A cooperative framework for segmentation of MRI brain scans

Automatic segmentation of MRI brain scans is a complex task for two main re asons: the large variability of the human brain anatomy, which limits the u se of general knowledge and, inherent to MRI acquisition, the artifacts pre sent in the images that are difficult to process. To tackle these difficult ies, we propose to mix, in a cooperative framework, several types of inform ation and knowledge provided and used by complementary individual systems: presently, a multi-agent system, a deformable model and an edge detector. T he outcome is a cooperative segmentation performed by a set of region and e dge agents constrained automatically and dynamically by both, the specific gray levels in the considered image, statistical models of the brain struct ures and general knowledge about MRI brain scans. Interactions between the individual systems follow three modes of cooperation: integrative, augmenta tive and confrontational cooperation, combined during the three steps of th e segmentation process namely, the specialization of the seeded-region-grow ing agents, the fusion of heterogeneous information and the retroaction ove r slices. The described cooperative framework allows the dynamic adaptation of the segmentation process to the own characteristics of each MRI brain s can. Its evaluation using realistic brain phantoms is reported. (C) 2000 El sevier Science B.V. All rights reserved.