ScanSoft TextBridge Pro Millennium Business Edition, Руководство пользователя

Страница: 115 / 130

Advanced Sample Sessions
6–11
11. Click the Auto button.
The Get Pages dialog box appears.
12. In the Get Pages dialog box, double click Scanning.tif.
TextBridge gets the file and processes it using the settings defined
in My Newsletter, including the zones saved in the template. The
Save As dialog box appears
13. Save the document as ScanningNews.rtf.
TextBridge formats and saves the document.
S
ESSION
3: T
RAINING
T
EXT
B
RIDGE
OCR
To assure the highest possible accuracy, TextBridge provides an
interactive training capability. This feature enables you to
participate in the OCR process and train TextBridge by verifying
correctly recognized words and correcting recognition errors. With
training, TextBridge achieves higher accuracy for this specific
page and any other pages like it.
Interactive training is especially effective for degraded documents,
such as faxes and multi-generation photocopies.
As you correct or accept TextBridge’s recognition decisions for a
page or two, you also train the program to improve its own
accuracy rate for later pages of the document. It is recommended
that you complete at least one full page of a multi-page
document to sufficiently train TextBridge about character
shapes and sizes for that document.
In addition, you can save and later reload training data to
assure that other documents of the same type are recognized with
the same high degree of accuracy.
In this session you learn to use interactive training to improve
OCR.
☞
For this session, use the image file named fax.pcx . This faxed
memo has a single column of text with fax headings.