Mesh statistics for robust curvature estimation
In this project we were investigating the accuracy of curvature estimators for triangle meshes. We have created a benchmarking framework that allows pefroming large scale experiments with randomized data. The framework, together with the source code, can be downloaded here:
http://graphics.zcu.cz/graphics_filesystem/curvature/CurvatureBenchmark.zip
The results will be presented at the SGP 2016 conference in the following paper:
Libor Váša, Petr Vaněček, Martin Prantl, Věra Skorkovská, Petr Martínek, Ivana Kolingerová: Mesh statistics for robust curvature estimation
There is also a supplementary material available, showing some of the models used in the experiment.
Since the publication, we have found a few bugs. Results of a rerun of the experiment can be found here.
Note: We are currently planning to re-run the experiment with updated settings, if you have comments on how to change it to better server its purpose, please let us know by contanting Libor Váša, lvasa@kiv.zcu.cz.
![](graphics_filesystem/curvature/repreimage.jpg)