Particle–based T-Spline Level Set Evolution for 3D object reconstruction with Range and Volume Constraints

Abstract

We consider an evolution process for implicitly defined surfaces, which are represented as the zero–levels of T-spline functions. The paper presents two novel contributions. First, we will use particles on the evolving surface in order to discretize the evolution equation. In particular we describe criteria for local and global resampling, which are needed in order to maintain a sufficiently uniform distribution of the particles. Second, we discuss volume and range constraints which can be added to the framework. More precisely, it is possible to specify a fixed volume (volume constraint) or to define a region which should or should not be contained in the final object (range constraint). These constraints can also be regarded as a priori knowledge of the data.