A Data Clustering and Streamline Reduction Method for 3D MR Flow Vector Field Simplification

• Main Authors: Carmo, Bernardo Silva (Author), Ng, Y H Pauline (Author), Prügel-Bennett, Adam (Author), Yang, Guang-Zhong (Author)
• Other Authors: Barillot, C (Contributor), Haynor, D R (Contributor), Hellier, P (Contributor)
• Format: Article
• Language: English
• Published: 2004.
• Subjects: Article
• Online Access: Get fulltext

With the increasing capability of MR imaging and Computational Fluid Dynamics (CFD) techniques, a significant amount of data related to the haemodynamics of the cardiovascular systems are being generated. Direct visualization of the data introduces unnecessary visual clutter and hides away the underlying trend associated with the progression of the disease. To elucidate the main topological structure of the flow fields, we present in this paper a 3D visualisation method based on the abstraction of complex flow fields. It uses hierarchical clustering and local linear expansion to extract salient topological flow features. This is then combined with 3D streamline tracking, allowing most important flow details to be visualized. Example results of the technique