Segmentation Of MR Brain Images
Jussi P. S. Parkkinen, Gregg Cohen, Milan Sonka, Nancy C. Andreasen · 2005
The segmentation of the image is a useful opera- tion in several medical image processing tasks. Two examples are the extraction of the lung tissue from the x-ray image and the segregation of gray matter, white matter, and cerebrospinal fluid in the brain images. Since the segmentation needs are of- ten related to the anatomy of the object, segmentation is most naturally done from the anatomical images like x-ray, Comput- erized Tomography (CT), or Magnetic Resonance (MR) images. An advantage of the use of the MR-images is that they can be acquired as multicomponent images and therefore the pixelwise segmentation can be based on more information than in the x-ray images, where the pixelwise segmentation is simply thresholding. From the segmented image one is able to derive different features to describe the substructures in the image. In this paper we describe two methods for pixel-wise segmen- tation of the MR-images of the brain. A set of useful shape- features derived from the edge-based segmentation of the subre- gions is given, also.