Reconstruction and Representation for 3D Implicit Surfaces

Radial Basis Function (RBF) Kernel method is currently the most useful method for carrying out 3D implicit surface reconstruction. However, fitting RBF to 3D scattered data has not been regarded as computationally feasible for large data sets. For this reason, this research conducts an in-depth investigation on implicit surface construction, along with self organizing map (SOM) network and kernel method both in theory and experiment. From research results, we can then use SOM network to obtain geometric features which describe the original model. The use of kernel methods makes calculation of the implicit surface more simple and efficient for performing broken surface reconstruction.