstract
Current state-of-the-art segmentation techniques of brain MR images improve segmentation accuracy by encoding spatial information through hidden Markov random field (HMRF) model. However, HMRF model has higher computational overhead compared to finite Gaussian mixture (FGM) model but the segmentation results are with no significant difference when applying to cleaner data. We believe this is because the spatial constraint is too simple to utilize the characteristics of the brain. In this paper, we propose a novel method to improve the neighborhood system of the HMRF model by better characterizing natural structures of human brain. Experiments on both real and synthetic 3D brain MR images show that the segmentation results of our method have higher accuracy compared to existing solutions.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 3209-3212 |
| Number of pages | 4 |
| Journal | IEEE Nuclear Science Symposium Conference Record |
| Volume | 5 |
| State | Published - 2004 |
| Externally published | Yes |
| Event | 2004 Nuclear Science Symposium, Medical Imaging Conference, Symposium on Nuclear Power Systems and the 14th International Workshop on Room Temperature Semiconductor X- and Gamma- Ray Detectors - Rome, Italy Duration: Oct 16 2004 → Oct 22 2004 |
Keywords
- Brain MRI
- Finite Gaussian mixture model
- Hidden Markov random field model
- Markov random field theory
- Segmentation
ASJC Scopus subject areas
- Radiation
- Nuclear and High Energy Physics
- Radiology Nuclear Medicine and imaging