Quantify and export

You need: the reviewer dataset role or higher to export annotations and quantification; curator to export images.

Before you start

• The objects you want measured are annotated and labelled.
• The images are calibrated, unless pixels are genuinely what you want.

Steps

1. Open the dataset's quantification page.
2. Choose a quantification profile — which metrics get computed. See the metric reference; computing metrics nobody will read costs time for nothing.
3. Explore the results: summary cards for the dataset, distributions, and the label tree.
4. Drill down to per-image and per-object values to sanity-check before you trust the aggregate.
5. Export the measurements for your own analysis.

Profile selector → summary cards → a distribution → per-image drilldown → download the report.

Reading the output honestly

Missing is not zero

The contextual metrics (nn_distance, mean_knn_distance) omit contours that have no siblings, because a nearest-neighbour distance is undefined for them. A missing value there is not a distance of zero. By contrast n_children is defined for every contour, and a 0 there is a real count.

Units follow calibration

Length metrics come out in the image's physical unit, area metrics in that unit squared, and ratio / count / colour / intensity metrics are unaffected. An uncalibrated image reports pixels.