How to import 80K scans in Access

February 24, 2010 at 02:51:58
Specs: Windows XP
Hi everyone,

I have a project to scan 80.000 cards in to pc and put them in a access database, so people can search for the cards on the easy way and it safes us time to do another things than looking for a card.

The software I use is Omnipage 17(those idiots wont support, because i pay for it and not steal it). I was looking around and found out that I can use .CSV files for that.

But it seems it has a lot of work with correcting the spelling of the scans that OmniPage makes.

So does anyone have experience with OmniPage or another scannings(OCR) program?

I also try to make a batcfile with help of Mechanix2Go (
But find out that is not usefull enough for the task.

Thank you in advance for you answere.



See More: How to import 80K scans in Access

Report •

February 24, 2010 at 04:46:26

When running OmniPage, did you set up a custom workflow or did you just run 1-2-3.

With a custom workflow you can include InteliTrain, which allows OmniPage to improve its OCR ability. This will significantly improve the OCR.

Use a workflow that includes manual correction with IntelliTrain, do the corrections until recognition is pretty much 100% correct, then you can stop the manual review and let OmniPage OCR the remaining cards without manual intervention.

If you are getting a large number of incorrect words, you should also look at the scanner settings. Try one card with different scanner settings and see which comes out the best. I seem to recall that Greyscale at 200 to 300 dpi gave good results while black and white was terrible on some documents and higher dpi did not improve recognition.

Another approach is to scan the cards as images and save each one with a logical name and just use Excel as a means to return a hyperlink to the card's image based on criteria you enter in Excel.


Report •

February 25, 2010 at 00:12:06
Hi Humar,

Thank you for your reaction.

I try it out, it was better with setting it to 400 dpi and also I used the workflow tool.

The only thing what I did not find is, is wich output is the best for import in Acces database. I used .CSV, but it doesnt work well.

Next thing: With workflow I cant use self made zones, If I use workflow it uses the whole paper instead the places I need.

Do you have some more Ideas for me.

Thank you.

Report •

February 25, 2010 at 04:32:10

When you create your own Workflow, after 'Scan Images' include a 'Zone Images' step. This gives you the option to choose a template that you have created, and to apply it, (with or without the zones being displayed for further manual adjustment).

I do not know what type of information is on your cards, but if it is all plain text, I would tend to avoid CSV and save as A Word RTF file. (If you put a blank card in every 1000 cards you can use the Save option to create a new file at every blank page, so ending up with 80, more manageable files.

Then use Word to do the remaining spelling corrections. I would then import into Excel and depending on what is in your cards, possibly use the 'Text to Columns' function to split the files up at spaces, tabs or at a specified character, and finally import into Access.


Report •

Related Solutions

Ask Question