10.28. Compute Feature Clustering

Group (Subgroup)

Statistics (Morphological)

Description

This Filter computes the radial distribution function (RDF) of a set of Features. A Feature is a contiguous region of like-segmented cells (for example a grain or particle). An RDF is a histogram that describes how the Features are spatially distributed relative to each other: for each distance bin, it counts how many other Feature centroids lie at that separation distance, then normalizes that count against the count expected if the same number of Features were scattered randomly in the same volume. Values above 1.0 at a given distance indicate the Features are more likely to be found at that separation than random (clustering or ordering); values below 1.0 indicate they are less likely (exclusion). Use this Filter to detect and quantify clustering, ordering, or repulsion in a population of Features.

An Ensemble is a group of Features that share a common phase (a distinct material or crystal structure). This Filter operates on a single Ensemble at a time, specified by the Phase Index parameter, and the resulting RDF is stored as Ensemble data.

All distances in this Filter are physical lengths expressed in the same length units as the Image Geometry spacing (for example microns). The algorithm proceeds as follows:

  1. Find the straight-line (Euclidean) distance from the current Feature centroid to every other Feature centroid of the selected phase.

  2. Put all calculated distances in a clustering list.

  3. Repeat steps 1-2 for all Features.

  4. Sort the distances into the specified number of bins, all equally sized in distance from the minimum separation to the maximum separation. For example, if the user chooses 10 bins, the minimum separation is 10 microns, and the maximum separation is 80 microns, each bin spans 7 microns.

  5. Normalize the RDF by the probability of finding the Features at that distance if they were distributed randomly in the bounding box.

The Filter also outputs the clustering list (every inter-Feature distance) and the minimum and maximum separation distances (both in physical length units).

Note: Because the algorithm iterates over all Features, each distance is double-counted. For example, the distance from Feature 1 to Feature 2 is counted along with the identical distance from Feature 2 to Feature 1.

Required Input Sources

Input Parameter(s)

Parameter Name

Parameter Type

Parameter Notes

Description

Selected Image Geometry

Geometry Selection

Image

The target geometry

Number of Bins for RDF

Scalar Value

Int32

Number of bins to split the RDF

Phase Index

Scalar Value

Int32

Ensemble number for which to calculate the RDF and clustering list

Remove Biased Features

Bool

Remove the biased features

Random Number Seed Parameters

Parameter Name

Parameter Type

Parameter Notes

Description

Set Random Seed

Bool

When checked, allows the user to set the seed value used to randomly generate the points in the RDF

Seed Value

Scalar Value

UInt64

The seed value used to randomly generate the points in the RDF

Stored Seed Value Array Name

DataObjectName

Name of array holding the seed value

Input Feature Data

Parameter Name

Parameter Type

Parameter Notes

Description

Phases

Array Selection

Allowed Types: int32 Comp. Shape: 1

Specifies to which Ensemble each Feature belongs

Centroids

Array Selection

Allowed Types: float32 Comp. Shape: 3

X, Y, Z coordinates of Feature center of mass

Biased Features

Array Selection

Allowed Types: uint8, boolean Comp. Shape: 1

Specifies which features are biased and therefore should be removed if the Remove Biased Features option is on; True values removed

Input Ensemble Data

Parameter Name

Parameter Type

Parameter Notes

Description

Cell Ensemble Attribute Matrix

AttributeMatrixSelection

The path to the cell ensemble attribute matrix where the RDF and RDF min and max distance arrays will be stored

Output Feature Data

Parameter Name

Parameter Type

Parameter Notes

Description

Clustering List

DataObjectName

Distance of each Feature’s centroid to every other Feature’s centroid

Output Ensemble Data

Parameter Name

Parameter Type

Parameter Notes

Description

Radial Distribution Function

DataObjectName

A histogram of the normalized frequency at each bin

Max and Min Separation Distances

DataObjectName

The max and min distance found between Features

Example Pipelines

  • PorosityAnalysis

DREAM3D-NX Help

If you need help, need to file a bug report or want to request a new feature, please head over to the DREAM3DNX-Issues GitHub site where the community of DREAM3D-NX users can help answer your questions.