Skip to content

IQUANA documentation

Intelligent QUantification, ANnotation and Analysis — a tool for AI-assisted segmentation, annotation and quantification of scientific image datasets, built at DFKI in partnership with HIFMB.

Most annotation tools stop at the mask. IQUANA carries the same objects through measurement and review, so what you export is an analysis rather than a pile of polygons — and every number stays traceable to the object it came from.

A point prompt lands on an object, the mask blooms in, it is accepted and labelled.

Where to start

The loop

IQUANA is built around a cycle rather than a labelling queue:

Step What happens
01 Upload Datasets with typed per-image metadata you can filter and group by.
02 Predict Run segmentation models across a dataset, one model per label if you need it.
03 Correct Review and correction queues surface what the model most likely got wrong.
04 Quantify Per-object metrics in physical units, driven by configurable profiles.
05 Export Measurements, provenance and label hierarchy out to your own analysis.

Each pass through the loop improves the annotations, and the corrected annotations are what you fine-tune the next model on.

Getting help

  • Something is broken, or you want a feature — please use the issue templates. Bug reports for the frontend, backend, AI service and installer all belong in the same tracker.
  • Something will not start — begin with Troubleshooting.
  • You want to bring IQUANA to your own domain, or are interested in a research collaboration — write to robert.leist@dfki.de.

We are a small research team and cannot promise to implement every request.