OtsuThreshold¶

MLModule¶

genre	`Segmentation`
author	`MeVis Medical Solutions AG`
package	`MeVisLab/Standard`
dll	`MLSegmentation`
definition	MLSegmentation.def
keywords	`segmentation`, `threshold`

Purpose¶

The module OtsuThreshold computes automatically a threshold according to Otsu and segments the image into 0 (lower than the threshold) and 1 (greater than the threshold).

A background threshold value can be set to let the module work only on a mask.

Details¶

See:

N. Otsu, A Threshold Selection Methods from Gray-Level Histograms,
IEEE Trans. Systems, Man and Cybernetics, 1979, Vol. 9, 62--66

Tips¶

Does only work with unsigned int 16 data type as input.

Windows¶

Default Panel¶

../../../Modules/ML/MLSegmentation/mhelp/Images/Screenshots/OtsuThreshold._default.png

Input Fields¶

input0¶

name: input0, type: Image¶: Input will be converted internally to unsigned int 16 for the histogram computation.

Output Fields¶

output0¶

name: output0, type: Image¶: Output is a binary segmentation of the image according to the computed threshold. Output image has unsigned int 16 data type.

Parameter Fields¶

Field Index¶

Apply: Trigger

Auto Apply: Bool

Background Threshold: Integer

Bin Size: Integer

Threshold:: Integer

Visible Fields¶

Bin Size¶

name: binSize, type: Integer, default: 128¶: Sets the number of internally used bins for the histogram computation.

Apply¶

name: apply, type: Trigger¶: If pressed, the module computes anew.

Auto Apply¶

name: autoApply, type: Bool, default: FALSE¶: If checked, the module computes anew on each field change.

Background Threshold¶

name: backgroundThreshold, type: Integer, default: -1¶

Sets a threshold for suppressing background values.

If this is set to a value < 0, this threshold has no effect (since the module only works on unsigned integer images).

Threshold:¶

name: threshold, type: Integer, default: 0¶: Shows the automatically computed threshold for the Otsu segmentation.