Robust Modelling and Analysis of Vascular Geometries from Biomedical Images

Si-Yong Yeo, Xulei Yang, Yan Nei Law, Tianxia Gong, Yi Su, and Li Cheng

Keywords

Segmentation, Biomedical , Shape Analysis, Curvatures, Vessel Geometries

Abstract

In this paper, a robust computational framework is proposed for the modelling and analysis of vascular geometries from biomedical images. The approach consists of the segmentation of vascular geometries using an active contour model and the extraction of geometric features. A robust image feature is derived based on geometric interactions between the active contour model and the image object boundaries. The derived image feature uses voxel interactions across the image domain, and gives a coherent representation of the vessel shapes in the image. The active contour model is therefore more robust to image noise and weak object edges. The regional shapes of the vessel structures are characterized using surface curvatures computed analytically based on a surface patch approximation technique. In particular, the curvature measures are derived and used in the morphological analysis of the vessel geometries. The experiments show that the computational modelling technique can be used for the robust segmentation and shape analysis of vessel structures in the images.

Important Links:



Go Back


IASTED
Rotating Call For Paper Image