fast::OtsuThresholding class

Otsu thresholding segmentation method.

Contents

Automatically determines the threshold to use for segmentation using Otsu's method. Supports multi-class thresholding by changing the number of classes argument in the constructor (default is 2 classes = 1 threshold). Maximum number of classes is currently 4.

The implementation uses 256 bins when calculating the histogram. Input images that are not of UINT8 type are normalized to [0, 255] range and cast to UINT8 before processing.

Inputs:

  • 0: Image (only first channel is used, multi-channel support not implemented).

Outputs:

  • 0: Image segmentation

Base classes

class ProcessObject
Abstract base class for all process objects.

Private functions

void execute() override
Create instance.

Function documentation

void fast::OtsuThresholding::execute() override private

Create instance.

Returns instance