Click the Define button to start “teaching” the classifier. Select a method of evaluating the binary objects (Manual, Nearest Neighbour, Bayes, Neural Network):
The classification is based on the measurement feature selected in the nearby pull-down menu, and the limit values of each class shall be put to the Classes table.
For these non-manual methods, become the additional options are available:
The default set of features contains only the main object characteristics: EqDiameter, Circularity, MeanIntensity, HueTypical, MeanSaturation.
This option modifies the list of features in the Features pull-down menu.
Specify the number of classes by pressing
and 
buttons.Select a class that you are going to define within the Classes table.
Pick reference objects within the image using the picker
.Use the histogram to define objects manually.
Repeat this procedure for each class.
When you click the Test button, number of objects that match each class criteria appears in the table.
Reset all samples which were picked previously with the
button.Press the
button to create a separate binary layer out of each class of objects.Press the Define button again to finish the definition.
Other Options
Classify ND Image When working with a ND2 file, this button enables you to classify all frames.
Show Scattergram This button displays a scattergram which visually displays the measured features.