BundleMAP: Anatomically localized classification, regression, and hypothesis testing in diffusion MRI

详细信息    查看全文

• 作者：Mohammad Khatami^a ; ^{khatami@cs.uni-bonn.de}Author Vitae ; Tobias Schmidt-Wilcke^b ; ^{tobias.schmidt-wilcke@bergmannsheil.de}Author Vitae ; Pia C. Sundgren^c ; ^d ; ^{pia.sundgren@med.lu.se}Author Vitae ; Amin Abbasloo^a ; ^{abbasloo@cs.uni-bonn.de}Author Vitae ; Bernhard Schö ; lkopf^e ; ^{bs@tuebingen.mpg.de}Author Vitae ; Thomas Schultz^a ; ^{schultz@cs.uni-bonn.de}Author Vitae
• 关键词：Disease detection ; Manifold learning ; Support vector machines ; Classification ; Regression ; Fiber tracking ; Diffusion MRI
• 刊名：Pattern Recognition
• 出版年：2017
• 出版时间：March 2017
• 年：2017
• 卷：63
• 期：Complete
• 页码：593-600
• 全文大小：2042 K
• 卷排序：63

文摘

Manifold learning is used to achieve a joint parametrization of fiber bundles from diffusion MRI. Diffusion parameters can be plotted along the bundle. Anatomically localized and interpretable features are extracted. Increased accuracy for supervised classification and regression is demonstrated. Increased power for hypothesis testing is shown.