Neural mesh ensembles

This work proposes the use of neural network ensembles to boost the performance of a neural network based surface reconstruction algorithm. Ensemble is a very popular and powerful statistical technique based on the idea of averaging several outputs of a probabilistic algorithm. In the context of surface reconstruction, two main problems arise. The first is finding an efficient way to average meshes with different connectivity, and the second is tuning the parameters for surface reconstruction to maximize the performance of the ensemble. We solve the first problem by voxelizing all the meshes on the same regular grid and taking majority vote on each voxel. We tune the parameters experimentally, borrowing ideas from weak learning methods.

[1]  Bernd Fritzke,et al.  Growing cell structures--A self-organizing network for unsupervised and supervised learning , 1994, Neural Networks.

[2]  John C. Platt,et al.  Elastically deformable models , 1987, SIGGRAPH.

[3]  Hans-Peter Seidel,et al.  Using growing cell structures for surface reconstruction , 2003, 2003 Shape Modeling International..

[4]  Hans-Peter Seidel,et al.  Neural meshes: statistical learning based on normals , 2003, 11th Pacific Conference onComputer Graphics and Applications, 2003. Proceedings..

[5]  Yizhou Yu.  Surface Reconstruction from Unorganized Points Using Self-Organizing Neural Networks , 1999 .

[6]  Hans-Peter Seidel,et al.  Multi-level partition of unity implicits , 2003, ACM Trans. Graph..

[7]  Hans-Peter Seidel,et al.  Neural meshes: surface reconstruction with a learning algorithm , 2004 .

[8]  Mario Botsch,et al.  Feature sensitive surface extraction from volume data , 2001, SIGGRAPH.

[9]  William E. Lorensen,et al.  Marching cubes: A high resolution 3D surface construction algorithm , 1987, SIGGRAPH.

[10]  I. P. Ivrissimtzis W-K.  Neural Meshes: Statistical Learning Methods in Surface Reconstruction , 2003 .

[11]  Georgios Papaioannou,et al.  A Fast Depth-Buffer-Based Voxelization Algorithm , 1999, J. Graphics, GPU, & Game Tools.

[12]  David W. Opitz,et al.  Generating Accurate and Diverse Members of a Neural-Network Ensemble , 1995, NIPS.

[13]  R. Schapire.  The Strength of Weak Learnability , 1990, Machine Learning.

[14]  Tony DeRose,et al.  Surface reconstruction from unorganized points , 1992, SIGGRAPH.

[15]  Alex Pentland,et al.  Closed-form solutions for physically-based shape modeling and recognition , 1991, Proceedings. 1991 IEEE Computer Society Conference on Computer Vision and Pattern Recognition.

[16]  Yoav Freund,et al.  Experiments with a New Boosting Algorithm , 1996, ICML.

[17]  Demetri Terzopoulos,et al.  Regularization of Inverse Visual Problems Involving Discontinuities , 1986, IEEE Transactions on Pattern Analysis and Machine Intelligence.

[18]  Arie E. Kaufman,et al.  Alias-Free Voxelization of Geometric Objects , 1999, IEEE Trans. Vis. Comput. Graph..

[19]  Chandrajit L. Bajaj,et al.  Automatic reconstruction of surfaces and scalar fields from 3D scans , 1995, SIGGRAPH.

[20]  Marc Levoy,et al.  Fitting smooth surfaces to dense polygon meshes , 1996, SIGGRAPH.

[21]  Robert E. Schapire,et al.  The Boosting Approach to Machine Learning An Overview , 2003 .

[22]  Richard K. Beatson,et al.  Reconstruction and representation of 3D objects with radial basis functions , 2001, SIGGRAPH.

[23]  Teuvo Kohonen,et al.  Self-organized formation of topologically correct feature maps , 2004, Biological Cybernetics.

[24]  Anath Fischer,et al.  Adaptive reconstruction of freeform objects with 3D SOM neural network grids , 2001, Proceedings Ninth Pacific Conference on Computer Graphics and Applications. Pacific Graphics 2001.

[25]  Marshall W. Bern,et al.  A new Voronoi-based surface reconstruction algorithm , 1998, SIGGRAPH.

[26]  Marek Teichmann,et al.  Surface reconstruction with anisotropic density-scaled alpha shapes , 1998, Proceedings Visualization '98 (Cat. No.98CB36276).