juglab / n2v

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

Channel last in Prediction #37

Closed tibuch closed 5 years ago

tibuch commented 5 years ago

The channel-wise normalization expects the images to have channel last. This is not necessarily given during prediction. Moving the channel to the last position in n2v.predict() should fix this.