Dimensionality and measurement capabilities

Use this reference to distinguish plane-local processing from true volumetric processing and to check the current measurement families. OpenHCS has no global 2D/3D switch. Dimensional behaviour belongs to each processing callable and its typed inputs.

How dimensionality is declared

The effective execution domain is determined by these declarations:

variable_components

Declares which microscopy components are assembled along the outer stack axis. Selecting Z describes the component identity; it does not by itself select volumetric processing.

ProcessingContract

Declares whether the assembled runtime stack is processed plane by plane, as a whole, through an explicit flexible control, or as a volume collapsed to a plane.

Image and object-label input modes

A callable can require the complete image payload or full-stack labels. These declarations preserve a nested volume when an outer runtime plane axis also exists.

Object-label domain

PAYLOAD means one object-ID domain across the complete label payload. PLANE means independent object-ID domains for the outer planes.

Array dimensionality alone is not an execution contract.

Current segmentation boundary

Route

Current domain

Consequence

CellProfiler-compatible Watershed

Complete image payload

Accepts a 3D image domain and emits payload-scoped labels, so an object can span Z planes.

IdentifyPrimaryObjects

Plane-local

Each outer plane is segmented independently.

IdentifySecondaryObjects

Plane-local

Secondary labels remain independent between outer planes.

IdentifyTertiaryObjects

Plane-local

Tertiary labels remain independent between outer planes.

Image and object morphology

Function and footprint dependent

Volumetric variants, including ball-footprint and 3D object operations, coexist with plane-local variants. Inspect the selected callable rather than assuming that the whole family shares one domain.

Plane-local labels are not stitched into volumetric objects. Use a route that produces a payload-scoped label domain when biological identity must persist through Z.

Current volumetric measurements

MeasureObjectIntensity accepts a payload-scoped 3D label domain and emits one row per volumetric object. Its 3D fields include Z positions such as centre of mass and maximum-intensity Z.

MeasureObjectSizeShape accepts full-stack labels. For 3D objects it emits the volumetric schema, including Volume, SurfaceArea, Center_Z, BoundingBoxVolume, Z bounds, extent, Euler number, axis lengths, and equivalent diameter. It does not relabel those values as 2D area features.

Occupied-volume routes report occupied volume, surface area, and total volume; physical surface-area scaling can use declared Z, Y, and X voxel spacing.

Measurement families

The current CellProfiler-compatible catalogue includes these measurement families:

  • image and object intensity;

  • 2D and 3D object size and shape;

  • texture and granularity;

  • radial intensity distribution and Zernike features;

  • colocalisation;

  • image and object overlap;

  • neighbours, skeletons, counts, and object relationships;

  • image quality;

  • area and volume occupied; and

  • calculated and classified outputs.

Availability of a family does not imply that every function or feature in that family is volumetric. The selected callable’s declaration and output schema are the exact authority. Use the desktop function search or the MCP capability search to inspect the callable available from the connected execution server.

Validation boundary

The source-backed Official30 corpus includes a 3D monolayer pipeline with volumetric Watershed segmentation followed by MeasureObjectIntensity and MeasureObjectSizeShape. The acceptance suite compares OpenHCS execution with native CellProfiler reference outputs under the policy documented in Measurement and runtime equivalence.

The corpus is compatibility evidence for the included pipelines, not a claim that every processing function has a 3D implementation. See Data dimensions and execution axes for the dimensionality mental model.