Welcome to Svetlana’s documentation!

Svetlana is a Napari plugin developed in Python, dedicated to the classification of segmentation masks. It has mainly been developed for biological imaging, but it can be applied to any kind of 2D, 3D or multichannel segmented image.

The classifier is based on light-weight convolutional neural networks and is efficient with few annotations. You can learn more about our method reading the related paper.

First check out the Installation section to know how to install it.

A Youtube tutorial video is available here. You can also have a look at Usage to know more about advanced features. Two demo images, identical to the ones of the tutorial video, are available in the annotation plugin using the button TRY ON DEMO IMAGE. Feel free to use them to familiarize yourself with the plugin.

If you use this plugin, please cite the paper:

Cazorla, C., Weiss, P., & Morin, R. (2023). Svetlana: a Supervised Segmentation Classifier for Napari.

  TITLE = {{Svetlana: a Supervised Segmentation Classifier for Napari}},
  AUTHOR = {Weiss, Pierre and Cazorla, Cl{\'e}ment and Morin, Renaud},
  URL = {https://hal.inria.fr/hal-03927879},
  NOTE = {working paper or preprint},
  YEAR = {2023},
  MONTH = Jan,
  PDF = {https://hal.inria.fr/hal-03927879/file/main_nature.pdf},
  HAL_ID = {hal-03927879},
  HAL_VERSION = {v1},


This project is still under active development. We welcome any suggestions for improvement from users.

Thank you