Markov Random Field Surface Reconstruction (MRFSurface) is a surface reconstruction and remeshing method. It has been tested on modestly sized datasets and the results are comparable to current state-of-the-art algorithms. In particular, it has been shown to be well suited for hole filling in surface scans of human faces. Currently, a compiled version is available for free use - or more preciseliy on a citeware basis. Use the software and be kind to cite us.

Recently, (October 2010) we have also released our combined alignment and reconstruction framework. It is included in the MRFSurface software package. Take a look at the examples page for usage.


The software is released as citeware. For the pure surface reconstruction the following reference is appropriate:

Rasmus R. Paulsen, Jakob Andreas Bærentzen, Rasmus Larsen, Markov Random Field Surface Reconstruction, IEEE Transactions on Visualization and Computer Graphics, pp. 636-646, July/August, 2010

For the combined alignment and surface reconstruction the following citation should be used:

Rasmus R. Paulsen and Rasmus Larsen. Anatomically Plausible Surface Alignment and Reconstruction. Proc. Eurographics UK conference on Theory and Practice of Computer Graphics. pp. 249-254, 2010.


The Oticon Foundation is thanked for funding Rasmus R. Paulsen. The 3D Craniofacial Image Research Laboratory at the School of Dentistry, University of Copenhagen is thanked for the use of their 3dMD scanner.