v0.3.2

• Add N2V2
• Lifted dependency requirements for compatibility with recent CSBDeep and TensorFLow versions
• Fixed tests

v0.3.1

Fix callback import.

v0.3.0

• Move from TensorFlow 1 to TensorFlow 2.
• Add test-time augmentation (tta) to prediction. By default it is turned off.

Note: TensorFlow 1 is not supported anymore.

v0.2.1

• Single Net per Channel Option
• StructN2V
• ModelZoo Export

v0.2.0

Add new functionality:

• Single U-Net per Channel
• StructN2V
• BioImage ModelZoo Export

Version 0.1.11

This releases fixes bug #73

v0.1.10

• Fix number of blind-spot computation: 0.198% of the pixels are manipulated per patch.
• Pin tensorflow and keras version numbers in README instructions
• Fix #42: Tiling during prediction keeps number of axes consistent with the input.
• Fix #45: Extract n patches from each sample S.
• Fix #43: Images are cast to 32bit float type during prediction.
• Add augmentation flat to training script entry point.
• Add BSD68 example notebook. This notebook reproduces the results reported in the paper (CVPR'19).

v0.1.8

• Add optional patch shuffling parameter
• Add progress bar to validation data preparation
• Normalize each channel separately
• Increase keras version to 2.2.5
• Relax csbdeep version requirement from 0.4.0 to 0.4.0<=csbdeep<0.5.0

v0.1.7

• Increase CSBDeep version from 0.3.0 to 0.4.0

pip release v0.1.6

This release includes:

• The data generator now sorts the filenames before images are loaded.
• This release includes entry points of the trainN2V.py and predict n2v.py scripts.