People who are blind or visually impaired have greater access to printed material than ever before, but a vast amount of material still remains unavailable in accessible formats. Optical character recognition (OCR), the process of transforming inaccessible printed and digital documents into accessible text, allows blind and visually impaired people to access material that would otherwise be unavailable to them. Kurzweil 1000 and OpenBook, both evaluated previously in AccessWorld are two popular OCR products, but each costs nearly $1,000. In this article, I look at two alternatives to high-cost OCR software: the ABBYY Fine Reader Professional version 11 ($169.99) and Text Cloner Pro version 11.5 ($99.95). Both programs require one gigabyte of RAM and both are compatible with Windows operating systems from Windows XP onward.
I tested each program using a variety of document types:
- Standard print documents printed from digital documents created with a word processor on 8.5 by 11 inch paper (some misspellings were intentionally included in these documents)
- Text-only pages from a standard paperback book
- Magazine articles with columns and images
- A scanned table from a book
- A utility bill printed on a small sheet of paper
I used a Plustek Opticbook 3800 as my scanner. I tested each program using NVDA, JAWS, and Window Eyes.
ABBYY Fine Reader Professional
Installation
To install ABBYY Fine Reader Professional, you can either download an install file from the ABBYY Fine Reader website or purchase the product on CD. There is also a 15-day free trial available for download on the Fine Reader Professional webpage. You'll need to provide some personal information (such as name and profession) before downloading the free trial, and you'll be limited to saving a maximum of 50 pages during the trial. The install program is very accessible as it is a regular Windows install program without any deviations from the standard format. Screen readers have no issue installing the program.
Documentation
Documentation for the ABBYY Fine Reader can be found packaged with the program (located in the program folder in the Guides sub-folder) or downloaded from the Fine Reader Guide webpage in accessible PDF format. The program itself also contains a very accessible Windows help file containing the contents of the guide. The guide is meant for a sighted audience, referring often to locations of elements and directing you to use the mouse to accomplish tasks. It does not, however, rely on inaccessible diagrams for instructions or images of controls, which makes the guide useful for a visually impaired or blind user.
The Scanning Process
When you first launch the program you will be presented with the New Task dialogue box where you can choose the type of scan you would like to perform. The first List Box allows you to choose the output format for the scanned material. You can choose between several formats including DOC, PDF, and e-book formats. The document will be opened in both the program that displays your chosen format and in the Fine Reader program itself.
Once you have selected your format, you must select several other options including language and color mode. The choices for color mode are Black and White and Color. Selecting Black and White returns results more quickly. For DOC, PDF, and e-book formats you must also choose a variant on the document type. For example, you can choose between EPUB, FB2, and HTML for the e-book format. Once you have selected your options, you can activate the "Scan" button to be taken to the Scanning dialogue box.
The Scanning dialogue box contains several options for adjusting the scanner settings. These include the Paper Size, Brightness, Color Mode, and Resolution. You can also indicate if you would like the program to check the orientation of the page being scanned, and whether you'd like to separate facing pages into single pages in your output file. Once you have made your selections, you can either activate Preview to create a quick image scan, or activate the "Scan" button to scan an image for OCR processing. After the scan is complete the Scanning dialogue box will reappear so that you can scan another page. To exit and view your OCR results, you can press the "Close" button. Depending on which output file type you have chosen, you either will be taken to the Fine Reader editing screen, or the program that is associated with your selected file type will be launched with a copy of the OCR results.
Scanning Results
Overall, the scanning results for ABBYY were overwhelmingly positive. A document printed on standard 8.5 by 11 inch paper was perfectly rendered. Misspelled words (such as "tke" for "the") were recognized as written. The scanning results were also nearly perfect when pages from a book were scanned. Multiple errors in one scan were due to shifting the book around during the scanning process. Further scans of the book were error-free.
The magazine article with multiple columns was rendered in a single column, with the second column following the first. The text was perfectly rendered. The text of a magazine article with an image was also rendered perfectly with the image recognized accurately. Fine Reader worked very well when scanning a table from a book. When the document was saved in DOC and HTML formats, the individual cells were rendered in a screen-reader navigable form.
The only document type that caused an issue was the utility bill. The bill was rather small and contained a bar code that was rendered as a line of the letter "I" in the OCR results.
