Review: Russian import has knack for scanning complex documents
The American economy is grounded on the principle that competition is good. Certainly, consumers will cheer the introduction of ABBYY Software House's FineReader Pro 5.0 desktop optical character recognition package to U.S. markets.
This Russian import (the company's American division is ABBYY USA) does certain things—particularly recognition and layout reproduction of complex pages—better than the market-leading OmniPage Pro 11 from ScanSoft Inc., and we can expect this competition in the desktop OCR market to spur both products to new heights.
Although FineReader and OmniPage both do a virtually flawless job of recognizing standard typed pages in good condition, we also test OCR products with complex pages, including color graphic elements, tables, labels and vertical text. In addition, we test OCR programs with degraded documents, such as faxes that have been crumpled.
We found that FineReader did a somewhat better job than OmniPage of retaining formatting of the complex documents we used in testing. Neither program passed the test without numerous mistakes. FineReader's mistakes, however, were less noticeable and easier to correct. The program did a decidedly better job of matching font types and sizes and of including rules between columns.
In our testing, we found that both programs did a comparable—and generally praiseworthy—job of dealing with degraded documents.
FineReader, however, offers several features you won't find in OmniPage. For starters, the program can recognize text that runs vertically down the page. Indeed, we found that it generally did a good job of accurately recognizing vertical text in tables, though it was much less effective in recognizing vertical text in page layouts or graphic elements.
We also appreciated FineReader's ample toolbox for editing pages. We were especially pleased by its unique eraser tool, which makes it a snap to clean up unwanted artifacts such as misrecognized or misplaced elements.
FineReader is also a little more worldly than OmniPage when it comes to language support. Specifically, FineReader offers spell checking in nearly twice as many languages as Omni.Page (34 vs. 18) and can handle documents in 176 languages, compared to 114 for OmniPage. What's more, FineReader recognizes a number of the major scripting languages (Basic, C/C++, Fortran, Java and Pascal) as well as simple formulas, feats beyond the capabilities of OmniPage. Like OmniPage, FineReader can save documents as PDF files. FineReader cannot, however, perform OmniPage's trick of recognizing PDF files and turning them into word processing files. (Readers should note, however, that in our recent review of OmniPage Pro 11 we found that the program did not do a particularly good job of converting PDF files to word processing files. See the review on Page 40 of the July 30 issue of Federal Computer Week.)
Both OmniPage and FineReader are scalable to accommodate the heavy OCR demands of busy federal agencies and departments by supporting batch processing, distributed processing (in which scanning is done on one computer, while recognition is processed on another) and multiprocessor support. With OmniPage, those capabilities are bundled into the standard $499 version. ABBYY, on the other hand, offers users a choice: If you need those capabilities, you can shell out $399 for FineReader Office. If you don't, you can save your money and spend only $99 for FineReader Pro. So why should any department even consider OmniPage? We did find OmniPage a bit more intuitive to use than FineReader. Also, in addition to recognizing PDF files, OmniPage offers two capabilities you won't find in Fine.Reader: voice readback and job scheduling. The latter feature is of special concern to many federal agencies and departments.
If job scheduling and PDF recognition are not must-have features for your organization, however, FineReader is well worth a look.
NEXT STORY: Videoconferences go wireless