Support several (optional) normalizations

nshaud / DeepHyperX

Deep learning toolbox based on PyTorch for hyperspectral data classification.

Other

439 stars 122 forks source link

Support several (optional) normalizations #34

Open nshaud opened 4 years ago

nshaud commented 4 years ago

Currently the dataset is normalized into (0,1). This is mostly fine but we should be able to use other normalizations or even no normalization at all.

TODO:

[ ] Add a normalization option
[ ] Support multiple normalizations (to define)

nshaud commented 4 years ago

@jfrob27 what sort of data normalization do you use on astro data?

mengxue-rs commented 3 years ago

Please see https://github.com/nshaud/DeepHyperX/pull/38#issue-532363161