作者: Koen L. Vincken , André S. E. Koster , Max A. Viergever
DOI: 10.1007/978-3-540-49197-2_44
关键词:
摘要: A multiscale method (the hyperstack) is proposed to segment multidimensional MR brain data. Hyperstack segmentation based upon the linking of voxels at adjacent levels in scale space, followed by a root selection find that represent segments original image. This paper addresses an advanced and labeling for hyperstack. In particular, attention will be paid extension scheme detection classification partial volume voxels. The result—a list probabilities each voxel—improves resulting segmentations.