Screen Reader Accessibility
The scanning process is the only portion of Fine Reader that is fully accessible. The main editing interface, on the other hand, is almost entirely inaccessible. Though it's possible to enter the main editing window where the text is displayed, many letters will be missing from the OCR result and there is no accessible way provided to correct the text. Therefore, I will focus my assessment on the accessibility of the scanning process, since it's possible to use Fine Reader if you save your results in another format.
NVDA had the best accessibility results. Almost all of the options in the New Task dialogue box read correctly and NVDA even read the explanation of each variant for those formats that had a multiple options. The "Help" button was unlabeled and the selections in the Color Mode list box were not displayed, but otherwise all the controls were rendered perfectly. In the Scanning dialogue box NVDA recognized all of the items. NVDA also read the dialogue boxes for saving documents correctly.
JAWS did not read the format types in the New Task dialogue box. It also did not read the description of the format variants when these were highlighted. All of the other buttons and options were read correctly accept for the "Help" button, which again was unlabeled. The Scanning dialogue box was correctly read in its entirety, and the dialogue boxes for saving files were also correctly read.
Window Eyes did not read the format selections in the New Task dialogue box. It also did not read the "Help" button, or the variant selections. It did read the label on the combo box for selecting Color or Black and White, but it did not read the options themselves. It did read the "Scan" button, the "Close" button, and the checkbox that determines if the New Task dialogue box appears when you start the program. Window Eyes also struggled with the Scanning dialogue box. The combo boxes that select Resolution,Scanning Mode,Brightness, and Paper Size were not read correctly. The labels of these boxes were read but the content was not. Window Eyes does recognize the buttons for "Scan," "Preview," "Restore Defaults," and "Close." It also correctly announces changes in the Brightness slider. Compared to its poor performance on the Scanning and New Task dialogue boxes, Window Eyes reads perfectly in the Saving dialogue boxes.
Text Cloner Pro
Installation and Interface
The Text Cloner install package, which uses the standard Windows format,is very accessible using a screen reader. It's possible to download a demonstration version of Text Cloner from the Premier Literacy Free Trials webpage. The complete version of Text Cloner can be purchased either as a download or as a CD from the Text Cloner Pro product webpage.
When you launch the program you will be presented with a blank document where you can either scan or type content. Text Cloner acts similarly to a basic text editor, and offers a list of common menus: File, Edit, View, etc., as well as program specific menus such as Scan and Table. There are very few program-specific hotkeys and these are restricted to activating the scanning process and the document summarization feature.
Documentation
The documentation is accessible from the Help menu. It can be accessed in a Windows-style help file or as a Word document that can be opened in Text Cloner. You can also find the Word document in the Text Cloner Pro installation folder where you can use your processor of choice to read it. The document provides a quick reference for keyboard shortcuts; most of the shortcuts listed are standard in Windows programs. For example Ctrl+N opens a new document and Ctrl+C copies highlighted text. The only program specific hotkeys mentioned are the scanning keys: F5 for black-and-white scanning and F6 for color scanning, as well as the shortcut Ctrl+Shift+U for document summarization.
After the Quick Reference section, the manual describes each menu in detail. The document is quite comprehensive, listing every option and its use. Frequent typos throughout the document such as missing words or added words made some instructions difficult to understand and detracted from the professional appearance of the program.
Scanning
There are two different scanning processes available in Text Clone. Black-and-white is used for books and text-only materials. Color is used for more complex material that includes images and tables. You can begin a scan by pressing F5 for black and white and F6 for color. Each of these choices has its own set of options that can be accessed from the Scanning menu. The options for each scanning process optimize the scan for the type of material being scanned. For example, the black-and-white scan provides the option of scanning a magazine, book, or other material. The color scanning process provides options to scan text/images/tables, text/images, or text only. The Scan menu also provides a scanner selection dialogue box labeled Select Scanner Driver, where you can select the TWAIN scanner of your choice. When you scan either with black-and-white or color, the program automatically begins recognition after the image scan has completed. You can also import a PDF file or image file for processing if the material that you need to scan is digital. The supported image file types are GIF, TIF, BMP, JPG, and FMF.
