Skip to content
PhenoMePhenoMe

Select properties

Property functions extract numerical features from images and masks. For what each built-in column means, see Property interpretation. For preset bundles (basic, standard, …) and mixing in custom code, stay in this section.

Select properties covers:

  • Quantifying phenotypes (cell size, intensity, shape) via presets or your own functions.
  • Using classical features as distance sources alongside embeddings.
  • Domain-specific metrics and correlation with embedding space.

Decision tree

Section titled “Decision tree”
Situation Read next
You want a ready set of features with one parameter Presets
You need a domain-specific function Writing functions
You need GLCM, OpenCV, multi-channel tricks, or troubleshooting Advanced recipes

Minimal example

Section titled “Minimal example”
import numpy as np


def mean_intensity(image2d, mask2d):
    if image2d is None:
        return {}
    return {"mean_intensity": float(np.mean(image2d))}


df = pheno.compute_properties(
    property_preset="basic",
    additional_property_functions={"image": [mean_intensity]},
)

See also

Section titled “See also”