Python implementation of the Mixed-Scale Dense Convolutional Neural Network.
If you use this code in a publication, we would appreciate it if you would refer to:
- Pelt, D. M., & Sethian, J. A. (2018). A mixed-scale dense convolutional neural network for image analysis. Proceedings of the National Academy of Sciences, 115(2), 254-259. link
If you use this code to improve tomographic reconstruction, we would appreciate it if you would refer to:
- Pelt, D. M., Batenburg, K. J., & Sethian, J. A. (2018). Improving Tomographic Reconstruction from Limited Data Using Mixed-Scale Dense Convolutional Neural Networks. Journal of Imaging, 4(11), 128. link
Development of the Mixed-Scale Dense Convolutional Neural Network method was supported by CAMERA, jointly funded by The Office of Advanced Scientific Research (ASCR) and the Office of Basic Energy Sciences (BES) within the United States Department of Energy's Office of Science. Development of the Python implementation is supported by Centrum Wiskunde & Informatica (CWI), with financial support provided by The Netherlands Organisation for Scientific Research (NWO), project number 016.Veni.192.235.
Installation
To install this code in a conda environment, run:
conda install -c conda-forge msdnetIn other environments, the code can be installed by running:
python setup.py installThe code requires the following Python modules: numpy, scipy, scikit-image, psutil, h5py, tqdm, numba >=0.41 . For compiling the code, the scikit-build module is required.
To run on GPU (recommended), a CUDA-capable GPU must be present and CUDA drivers must be installed. In addition, please make sure that the version of the cudatoolkit package installed by conda matches the CUDA version of your drivers. Specific versions of cudatoolkit can be installed by running (where 'X.X' is the CUDA version, e.g. '10.0'):
conda install cudatoolkit=X.XUsage
Please see the included example scripts for usage information. The scripts and further documentation can be found here: https://dmpelt.github.io/msdnet/.