Diagnosis of Connective Tissue Disorders based on Independent Component Analysis of Aortic Shape and Motion from 4D MR Images

Michael Sass Hansen, Fei Zhao, Honghai Zhang, Bjarne Kjær Ersbøll, Andreas Wahle, Thomas Scholz, Milan Sonka

    Research output: Chapter in Book/Report/Conference proceedingArticle in proceedingsResearchpeer-review

    Abstract

    Independent component analysis (ICA) is employed for com\$\backslash\$-puter-aided diagnosis (CAD) allowing objective identification of subjects with connective tissue disorder from 4D aortic MR images. Stationary independent components assist in the disease detection, which is the first implementation reported in the field of medical imaging. Prior knowledge of the source distribution is utilized using an appropriate ordering of the components. The CAD method distinguished between normal and diseased subjects with a classification accuracy of 96.8\verb+~+\$\backslash\$%.
    Original languageEnglish
    Title of host publicationThe 1st International Workshop on Computer Vision for Intravascular and Intracardiac Imaging. CVII 2006
    PublisherIEEE
    Publication date2006
    Publication statusPublished - 2006

    Cite this