NormFet Software for Approximating Normals and Feature Sizes from Noisy Point Clouds

Tamal K DeyJian Sun , Luke Molnar


NormFet  software can take a point cloud data possibly contaminated with noise and then compute the normals and feature sizes at the sample points. The output is a point cloud equipped with normals and feature sizes that can be fed into other software that need such information such asAMLS for surface reconstruction. It is important to note that, by default, normals are estimated only at a subset but sufficiently many of points which is alright for AMLS.  Feature sizes are estimated at all points. The software is based on the theory developed in the paper given below.

Codes (both binaries and source) are  available for Linux,  and Windows. Please send an email to to get the password to access the download area.

                                                   NormFet Download area


T. K. Dey and J. Sun. Normal and Feature Estimations from Noisy Point Clouds. Technical Report
OSU-CISRC-7/50-TR50, July, 2005.

Additional recommendation:

T. K. Dey, S. Goswami.  Provable Surface Reconstruction from Noisy Samples. Comput. Geom. Theory and Applications, to appear.

 Other papers

Other Software : AMLS , Cocone, Qualmesh , Segmatch

The output of the NormFet software can be viewed with GEOMVIEW

This software was written using CGAL libraries.

Disclaimer: We do not intend to be responsible for the maintenance of the software.

Copyright: Jyamiti group at the Ohio State University. No commercial use of the software is permitted without license.