Overlap Measures

Class: NodeMaskOverlapMeasures

Node Icon

Compute Jaccard overlap, Dice coefficient, volume similarity, false negative error and false positive error similarity of two or more binary masks.

Venn diagram

The measures are defined as follows:

\(\textrm{Jaccard} = \displaystyle 2\frac{S \cap T}{S\cup T}\)

\(\textrm{Dice} = \displaystyle 2\frac{S \cap T}{S + T} = \frac{2\times Jaccard}{1+Jaccard}\)

\(\textrm{Volume Similarity} = 2 \displaystyle \frac{S-T}{S+T}\)

\(\textrm{False Negative Error} = \displaystyle \frac{T \setminus \!\!S}{S}\)

\(\textrm{False Positive Error} = \displaystyle \frac{S \setminus \!\!T}{T}\)

Example Workflows

Overlap measures example

Inputs

Masks

At least two binary masks of the same size.

Type: Image4DBool, Required, Multiple (Minimum = 2)

Outputs

Measures

A Data Table with seven columns: Mask 1, Mask 2, Dice, Jaccard, Volume Similarity, False Negative Error, False Positive Error

Type: DataCollection

Settings

Slicewise Boolean

Perform the analysis slice by slice. Each slice will result in one row in the output result data table.

Slice Direction Selection

If slicewise analysis is selected this specifies the orientation of slices. E.g. if slice orientation is XY, slice interpolation will be in the z-direction.

Values: XY, XZ, YZ