ShapeProperties

ShapeProperties computes deterministic size and shape descriptors for each non-zero label in a 2D label image.

Inputs are label_image. Outputs include label, area, pixel-edge perimeter, bbox_area, extent, aspect_ratio, and equivalent_diameter.

Use it when segmentation outputs need lightweight morphology features without adding heavier measurement dependencies.

Minimal Example

from bioimageflow_core import Arguments
from bioimageflow_measurement_tools import ShapeProperties

rows = ShapeProperties().process_row(Arguments(label_image="labels.tif"))

Inputs

• label_image: 2D label image with background 0.

Outputs

• one dataframe row per non-zero label.

• object_count: number of measured objects.

• table columns: label, area, pixel-edge perimeter, bbox_area, extent, aspect_ratio, and equivalent_diameter.

Dependencies and Core Libraries

imageio, NumPy, pandas, and deterministic package-local geometry helpers.

Assumptions

Labels are integer object IDs and background is zero. The current implementation is intentionally deterministic and 2D-oriented.

Expected Results

Synthetic rectangle fixtures produce exact areas, bounding boxes, aspect ratios, and equivalent diameters.

Failure Modes

Unreadable images or unsupported dimensions fail through the image reader or measurement code. Empty label images return an empty table and zero objects.