Authors: Alexander Morozov, Davide Zgyatti, and Petr Popov, iMolecule, Skoltech.
Equidistant and Uniform Data Augmentation for 3D Objects.
Python library for 3D data augmentation, based on Eurecon algorithm ([1,2]).
Tests were performed on:
Easiest way to run the augmentation via Eurecon is running eurecon.py with the following command line:
python eurecon.py -r
Detailed instructions on how to run Eurecon to reproduce an augmented ModelNet40 dataset are given in augment_ModelNet40.py script.
python eurecon.py -r 0.5 -p 1 -a resources/tesselation_vertices_layer_0.txt -i examples/test.off -o examples/test_output
Morozov A., Davide Z., Popov P., "Equidistant and Uniform Data Augmentation for 3D Objects", IEEE Access, 2022.
[1] Popov, Petr, and Sergei Grudinin. "Rapid determination of RMSDs corresponding to macromolecular rigid body motions." (2014): 950-956.
[2] Popov, Petr, and Sergei Grudinin. "Eurecon: equidistant uniform rigid-body ensemble constructor." Journal of Molecular Graphics and Modelling 80 (2018): 313-319.