Nuance scansoft omnipage pro 14 Manual Do Utilizador

70

You can use training to improve recognition of special symbols such as @, 
® and © or to recognize supported accented letters more reliably. The 
purpose of training is not to teach the program to read characters from 
non-supported languages or alphabets.

OmniPage Pro 14 offers two types of training: manual training and 
automatic training (IntelliTrain). Data coming from both types of 
training are combined and available for saving to a training file.

When you leave a page on which training data was generated, you will be 
asked how to apply it to other existing pages in the document.

To do manual training, place the insertion point in front of the character 
you want to train, or select a group of characters (up to one word) and 
choose Train Character... from the Tools menu or the shortcut menu. 
You will see an enlarged view of the character(s) to be trained, along with 
the current OCR solution. Change this to the desired solution and click 
OK. The program takes this training and examines the rest of the page. If 
it finds candidate words to change, the Check Training dialog box lists 
these. Incorrect words should be re-trained before the list is approved.

For guidance on using the Train Character and Check Training dialog 
boxes, please consult their context-sensitive help or the online help topic 
Manual training and its related topics.

IntelliTrain is an automated form of training. It takes input from the 
corrections you make during proofing. When you make a change, it 
remembers the character shape involved, and your proofing change. It 
searches other similar character shapes in the document, especially in 
suspect words. It assesses whether to apply the user correction or not.

You can turn IntelliTrain on or off in the OCR panel of the Options 
dialog box. 

The following shows how IntelliTrain works, using the original image. 
Our example involves the letters c and e. With some typefaces and