Segmenting Datasets

You can segment additional datasets with a trained classifier at any time in the Segmentation Trainer. See Training Classifiers for information about training a classifier.

You can also use the Segment with Classifier feature, which is available on the Segment tab, to preview and segment image data with a trained classifier (see Segmenting Image Data with Trained Classifiers.). This feature also lets you compare your results with the results of classifiers available in the Infinite Toolbox.

1. If required, choose Tools > Segmentation Trainer on the menu bar.

   The Trainable Segmentation dialog appears.

2. Select the required classifier and review the requirements in the Summary box. If necessary, you can filter the list of classifiers.

   See Classifier Panel for information about filtering classifiers, classifier requirements, and other options. You should note that items marked in red are required, while items marked in yellow are optional.

3. Click the Input tab.

   Information in the Input panel includes the engine used by the selected classifier and the inputs.

4. Add the required dataset or datasets to the Datasets inputs as follows:
   • Select the required dataset or datasets in the Data Properties and Settings panel and then click the Add button.
5. Preview the segmentation, recommended.
   • To preview a segmentation, first scroll to a representative image slice in the dataset and then click the Preview button at the bottom of the Segmentation Trainer dialog.

     Previews are calculated only on the visible portion of the image only. You can modify the appearance of the segmentation on the Result panel, if required (see Result Panel).

6. If required, you can modify the features tree and re-train the classifier.
7. Click the Segment button.
8. Export your results, as required.

   See Export for information about exporting segmentation results, regions, and confidence maps.