Noise2Model

Image restoration for fluorescence microscopy with physics-based noise modeling

Contents

Quick install

Develop/edit locally:

pip install -e .

Quick start (examples)

Import core helpers:

from Noise2Model.core import get_noisy_pair  # see [Noise2Model.core.get_noisy_pair](http://_vscodecontentref_/0)

Load/display HDF5 dataset example:

  • Use the dataset class Noise2Model.data.SIDD_HDF and helper Noise2Model.data.load_and_display_hdf5_image.

Train / evaluate

  • Models live in Noise2Model.models.NMFlow.
  • Trainers are in Noise2Model.trainer.NoiseFlowGANTrainer.

Networks & utilities

  • Key network class: Noise2Model.networks.ResidualNet
  • Helpers & IO: Noise2Model.utils.FileManager

Notebooks & docs

The repository includes interactive notebooks under nbs/ (source for documentation). The package exposes an nbdev index via setup entry points configured in setup.py.

Contributing

Follow the repo conventions in settings.ini and keep code/tests in the usual locations. See the included LICENSE for license and contribution rules.

Authors

If you have any questions or need further assistance, please open an issue on GitHub or contact us directly at:

  • Project Lead: Biagio Mandracchia
  • Contributors: Inés Varon Peña, Sara Cruz-Adrados, Rosa-María Menchón-Lara