Premium
Abstractive Representation and Exploration of Hierarchically Clustered Diffusion Tensor Fiber Tracts
Author(s) -
Chen Weri,
Zhang Song,
Correia Stephfan,
Ebert David S.
Publication year - 2008
Publication title -
computer graphics forum
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.578
H-Index - 120
eISSN - 1467-8659
pISSN - 0167-7055
DOI - 10.1111/j.1467-8659.2008.01244.x
Subject(s) - diffusion mri , fiber bundle , cluster analysis , computer science , visualization , fiber , fractional anisotropy , artificial intelligence , segmentation , orientation (vector space) , representation (politics) , tensor (intrinsic definition) , fiber tract , pattern recognition (psychology) , computer vision , mathematics , geometry , materials science , magnetic resonance imaging , political science , medicine , politics , law , composite material , radiology
Diffusion tensor imaging (DTI) has been used to generate fibrous structures in both brain white matter and muscles. Fiber clustering groups the DTI fibers into spatially and anatomically related tracts. As an increasing number of fiber clustering methods have been recently developed, it is important to display, compare, and explore the clustering results efficiently and effectively. In this paper, we present an anatomical visualization technique that reduces the geometric complexity of the fiber tracts and emphasizes the high‐level structures. Beginning with a volumetric diffusion tensor image, we first construct a hierarchical clustering representation of the fiber bundles. These bundles are then reformulated into a 3D multi‐valued volume data. We then build a set of geometric hulls and principal fibers to approximate the shape and orientation of each fiber bundle. By simultaneously visualizing the geometric hulls, individual fibers, and other data sets such as fractional anisotropy, the overall shape of the fiber tracts are highlighted, while preserving the fibrous details. A rater with expert knowledge of white matter structure has evaluated the resulting interactive illustration and confirmed the improvement over straightforward DTI fiber tract visualization.