It goes without saying that we would all like to save a few dollars when it comes to purchasing assistive technology and that we are sometimes frustrated by the difference in cost of products designed for blind people compared to the cost of commercial technologies. For example, optical character recognition (OCR) packages designed for blind people—such as OPENBook from Freedom Scientific or Kurzweil 1000 from Lernout & Hauspie—retail for $995, while the leading commercial product, OmniPage Pro 11.0 from ScanSoft, retails for $495. This article examines the features and functionality of OmniPage Pro Version 11.0 as an OCR alternative in contrast with the blindness products, and offers customization hints to increase the accessibility of OmniPage Pro for those who would like to try it.
Quick Tour of OPENBook and Kurzweil 1000
Blind people generally use OCR in order to access print materials such as correspondence, books, and magazines. You typically place a document on the scanner, press a key and wait while the page is scanned and recognized. By default the page is then read aloud to you in high quality synthesized speech. As more pages are scanned and recognized they are read aloud. Graphics are discarded because they cannot be converted to text. All commands are accessed through conventional Windows menus and via the keyboard. Other applications including braille translation software can be launched from OPENBook or Kurzweil 1000, and both offer word processing and other unique features as part of their user interface.
Quick Tour of OmniPage Pro
Sighted people use OCR to manipulate and manage images and text. For example, OCR is often used to convert documents such as brochures, forms, and manuals to and from Adobe's Portable Document Format (PDF). Processing a document typically includes: scanning an image or importing one from a file, performing recognition on both text and graphics associated with the image, proofreading and correcting the text against the original image, and saving the corrected results to one of many supported formats.
OmniPage utilizes three on-screen work areas described as: Original Image, Document Manager and Text Editor. Some commands are accessed through conventional Windows menus, whereas others are located in a menu on a toolbar which must be clicked with a mouse or with the mouse movement keys of a screen reader.
Use with Screen Readers
OPENBook and Kurzweil 1000 are self-voicing applications with full speech support for all functions of scanning, recognition, and reading. This makes them ideal for users who do not have a screen reader or who are not comfortable working with Windows applications.
I evaluated OmniPage Pro with JAWS for Windows 3.7 (JFW) from Freedom Scientific and with Window-Eyes 4.1 from GW Micro on a Pentium 600 with a Hewlett-Packard 4P scanner. Both screen readers accurately tracked menu options, correctly identified the active work area, and read the status of items such as check boxes and grayed menu options. Both screen readers were unable to identify the Toolbar, and I was unaware of this feature until someone told me it was there. JAWS includes scripts for OmniPage 10, which provide keyboard access to the toolbar. However, changes in Version 11 caused these scripts to work unreliably in some cases. For example, after accessing the toolbar with the JAWS keystroke, I was unable to return to the Text Editor using the keyboard. Therefore, with both screen readers I found it most efficient to access the toolbar through the following steps: 1) move the mouse cursor to the toolbar and click the mouse, 2) use the Tab key to move between menu items and use cursor keys to select options, and 3) move the mouse cursor back to the Text Editor and route the pointer to the mouse. This may sound complicated, but it is actually a typical process when application developers do not provide keyboard shortcuts for menu or toolbar options. Thus, users need to be familiar with access strategies for Windows applications, including the use of screen reader mouse emulation keys.
Installation
OPENBook and Kurzweil 1000 should provide speech prompts throughout the installation process. My experience, however, is that installation can be problematic with any self-voicing application for a variety of reasons, and even experienced users may need technical assistance to get everything working together. Installation of OmniPage Pro requires knowledge of screen reading commands, since this application does not talk automatically. It was possible to identify all choices and edit fields with the exception of the three edit fields for the license key. This key is printed on the CD jacket and must be entered to complete installation. The computer will be rebooted if OmniPage needs to configure the Windows Installer. Otherwise, the process is almost entirely automatic. Product registration can be completed online during installation or from the web site.
Documentation and Support
Both OPENBook and Kurzweil 1000 provide extensive online help and product documentation in electronic and other formats. All documentation is written to be understood by a blind person with dialogs and prompts described in terms of what you will hear through the synthesizer.
