OCR Software / Data Extraction

[Berics Accounting]

Hi All

Im after some scanning software that will scan multiple documents and extract key (pre-defined) pieces of data to use as file names.

For example, if i were to scan a bank statement it would recognise it as a bank statement and store it with perhaps the statement date, or if i were to scan an invoice, it would recognise it as an invoice and save it using the invoice number.

I know this is completely possible as I have used it in a previous job, but havent got the faintest idea how to find it.

Any suggestions?

Thanks

 

[janib4all]

Don’t have clue however would be interested in same if someone suggests something; I get few pending tasks related with OCR and data entry and I don’t like manual work.

 

[MrTomFarrow]

cVision do something like that. It's likely to be pricey, but you should reach out to them.

Take a browse around the net, looking for "intelligent document recognition", I'm sure that there are more packages.

Let me know how it goes. