If implemented in your program, end users can select a training file when performing OCR processing. Training files contain accepted or corrected words that the OCR engine highlighted as questionable during one or more interactive training sessions. They improve the process of recognizing image documents similar to the one processed during the training sessions.
You should make sure that your users include a training file that is specifically designed for the type of image document they expect to OCR. Including an inappropriate training file can actually decrease recognition accuracy.
Note: OCR processing is available with Imaging for Windows Professional Edition V1.0, V1.1, and V2.0 only.