Training Italic decimal numbers in NI Vision OCR Training Interface

The NI Vision OCR Training Interface is used to train characters, save character sets, and verify characters by comparing them to a reference character. This article will go through how to train decimal number that are in Italic style and gets overlapped with the nearby number which creates difficulty in training individual characters using NI Vision OCR Training Interface.

Requirements:

• Vision OCR Training Interface

Navigate to Start -> Vision OCR Training and double click on that to open the Vision OCR Training Interface Application as shown below:

Go to File -> Open Images to load the image.

When training an italic character like the below image.

The difficulty is faced while training the numbers as they overlap on each number when training and decimal point makes it more difficult to train.

You can overcome this problem by training two sets where one set consists of only italic numbers (1,2,3,4,5,6,7,8,9,0) and other set containing italic number followed by decimal point (1. ,2. ,3. ,4. ,5. ,6. ,7. ,8. ,9. ,0. ).

Once trained confirm whether all the characters are identified correctly, similar to the below example:

If the Text Read is wrong similar to the below image.

In order to troubleshoot this, follow the instructions below:

1. Select the particular set of number as highlighted in below image.

2. Check for the Text Read and identify the incorrect character. From the below image the incorrect character is identified as “5.”

3. Select the incorrect character.

4. Check whether the Text Read still reads incorrect character. Select “Train Incorrect Characters” and enter the correct string and click “Train”.

5. Now repeat step 1 and check if the Text Read correct string. If not repeat steps 2 to 5 until the issue gets resolved.

Note: The more you train the characters the more precise will be the result.

After successful training, you should be able to read all characters correctly similar to the below result when following the steps below.

Highlight all characters in the image.

Check for Text Read for correct characters.