Semi-automated extraction of brain contours from MRI

H. T. Ong, J. Tieman, Marilyn Albert, F. Jolesz, T. Sandor

Research output: Contribution to journalArticle


We compared brain volumes computed by trained operators using the BRAIN-MAP algorithm, which automatically extracts the contours of the brain from gradient-echo magnetic resonance images. The images of 19 subjects randomly selected from a group of normals and a group of patients with dementia were included. BRAIN-MAP found brain perimeters that were on the average ca. 3% tighter than those obtained by two experienced operators. Between-operator and within-operator reproducibility of the analyses were also estimated and found to be (-0.13 ± 0.51)% and (-0.63 ± 0.08)%, respectively. Replicate volume measurement by the computer alone provided a reproducibility of (0.44 ± 0.46)%.

Original languageEnglish (US)
Pages (from-to)797-803
Number of pages7
Issue number11
StatePublished - 1997
Externally publishedYes



  • Brain volume measurement
  • Contour extraction
  • Image processing
  • Magnetic resonance imaging

ASJC Scopus subject areas

  • Radiology Nuclear Medicine and imaging
  • Clinical Neurology
  • Radiological and Ultrasound Technology

Cite this

Ong, H. T., Tieman, J., Albert, M., Jolesz, F., & Sandor, T. (1997). Semi-automated extraction of brain contours from MRI. Neuroradiology, 39(11), 797-803.