Image segmentation by spatial fuzzy clustering

Antonio Vergari (antonio.vergari@uniba.it) & Francesco Tangari 

Laura Caponetti (laura.caponetti@uniba.it) & Giovanna Castellano (giovanna.castellano@uniba.it)

This plugin was implemented in ImageJ

RGB image

1.38x (used to develop this plugin)

Put the sfcm_clustering.jar into the ImageJ plugins folder and re-launch it. Under the menu Plugin>Segmentation

you will find four new plugins:

- Jarek Sacha's K-Means Clustering the original K-Means plugin by Jarek Sacha

- K-Means Clustering

- Fuzzy C-Means Clustering

- Spatial Fuzzy C-Means Clustering

This plugin implements the following algorithms for color image segmentation:

• Jarek Sacha's K-Means Clustering the original K-Means plugin by Jarek Sacha

• K-Means Clustering a refactored K-Means version that allows the user to chose the color space, the initialization criterion for the centroid matrix and some new visualization methods

• Fuzzy C-Means Clustering plugin that segments an image using the Fuzzy C-Means

• Spatial Fuzzy C-Means Clustering implementing a spatial version of the Fuzzy C-Means, more robust to noise

Here is a brief description: 

1. Select a RGB color  image; 

2. Click on the plugin  command; 

3. Change, if it is the case, the parameters of the plugin configuration by sliders; 

4. Click on the button "OK" to apply the  segmentation process with the specified parameters.