OMERO-Screen Napari

Napari widgets for browsing plates, inspecting cells and labelling training data.

AI / Agents

Developers

hh65

omero-screen-napari

Interactive Napari widgets for exploring high-content microscopy data stored on an OMERO server, generating cell galleries, and building training datasets for machine learning classifiers.

Status

Version:

What this package does

omero-screen-napari adds four widgets to the Napari image viewer that cover the full workflow from raw plate data to annotated training datasets:

Widget	Purpose
Welldata Widget	Load and visualise well images from an OMERO plate, with caching and stitching support
Gallery Widget	Extract individual cell crops as a montage grid and define new classifiers
Training Widget	Annotate crops with class labels and save sessions to disk
Aligned Plate Widget	Overlay images from multiple spatially registered plates

Two companion command-line tools work without opening Napari: omero-train manages the training database, and omero-screen-images renders cell galleries and whole-well overviews (for QC and figure panels) with a JSON manifest of every run.

Documentation

Full user documentation (including workflow guides for non-technical users) is available at the omero-screen documentation site.

Quick links: - Installation - Welldata Widget — loading images and stitching - Gallery Widget — cell crops and classifier creation - Training Widget — annotating cells - Session Manager & Direct Load - omero-train CLI reference - omero-screen-images — galleries and well overviews without Napari

Installation

This package is part of the omero-screen monorepo. The recommended way to install it is via the workspace:

git clone https://github.com/HocheggerLab/omero-screen.git
cd omero-screen
uv sync

To install the napari plugin standalone:

uv pip install omero-screen-napari

After installation, start Napari and the four widgets will appear under Plugins → Omero Screen Napari.

OMERO connection

The plugin connects to an OMERO server using credentials from a .env file in the project root:

USERNAME=your_omero_username
PASSWORD=your_omero_password
HOST=your_omero_server

omero-train CLI

A command-line tool for managing training databases without opening Napari:

omero-train list                    # show all classifiers
omero-train stats mitosis-rpe       # detailed stats for a classifier
omero-train export mitosis-rpe      # export annotations to CSV
omero-train delete mitosis-rpe      # delete a classifier and its data

See CLI reference for full details.

omero-screen-images CLI

Galleries and whole-well overviews without Napari, reproducible from a script or notebook:

omero-screen-images gallery 5108 --wells E2,G5 --classifier-column classifier_nuclei4 \
    --class micronuclei --channels DAPI --grid 5x5 --json
omero-screen-images well 5108 --wells B2,E2 --layers DAPI,Tub,nuclei_masks --zoom 4
omero-screen-images batch plan.csv --channels DAPI

See omero-screen-images for the workflows.

Authors

Created by Helfrid Hochegger — hh65@sussex.ac.uk

License

MIT — see LICENSE for details.