Abstract: Implicit surface fitting to a set of scattered points plays an important role in solving many
problems occurring in virtual reality and so on. This paper presents an implicit surface fitting algorithm
using radial basis functions with an ellipsoid constraint. This method can fit the data accurately by use
of information of point clouds. The fitted shape can still capture the main features of the object when
the data sets are sparse. The fitted surface may contain many redundant components and not represent
the actual shape properly with longer time when fitting large data set. Large-scale scattered data is
needed to be subdivided adaptively, these subdivided data fit simultaneously and these simpler implicit
surfaces blended smoothly. The experimental results show that the algorithm is good in preserving
feature and can improve the efficiency obviously.

Key words: implicit surfaces, radial basis function, ellipsoid constraint, blending