Example Based Lesion Segmentation.

摘要: Automatic and accurate detection of white matter lesions is a significant step toward understanding the progression many diseases, like Alzheimer’s disease or multiple sclerosis. Multi-modal MR images are often used to segment T2 that can represent regions demyelination ischemia. Some automated lesion segmentation methods describe intensities using generative models, then classify with some combination heuristics cost minimization. In contrast, we propose patch-based method, in which found examples from an atlas containing multi-modal corresponding manual delineations lesions. Patches subject matched patches memberships based on patch similarity weights. We experiment 43 subjects MS, whose scans show various levels lesion-load. demonstrate improvement Dice coefficient total volume compared state art model-based indicating more delineation