# References
This section provides a list of example Cytomine tasks you can use as a starting point or for testing.
# Dummy Tasks
These are tasks to showcase different parameter types used by the app engine.
How to use
All dummy tasks live in the cytomine/dummy-tasks (opens new window) repository. Download the zip archive of the task you want to test directly from GitHub, then upload it to your Cytomine instance via the app engine interface.
| Task | Parameter Type | Repository |
|---|---|---|
| Array Identity | Integer array | identity integer array (opens new window) |
| Boolean Identity | Boolean | identity boolean (opens new window) |
| Datetime Identity | Datetime | identity datetime (opens new window) |
| Enum Identity | Enum | identity enum (opens new window) |
| File Identity | File | identity file (opens new window) |
| Geometry Identity | Geometry | identity geometry (opens new window) |
| Image Identity | Image | identity image (opens new window) |
| WSIDICOM Image Identity | WSI DICOM image | identity wsi dicom image (opens new window) |
| WSIDICOM Image Array Identity | WSI DICOM image array | identity wsi dicom image array (opens new window) |
| Integer Identity | Integer | identity integer (opens new window) |
| Number Identity | Number (float) | identity number (opens new window) |
| String Identity | String | identity string (opens new window) |
# Other Tasks
These are real-world task integrations combining algorithms and models for image analysis workflows. They are good references for understanding how to structure a non-trivial Cytomine task.
# VALIS experiment
Repository: maxime915/app_engine_valis_experiment (opens new window)
Performs image registration using the VALIS (opens new window) library. It aligns multiple whole-slide images onto a common coordinate space, enabling cross-slide comparison and analysis.
# Stardist task
Repository: cytomine/task-stardist (opens new window)
Detects and segments cell nuclei using the StarDist (opens new window) deep learning model. StarDist represents nuclei as star-convex polygons, making it well-suited for densely packed nuclei in fluorescence and H&E images.
# TIA Centre Cytomine tasks
Repository: TissueImageAnalytics/cytomine-app (opens new window)
A collection of Cytomine tasks developed by the TIA Centre (opens new window), University of Warwick, wrapping pre-trained models from TIAToolbox (opens new window). The repository currently contains two tasks:
- cytomine-hovernet: nucleus instance segmentation and classification using HoVer-Net (opens new window), trained on the PanNuke dataset.
- cytomine-kongnet: nucleus detection using KongNet, trained on the MONKEY Challenge (opens new window) dataset.