Partition MR Images by Extended EM Segmentation
We propose an extended EM segmentation algorithm for partitioning brain tissue into distinct tissue types. The method extends the traditional EM segmentation to a power transformed version. The use of the algorithm for segmenting cortical gray matter was tested using T1-weighted images and compared with results derived from, the SPM software. We also used the extended algorithm to isolate gray matter voxels in functional magnetic resonance images collected in a retinotopic mapping experiment. Intensity time courses associated with gray matter voxels were analyzed by independent component analysis. The empirical results showed that intensity in the gray matter voxels gave taskrelated component maps which were more concentrated on target brain areas.
brain tissue EM segmentation statistical clustering.
Juin-Der Lee Michelle Liou Philip E. Cheng Chien-Chung Chen
Institute of Statistical Science, Academia Sinica, Taipei 11529 Neurometrics Institute, Berkeley, USA
国际会议
8th International Conference on Neural Information Processing(ICONIP 2001)(第八届国际神经信息处理大会)
上海
英文
468-473
2001-11-14（万方平台首次上网日期，不代表论文的发表时间）