Documentation for OmniPage Pro is supplied in print and as a PDF file on the CD. I converted the PDF file to text using the recognition features of OmniPage. There were a significant number of errors both in recognition and decolumniation, which affected readability of the document. I also sent the manual as an e-mail attachment to the Adobe PDF conversion site at pdf2txt@adobe.com with significantly better recognition and decolumniation results. The manual makes extensive use of diagrams, screen shots, and pictures of key labels in place of text descriptions. As a result, it was difficult to determine how most features worked without experimentation, and it was not possible to use the tutorial that was referenced throughout the manual. OmniPage does provide context-sensitive help, which I found very useful. It also has a Tips section in the help menu, which was accessible. When all is said and done, though, most of my help came from other blind users on e-mail lists who answered my questions and offered their tricks for using the program.
Customizing OmniPage Pro
I made the following changes to the default settings in OmniPage Pro to optimize its use with a screen reader:
- Uncheck the box for displaying Tips on the opening screen. I found them useful but annoying.
- Open the View menu with Alt-V. Uncheck the Ruler, Document Manager, and Original Image. This leaves only the Text Editor on the screen.
- From the View menu select Zoom and change the value to 100%. The Text Editor will now be a full screen view.
- Open the Tools menu with Alt-T. Select Options then select All. This is a multi-tab dialog box. Each tab can also be accessed directly from the Options menu. Choose settings to your liking unless otherwise specified.
- In the Direct OCR tab, make sure that Direct OCR is checked. This allows OmniPage to be customized and launched from a word processor or text editor. If the desired application is installed but not listed, it can be added here. When a registered application is opened, OmniPage will add itself to the File menu of that application.
- On the Proofing tab, disable Automatic Proofreading. This feature can be started manually from the Tools menu but was very difficult to use because recognition errors were not displayed in context. Saving recognition results in a text file before running the Proofreader makes it easier to determine errors, but the process was still very slow.
- In the Custom Layout tab select No Graphics.
- In the Text Editor tab uncheck the boxes for non-printing characters and markers. If checked, screen readers see spaces and hard returns as ANSI characters and will not read them correctly. (Every line will be interpreted as a word and spelled.)
- When options are set, select Save and enter a file name with no extension. Then select OK.
- From the View menu select OmniPage Toolbox and then select Auto OCR. Each page will be scanned, recognized, and saved. There is also an option for Manual OCR, which allows processing of multiple pages.
- Move the mouse cursor to the Toolbar. Locate the word "Scan" or the word "Load" (whichever is visible) and click on this word to access the menu. From the Save menu select Clipboard to avoid being prompted for a filename after each page. Results are stored and can be saved at any time during the session, even if Clipboard is chosen.
- Leave the Toolbar by moving the mouse back to the Text Editor and routing the cursor to it.
- Open the Process menu with Alt-P. Options will vary, depending on whether Auto or Manual OCR is enabled. If the steps just given have been followed, pressing Start will begin the scanning process.
Making A Choice
OPENBook and Kurzweil 1000 were designed specifically for use by blind people. The developers have worked to make a complex process very easy and accessible by providing a simple yet powerful interface and clear documentation. The cost may be well worth it for those who are not comfortable with Windows applications and for those who primarily want to scan and read on demand.
Configuring and learning OmniPage Pro was challenging and frustrating at times because of the cluttered screen display, array of options that vary in accessibility, inconsistency of keyboard access, and visually oriented documentation. But for those with tenacity, patience, curiosity, and plenty of time to experiment, OmniPage Pro may be well worth the effort in terms of cost and features.
Product Information
Product: OPENBook 5.0.
Manufacturer: Freedom Scientific Blind/Low Vision Group, 11800 31st Court North, St. Petersburg, FL 33716; phone: 800-444-4443 (U.S. and Canada); fax: 727-803-8001 e-mail: info@freedomscientific.com; web site: http://www.arkenstone.org/main.html. Price: $995.
Product: Kurzweil 1000 5.0.
Manufacturer: Kurzweil Educational Systems Group, Lernout and Hauspie Speech Products N.V., 52 Third Avenue, Burlington, MA 01803; phone: 800-894-5374 or 781-203-5000; fax: 781-203-5033; e- mail: education.info@lhsl.com; web site: www.lhsl.com/education/. Price: $995 with FlexTalk speech, $1,195 with DECtalk speech.
Product: OmniPage Pro 11.0.
Manufacturer: ScanSoft, Inc. 9 Centennial Drive, Peabody, MA USA 01960; web site: www.caere.com.
