Customer Reviews

Abbyy FineReader 5.0 Pro OCR

5 star:		(11)
4 star:		(4)
3 star:		(0)
2 star:		(1)
1 star:		(3)

 

 

I scan books that are in the public domain, which is a demanding task for OCR software. The bottom line is that the Abbyy FineReader Pro v5.0 is THE BEST OCR software on the market right now. It far surpasses Scansoft's Textbridge Millennium. The only thing that even gives FineReader Pro 5.0 any competition is Scansoft's OmniPage Pro 11, but FineReader is still far better. Since Scansoft bought out Textbridge (from Xerox) and bought out OmniPage (from Caere), you would think they could develop a superior product, but that's not so. Omnipage appears to be billed as an upgrade to Textbridge (I'm guessing that Scansoft is sending Textbridge to the graveyard). Scansoft claims in an advertisement that they spent a YEAR combining the programs. However, Scansoft did not even include rudimentary features of Textbridge in OmniPage 11. For instance, if you scan a book two pages at a time with the binding flat against the scanner (this is very common for older books that need to remain intact), Textbridge Millennium could split apart the pages using a dual-page mode. This was an extremely helpful if not critical feature for scanning books. However, this feature does not exist in OmniPage 11. FineReader on the other hand, can do this and more. It works great with Auto Document Feeders and it is far more accurate than OmniPage 11. Finally, FineReader's error/spell checker is once again far superior to OmniPage 11. The only single advantage that OmniPage 11 has over FineReader 5.0 is that it can import Adobe PDF documents for further OCR work. Otherwise FineReader Pro 5.0 is the way to go.

Also, regarding a former review, FineReader allows you to exclude optional-hyphens in the output file. Check in the advanced section of the output format. Including them is not a mistake. It's actually quite impressive that the program knows those that are optional-hyphens and those that are not.

Not a term I use too often, but prior to getting my hands on Abbyy's FineReader 5.0 Pro I was always under the impression that OCR software was a joke; an opinion bolstered by mine and others' experience of applications like Text-Bridge. Not FineReader. This stuff is in its ability to convert scanned documents into text files that can then be formatted in your word-processor. With FineReader 5.0 Pro I can do what would take me over a week - assuming I didn't go loopy in the process of transcribing it - in under an hour; converting about 20,000.00 words worth, into a useable MS-Word or RTF file.

FineReader makes very few mistakes. Some worth mentioning are the following:

Text hyphenated in order to allow a word to wrap in the original: This will give word splits with optional-hyphen symbols that will have to taken out individually in your word-processor file.

Quotation marks will default to text-editor style without any left and right handing: This can be corrected globally in a given document by ( in MS- Word ) using the replace all function and entering the normal quotation marks in the find and replace boxes. This applies to both single and double quotes.

Text set in columns with cross-column footnotes will create a line or two of rubbish, but the columns will be successfully interpreted as a single one in your word-processor document, providing you remember to save the processed file without any formatting. I reckon you should save all processed files without retaining the original formatting to make life easier.

You can also output a processed file as a PDF and an HTML page. Be sure to consider the settings for these export options, otherwise you can end up with an enormous PDF or an HTML page that won't work in older browsers.

I would also encourage a user to dispense with all the Wizard toolbars and get familiar with the commands via the menu bar. This is a better and more flexible way to work, and potentially less confusing. And buy a decent scanner, too. Epson do good ones such as their perfection series.

If you want an OCR application that really works, get FineReader 5.0 Pro. You'll recoup its incredibly modest cost in one job.

After looking around the Internet for information on affordable PC desktop OCR products, it seemed clear to me that Abbyy's "FineReader 5.0 Pro" was probably the best product for my needs. I have not been disappointed. I read a lot of negative "press" about OCR capabilities in general and I was not impressed by the OCR software bundled with my Canon 620P Scanner. You should also keep in mind that this was my first experience with any kind of OCR software.

My immediate requirement was to translate a five year-old, 57 page document to MS Word, so that it could be updated. The source document was lost sometime in the past. As a consequence, no one was interested in starting "from scratch" to update the document. The survivng hardcopy of the document was produced by a laser printer in black and white on 8 1/2 by 11 plain paper. The content consisted of text, logos, graphics, and a number of simple tables.

