NNhealpix

This is a package to apply neural networks and convolutional neural networks to spherical signal projected on the healpix grid.

Requirements

• Healpy
• Matplotlib
• NumPy
• Numba
• SciPy
• Tensorflow 2+

If you prefer to use PyTorch instead of Keras/TensorFlow, aasensio has implemented a port of the library available here: github.com/aasensio/sphericalCNN.

Install

The code is still under development. To install, use the following command:

[sudo] python setup.py develop [--user]

To automatically install the requirements, use the following command:

pip install -r requirements.txt

Testing

To run a suite of tests, you must have either nosetests or pytest. Just run nosetests or pytest within the NNhealpix folder.

Citing

If you use this code, please cite the following paper: Convolutional neural networks on the HEALPix sphere: a pixel-based algorithm and its application to CMB data analysis, Krachmalnicoff, N. & Tomasi, M. (A&A, Aug 2019).

@article{ KrachmalnicoffTomasi2019,
	author = {{Krachmalnicoff, N.} and {Tomasi, M.}},
	title = {Convolutional neural networks on the HEALPix sphere: a pixel-based algorithm and its application to CMB data analysis},
	DOI= "10.1051/0004-6361/201935211",
	url= "https://doi.org/10.1051/0004-6361/201935211",
	journal = {A\&A},
	year = 2019,
	volume = 628,
	pages = "A129",
}

License

The library is released under a MIT license. See the file LICENSE for more information.