Segmentation of large brain lesions

IEEE Trans Med Imaging. 2001 Jul;20(7):666-9. doi: 10.1109/42.932750.


This paper describes a region-growing algorithm for the segmentation of large lesions in T1-weighted magnetic resonance (MR) images of the head. The algorithm involves a gray level similarity criterion to expand the region and a size criterion to prevent from over-growing outside the lesion. The performance of the algorithm is evaluated and validated on a series of pathologic three-dimensional MR images of the head.

Publication types

  • Evaluation Study

MeSH terms

  • Algorithms*
  • Brain / pathology*
  • Brain Diseases / pathology*
  • Fourier Analysis
  • Image Processing, Computer-Assisted
  • Imaging, Three-Dimensional
  • Magnetic Resonance Imaging / methods*
  • Sensitivity and Specificity