It took me approximately 2.5 hours to scan the entire document and about 2 days, full time, to complete the project. This time estimate includes the time to get familiar with the program.

Without counting the total number of words in the document, I would estimate that FineReader was 99% accurate translating the text. The only place it had any difficulty was in procesing some of the section header numbers in the table of contents (TOC). For exampe, the program translated the image "8.x" to "&.x" in several instances and also confused "I"s and "1"s several times. Given the italicized font of the characters images and the errors are easy to understand. The few errors I did find were quickly identified and easily corrected manually in MS Word.

Tables took a little more manual work. FineReader's layout analysis logic had no problem identifying all of the tables, but, as noted in the software documentation, the program may not always be able to convert table cells with multiple lines of text and preserve the table row and column alignment of the orignal document. In my case, these problems were easy to fix using the column and row identifier tool provided in the software. Once I marked the location of rows and columns on the scanned image of the table, the tables were translated correctly, requiring only minor editing "touch-up" to align text in table cells to match the original document.

The only problem I have encountered with this software product was the result of the distribution package I received. Abbyy offers a "try and buy" distribution that consists of a 3.5" floppy diskette and a CD-ROM. The combination of media provides copy protection and is not problem unless, like me, you have a laptop PC that can support either a CD-ROM or a 3.5" floppy diskette, but not both at the same time. In short, I called Abbyy's technical support line and they told to return the floppy diskette and CD-ROM to them and they would send me a single CD-ROM. I isntalled and used the single CD-ROM I received from the vendor with no further problems.

The "message" is that you should make sure you understand which distribution of the software you are purchasing. The box did state, in the system requirements section, that the program required a 3.5" diskette and a CD-ROM drive. I did not deduce, however, that both devices were required concurrently.

The manual that came with my distribution was entirely adequate for my purposes, suffering only from the occasional grammatical problem common to many manuals for products developed in countries where English is not the native language. (FineReader was developed by Abbyy, a Russian company with headquarters in Moscow.)

I am an experienced computer user and typically install a new software product and then try to use it immediately without bothering to read the documentation. I believe you could take this approach with FineReader, but I recommend reading the documentation first, particularly, if, like me, you are not familiar with the basic flow of the OCR process. The manual isn't that long and you probably won't need to read all of it, to get enough information to starting using FineReader productively.

Because I am not an experienced OCR user, I am reluctant to say that this product will meet the needs of everyone or every job. I was delighted, however, with what the product was able to do for me. As I said at the beginning of my comments, I was not disappointed by Abbyy's "FineReader 5.0 Pro".

About 5 years ago, I had a vision of converting all my magazine clippings and other research material to digital form so it would take up less room and be easier to find and consult. It took at long time to find the hardware and software to convert from vision to reality and OCR was the hardest part. FineReader is the only OCR package that could fill the bill, and I've tried them all! Compared to earlier versions, FR 5 is more stable, more capable on difficult scans, and more flexible in its features. On ordinary-hard material it will do the whole thing without intervention and with good accuracy. When the scan copy is really poor or exceptionally complex, or when the ultimate in accuracy is critical, then some hand editing of the recognition blocks and/or hand correction of the results is called for -- but FR 5 makes it as easy as possible. I do a lot of scanning of statistical data in tables and its abilities for this are truly amazing.

For my purposes I need to save the output to a PDF file, and until OmniPage 11 (more expensive and not nearly as good as FR), FR was the only consumer-oriented package that would do this in one step. It's a huge advantage.

Limitations? There are some: It "loses" my HP 6250 USB scanner sometimes (which may be a problem on the scanner end, of course). There's no control over how much it compresses images -- too much for good quality. Its ability to handle equations is very limited. It has trouble with "old-style" (non-lining) numerals. If it mistakenly recognizes something as numerical it won't let you correct it to alphabetical or vice versa. There's no way to force it to recognize everything in a block as numerical (as would be helpful sometimes for tables). It's a bit slower than FR 4 (due, I presume, to more sophisticated recognition algorithms). The user-interface is fairly complicated (unless you simply use the auto modes) -- but no more than is necessary for this very powerful package with its many options. While I'd love to see ABBYY work on all these areas, I'm absolutely convinced that they are the only people who have any idea how to make an OCR package fit for adults. If you need OCR, you need FineReader 5.0.

