MaskComparison

MacroModule

genre	Statistics
author	MeVis Medical Solutions AG
package	MeVisLab/Standard
definition	Arithmetics.def
keywords	segmentation, statistics, metric

Purpose

The module MaskComparison provides three metrics for evaluating differences between two mask images.

Usage

Connect two binary images with same extent and voxel-to-world matrix to the fields inMask1 and inMask1.

Press Update to compute the metric values.

Windows

Default Panel

Input Fields

inMask1

name: inMask1, type: Image

The field inMask1 must be connected to a binary image containing only 0 and 1 values.

inMask2

name: inMask2, type: Image

The field inMask2 must be connected to a binary image containing only 0 and 1 values.

Parameter Fields

Field Index

Dice Coefficient: Float

instanceName: String

Jaccard Index / Intersection Over Union (IoU): Float

Overlap Coefficient: Float

Results are valid: Bool

Update: Trigger

Visible Fields

Dice Coefficient

name: dice, type: Float, persistent: no

Shows the Dice coefficient of the two input masks.

The Dice coefficient is computed using the formula:

2 × ∑(A ∩ B) / (∑(A) + ∑(B))

Overlap Coefficient

name: overlap, type: Float, persistent: no

Shows the overlap coefficient of the two input masks.

The overlap coefficient is computed using the formula:

∑(A ∩ B) / min(∑(A), ∑(B))

Jaccard Index / Intersection Over Union (IoU)

name: jaccard, type: Float, persistent: no

Shows the Jaccard coefficient of the two input masks.

The Jaccard coefficient is computed using the formula:

∑(A ∩ B) / ∑(A ∪ B)

Results are valid

name: resultsValid, type: Bool, persistent: no

Shows whether the results are valid.

Update

name: update, type: Trigger

When pressed, the metric results are updated.

Hidden Fields

instanceName

name: instanceName, type: String, default: MaskComparison