9308839 Merickel The goal of this project is to use a knowledge based method to extract contours and surfaces from medical images, such as magnetic resonance images. The approach that is proposed is a "top down" one that could be compared with the mental processes used by radiologists instead of the more traditional "bottom up" approach used in image segmentation. The use of adaptive contours will be extended to 3-D volumes. The investigators will study two specific cases; namely, vessel walls and brain regions. This research will address a fundamental problem in image processing and computer vision research. For example, in the biomedical area, image segmentation is important for the quantitative and noninvasive analysis of tissue and structure volumes, such as might be needed for the evaluation of vessel integrity due to atherosclerosis and brain structures for surgical planning. ***
