search

juglab / n2v

This is the implementation of Noise2Void training.
Other
392 stars 107 forks source link

One Net per Channel #33

Closed tibuch closed 4 years ago

tibuch commented 5 years ago

Add a 'super-conservative' parameter which has one U-Net per channel. Essentially multiple U-Nets per model.

tibuch commented 4 years ago

https://github.com/juglab/n2v/tree/315f1b36c7f542c21d1d77e1411fb4795976b47a implements the single unet per channel option.

This still has to be tested and an example should be added.