Training Incorrect Characters

Use Train Incorrect Characters to train incorrect characters in a segmented object, or to avoid training characters individually. Although you must enter character values for all segmented objects in Correct String, OCR trains only objects that do not have a match or have an incorrect match in Text Read.

For example, if you analyze an image that contains the letters A, P, and R and you train these letters, saving them to a character set file, you can later use the character set file to train characters on another image. If the second image contains the letters A, P, R, and O, OCR displays the recognized characters, A, P, and R, and replaces O with the substitution character in Text Read. You can use Train Incorrect Characters to train O without selecting it individually.

You also train incorrect characters when OCR displays the wrong character value for a segmented object. For example, if an ROI contains the letters A, P, R, and O and Text Read includes the letter B instead of P, you can use the Train Incorrect Characters option to correctly train the letter P.

Complete the following steps to train incorrect characters:

1. Access the OCR Training Interface.
2. Select File»Open Images, and select the image or images you want to use for training. You can select multiple image files by pressing the <Ctrl> key and clicking each file. You can also enable the Select all files checkbox to open all images in the directory you specified.
3. Click Open.
4. Use the navigation buttons to locate the image you want to use for training.
5. On the image, draw an ROI around the characters you want to train.

   OCR segments objects in the ROI, drawing character bounding rectangles around them according to the settings on each of the tabs at the bottom of the training interface. Text Read displays recognized characters and the substitution character based on the character set file you are using. If you have not opened a character set file, Text Read displays the substitution character for each of the segmented objects in the ROI. For example, if the ROI contains three segmented objects, Text Read contains three substitution characters. Any object that is surrounded by a character bounding rectangle is a segmented object.

6. Use the Threshold, Advanced Threshold, Size & Spacing, and Read Options tabs to set up the parameters you want to use in the training process. Adjust threshold methods and settings and make changes in the other tabs to configure OCR to draw character bounding rectangles around objects in the ROI appropriately. OCR displays objects in blue.
7. Click Train Incorrect Characters.
8. Enter the appropriate character values in Correct String, including the previously trained and recognized characters, and click Train.