Automatic segmentation of the articular cartilage in knee MRI using a hierarchical multi-class classification scheme

摘要: Osteoarthritis is characterized by the degeneration of articular cartilage in joints. We have developed a fully automatic method for segmenting knee MR scans based on supervised learning. A binary approximate kNN classifier first roughly separates from background voxels, then three-class assigns one three classes to each voxel that classified as classifier. The resulting sensitivity and specificity are 90.0% 99.8% respectively medial compartments. show an accurate segmentation achievable using low-field scanner.