Skip to content

Features#

PyPI

CAREamics is an on-going project and the API is not yet stable.

Officially supported features (v0.0.15)#

  • Installation from conda and mamba
  • Data formats
    • in-memory arrays
    • tiff files
  • Training
    • training in memory or from multiple files
    • choose number of epochs and steps per epochs
    • WandB and TensorBoard monitoring
  • Prediction
    • Tiled prediction
  • Algorithms
    • Content-aware image restoration (CARE) [ref]
    • Noise2Noise [ref]
    • Noise2Void [ref1, ref2]
    • N2V2 [ref]
  • napari UI

Current work#

  • Next-generation dataset
    • data formats:
      • Zarr
      • CZI
    • flexible, modular
    • arbitrary data organization
    • on-line statistics estimation
    • data masking
    • background patch rejection
  • Prediction
    • direct tile writing (e.g. Zarr)
  • Algorithms
  • Stand-alone UI