Scanning Results
Printed documents on 8.5 by 11 inch paper scanned rather well. Some letters did not appear properly but everything was legible and there weren't any full lines of unrecognizable text. The document with purposeful misspellings was rendered perfectly by the program with the misspellings intact. I used the black-and-white scanning process with the Other option checked in the material option submenu. Scanning book pages produced passable results. Every scan had a few errors where letters were not recognized correctly. Repeated scans would correct errors in one area but introduce them in another.
With the table, the scan was very good and almost free of errors. I was able to save the table in HTML format and use my screen reader's table navigation features to navigate it. For a low-cost OCR solution, Text Cloner Pro is impressive in its ability to correctly render tables. I used the color scanning process and the Text, Tables, and Images option when scanning this material.
For the multi-column magazine article, Text Cloner Pro provided an excellent result with proper formatting of the columns, one beneath the other for easier reading, and the image correctly recognized and placed. I used the color process with the Text and Images option selected.
An electric bill was the only type of material that caused significant issues. The bill contents were not displayed correctly in the proper format even after several scans to obtain the optimum image. The portion of the bill that contained only text was rendered correctly but the portion with the amount of the bill and other pertinent information was extremely poorly rendered. That side of the bill did contain a bar code, which may have contributed to these issues.
Editing
Text Cloner Pro is a somewhat robust text editor. The area where content appears after being typed or scanned is similar to what you'd find in WordPad or Microsoft Word. Many common editing options are included in the Edit menu such as Find and Replace,Thesaurus, and Spell Check. The program supports headers and footers as well as bulleted lists. You are also able to create your own tables using the Table menu and insert images and text from other documents using the Insert menu. The program also includes the ability to change fonts and set special attributes on the paragraph level.
PDF Conversion and Summarization
Text Cloner Pro is able to perform OCR on PDF files and to summarize a document. In the Scan menu there is an option labeled Import PDF that allows you to import a PDF for processing. You must first select the PDF using the "Select PDF" button in the Import PDF dialogue. Next you must determine the pages on which you wish to perform OCR. After this you must press the "Process" button to begin OCR on the PDF file. The PDF conversion dialogue remains on the screen until conversion is complete, whereupon you are placed back in the document. The accuracy of conversion will depend on the quality of your PDF file. I discovered that some PDF files scanned incredibly well while others contained some errors.
Summarization takes a document and removes much of the content, displaying a percentage of the whole. For example, several sentences will be displayed followed by several sentences from a paragraph several pages later in the document. The summarize dialogue contains a button to begin summarization and a slider to establish the percentage of the file that will appear in the summary.
Screen Reader Accessibility
NVDA did not perform well when using Text Cloner. It would read letters multiple times when editing letter-by-letter or word-by-word. When reading by line, often words would be left out. Because of these issues, if you're using NVDA, it's necessary to save your document before you can properly edit it. NVDA also could not read the slider that determined the percentage to display when summarizing a file and the labels on the edit boxes that determined the width and height of tables in the Table Creation dialogue box. Otherwise, NVDA performed well; menus were read correctly as were other dialogue boxes.
Jaws and Window Eyes both performed extremely well with Text Cloner Pro. Both screen readers allowed navigation by character and word without issue and all dialogue boxes were read correctly. Both screen readers announced items in nearly an identical way so that using either works equally well.
The Bottom Line
Text Cloner Pro and ABBYY Fine Reader provide optical character recognition for a fraction of the cost of products such as OpenBook and Kurzweil 1000. They do, however, have limitations that aren't found in the more expensive products. Text Cloner Pro has an accessible interface if you are either using Window Eyes or Jaws, but its recognition results often are poor and require much editing or rescanning to acquire a perfect image. In contrast, ABBYY Fine Reader has excellent scanning results but an interface that is extremely inaccessible. At the moment, ABBYY Fine Reader seems the best choice since you can use work arounds to provide an accessible scanning experience. In the future, there is the chance that Fine Reader will become even more inaccessible than it is now. Text Cloner Pro will always remain accessible because its purpose is to provide a screen-reader accessible OCR experience. Text Cloner Pro should see improvements in OCR accuracy in the future making it a better choice when considering OCR products for their consistency and longevity.
Product Information
Product: ABBYY Fine Reader Professional
Price: $169.99
Available from: ABBYY
Phone:866-463-7689
Product: Text Cloner Pro
Price: $99.95
Available from: Premier Literacy
Phone: 815-927-7390
