The influence of gradient estimation on the extraction of boundary point cloud

Abstract

To extract a tubular object boundary from a volumetric image is important to compute its morphometric properties, like the estimation of the boundary curvature, or the radius of a tubular object, for example, the radius is one of the descriptions of blood vessel for detecting the soft plaque. How to estimate the gradient of the volumetric data has an influence on the computation results of the morphometric properties. Extract the points of maximum gradient along the gradient direction in 3D as the boundary point cloud of an object is used by [4]. The boundary points are computed by trilinearly interpolated the volumetric datasets, and apply the parabolic interpolation to find the maximum gradient along the gradient direction. The extraction of boundary point cloud depends on the estimation of the image gradients. This paper is to compare the tricubic B-spline and trilinear interpolation algorithms on the estimations of morphometric properties of the volumetric dataset.