Python 3 software to locally classify crystalline structures in 3D pointcloud data as retrieved from moleculardynamical simulations, colloids or complex plasmas.
- Calculation of a feature vector for local classification of crystalline structures as described in Dietz et al.
- Training of a neural network with artificial crystal lattices of fcc, bcc, and hcp
- Numpy
- Scipy
- Scikit-learn
- Pandas
- Sympy
- Numba
- Multiprocessing (optional)
- Download the repository.
- Install the anaconda Python 3 distribution
- The package should work out of the box in a standard anaconda installation
If requested, I might create a complete Anaconda package for this repository.
If you use this package for a publication, we would be very happy to be cited:
Dietz, C., T. Kretz, and M. H. Thoma. "Machine-learning approach for local classification of crystalline structures in multiphase systems." Physical Review E 96.1 (2017): 011301.