CLIJ2

Logo

GPU accelerated image processing for everyone

CLIJ2 home

trainWekaModelWithOptions

Trains a Weka model using functionality of Fijis Trainable Weka Segmentation plugin.

It takes a 3D feature stack (e.g. first plane original image, second plane blurred, third plane edge image)and trains a Weka model. This model will be saved to disc. The given groundTruth image is supposed to be a label map where pixels with value 1 represent class 1, pixels with value 2 represent class 2 and so on. Pixels with value 0 will be ignored for training.

Default values for options are:

Categories: Segmentation, Machine Learning

Availability: Available in Fiji by activating the update sites clij and clij2. This function is part of clijx-weka_-0.30.1.19.jar.

Usage in ImageJ macro

Ext.CLIJx_trainWekaModelWithOptions(Image featureStack3D, Image groundTruth2D, String saveModelFilename, Number trees, Number features, Number maxDepth);

Usage in object oriented programming languages

Java
// init CLIJ and GPU
import net.haesleinhuepf.clijx.CLIJx;
import net.haesleinhuepf.clij.clearcl.ClearCLBuffer;
CLIJx clijx = CLIJx.getInstance();

// get input parameters
ClearCLBuffer featureStack3D = clijx.push(featureStack3DImagePlus);
ClearCLBuffer groundTruth2D = clijx.push(groundTruth2DImagePlus);
int trees = 10;
int features = 20;
int maxDepth = 30;
// Execute operation on GPU
CLIJxWeka2 resultTrainWekaModelWithOptions = clijx.trainWekaModelWithOptions(featureStack3D, groundTruth2D, saveModelFilename, trees, features, maxDepth);
// show result
System.out.println(resultTrainWekaModelWithOptions);

// cleanup memory on GPU
clijx.release(featureStack3D);
clijx.release(groundTruth2D);
Matlab
% init CLIJ and GPU
clijx = init_clatlabx();

% get input parameters
featureStack3D = clijx.pushMat(featureStack3D_matrix);
groundTruth2D = clijx.pushMat(groundTruth2D_matrix);
trees = 10;
features = 20;
maxDepth = 30;
% Execute operation on GPU
CLIJxWeka2 resultTrainWekaModelWithOptions = clijx.trainWekaModelWithOptions(featureStack3D, groundTruth2D, saveModelFilename, trees, features, maxDepth);
% show result
System.out.println(resultTrainWekaModelWithOptions);

% cleanup memory on GPU
clijx.release(featureStack3D);
clijx.release(groundTruth2D);

Back to CLIJ2 reference Back to CLIJ2 documentation

Imprint