7.5.4.2   Teaching

This section allows you to 'teach' characters which are displayed in the window right next to the text field.

Steps for teaching:

1. Use the text field to enter characters that correspond to the segmented characters in the image. As you enter the characters into the text field, the image of the corresponding character is displayed on the right side of the text field. The characters are highlighted in the Graphics Window as well.
2. Click the button Add To Training Data.
3. Repeat this process for each sample image.
4. Click the button Train Now in the section Training.
5. When the classifier has been trained for the first time, results of the OCR are available in the text field at the bottom of the Teaching dialog to assist with further teaching.
6. If the results of the OCR are correct, click the button Suggestion to quickly add the results to the text field during further teaching.

If you want to view the results of the OCR classification, continue to the Results tab.

If you want to improve the OCR classification, you can change Basic Features that may improve a classification and train again. You can also select a different classifier and change parameters that influence single classifiers via the dialog Advanced Training Parameters and Features.

If you want to save data for archival purposes during training, you can activate Save Ground Truth. If activated, each time data is added to the training file, a copy of the source image, assistant settings, and teach data will be stored in a subdirectory next to the training file.