Unlike most OCR programs this one works. Toss out your bundled OCR software that came with your scanner. It only provides you with an opportunity to practice your typing skills. FineReader 5.0 Recognizes and learns from it's mistakes.

And it doesn't cost nearly what some other, less capable programs do.

One note... it does have some issues with Windows XP. There is a 22meg service pack available from the publisher's US website along with a registry hack to make it behave. I have used it successfully with Win98SE, WinME, and XP though.

Enjoy!

Compared to the American-produced products I have seen in the past, this is a major step forward. Very high accuracy in a wide range of languages. My only beefs are two: that it would be nice to have a feature to help to deal with double-sided pages fed through an automatic document feeder, and, that it would be nice to be able to output directly to WordPerfect and maybe some dtp programs. Otherwise, hard to see how it could get much better.

I bought Scansoft Textbridge Pro Millennium (...) at a physical store and was REALLY dissatisfied with the results of even simple scans. It was certainly NO better than the Presto PageManager 98 that came with my Epson 1200U scanner. THEN I went and read the reviews of the product here on Amazon. OOPS! I would never have purchased it if I'd read the reviews here first! When will I learn to check Amazon before I buy anything? Fortunately, the publisher agreed to accept a return.

Then I purchased ABBY FineReader 5.0 from Amazon. Well, the differences between it and Textbridge were ASTOUNDING! In side-by-side tests of a scanned newspaper article, FineReader reproduced it perfectly as an MS Word file. Textbridge's OCR engine made a jumbled mess of it all.

I've owned several different OCR programs (since the inception of OCR), Abbyy FineReader 5.0 Pro OCR is the best one I've ever used. I use it for reading both English and Greek, it is flawless. It is especially good at reading (flawlessly) small fonts (I've tested it on 8 pt. and newspaper articles). I can copy dual pages from books, and it will automatically format them in portrait, again flawlessly. This is the first OCR software I've had that requires only minor editing (only format, never extra wierd characters). I suspect the great price is because v.6 is being released.

I've been using FineReader Pro 5.0 for a week and I'm really pleased with the excellent job it does. My most critical need was to scan "grainy" pages--printouts of microfilm pages. These pages are scarred with scratches and blobs. FR does an unbelievable job of interpreting the text accurately. It lets you know where it has guessed, and it guesses right in the great majority of cases. It also highlights where guessing is impossible and lets me correct the text while looking at a blowup of the problem area. This saves me from having to locate the problem on the paper original.

Another useful feature: the ability to scan the paper sideways and rotate it inside FR.

All things considered, an extremely effective tool that gets the job done. It's hard to imagine software that would do the trick better.

One minor quibble, since nothing is perfect. There are discrepancies between the printed manual and the software. For example, the manual shows the block cutting tool as being a button on the interface. In fact, it's in one of the drop-down menus.

I bought this software, primarily on the basis of the reviews on Amazon. It installed easily...and I should note that although I've used several scanning packages, this is the first one I've bought. The capabilities of the software are truely awesome...some real virtuousity is clearly evident. Tables and text scan easily and flawlessly into every kind of file I can think of. Amazing.

The problem I have with Finereader 5.0 Pro it is really two-fold. First, the manual is pathetic...nowhere would one discover how to use the product, much less what it is capable of delivering, if s/he relied on the manual. It just doesn't make sense and has loads of omissions. Second, the email support is literally useless...I don't really think there is anything but an email address at the other end... No response, at all. I was about 15 minutes from returning the whole box when I tried something completely illogical to change the scanner settings, when, Voila! The thing started working.

Whatever, you do...don't rely on the manual. A good technical writer with several weeks' experience with this package could earn this company a lot of grateful and loyal customers. Are you ABBYY guys listening?