Takeshi Matozaki
Department of Electronics & Communication Engineering, Musashi Institute of Technology, 1-28-1 Tamazutsumi, Setagaya, Tokyo 158-8557, Japan
(Received September 15, 2001; Accepted October 22, 2001)
Keywords: Brain MR Image, Automatic Extraction, Clumsy Painter Method, Brain Parenchyma
Abstract. For simulation of surgical operations, extraction of selected regions from MR images requires a high level of skill and experience from the physicians. We have developed a unique automatic extraction algorithm for extracting three dimensional brain parenchyma from MR head images. This method has advantages compared with the morphological filtering (SERRA, 1982, 1986; Haralik et al., 1987) and the region growing method. As first step, this method was applied to binarized images, but there were some problems in that the results of extraction varied much with the value of the threshold level (SATO et al., 1997). As second step, the method was applied to gray scale image, but misextraction was caused because density in the contour of object change, locally. Finally, images having information corresponding to the property of human tissues, respectly, are introduced. The characteristics of this method is that an extracted form is not an exterior of the object, but the form of a set of picture element having a physical property. We could improve extraction accuracy. As a results, the over extracted ratio was improved by 36%, and the under extracted ratio was improved by 20%.