Desperate for recognition




In my seventy-fourth year I am having a little difficulty making the jargon of computers stick in the grey cells but, with the monthly help of your magazine, I am staying abreast of most of the items that fall within my interests. I enjoy your Help Screen columns and am endeavouring to compile an index to enable me to find those items which have rescued me from the depths of despair on many occasions.
I have been struggling with an Epson GT-5000 Scanner with Epson Scanit (Twain) program, which the distributors assured me would produce Word documents through MS-Word. I can produce a picture of the document, but cannot make Word convert that picture into a .doc. file. Maybe I am asking too much. The main object is to streamline my filing system. I am the secretary for several charitable organisations in the district, and I wish to write a letter and when the reply arrives place it on the same file as the original. When I try to do this I find it takes heaps of memory and I get two letters per floppy.
My equipment consists of a 486DX2-66 with 8Mb RAM, 515Mb hard disk and Windows 95. I have plenty of available memory. I would be grateful if you could tell me whether that which I want to do is achievable.
- Vic Fifield


What you are trying to achieve is quite possible with the setup that you have. The problem is that you are taking a snapshot, like a photocopy, of the documents and storing them on your disk, rather than interpreting the picture. It's not Word itself that does the conversion. In order to scan a document as text, you need optical character recognition (OCR) software, which is now available quite cheaply and works quite well.
This type of software analyses your documents as they are scanned, looking for patterns that it can recognise as typed characters. It then outputs these to a file that can be edited by Microsoft Word or any other text editor. You will find that documents stored as words, instead of pictures, take up far less space on the disk. If some documents are badly printed or handwritten, you might still have to store the scanned picture, as the OCR software may not be able to translate them into text. You should always run the spelling checker over a scanned document to correct any minor glitches.
My favourite OCR software is OmniPage, but it is quite expensive. I actually use Recognita Plus, which is perhaps not as accurate but is a good deal cheaper. However, if the dealer assured you that this software would scan and produce editable text, ask them to show you how it works, or get them to swap the software for some that will do what you need.
- Ian Yates


Category: Publishing and presentation
Issue: Oct 1996
Pages: 173

These Web pages are produced by Australian PC World © 1997 IDG Communications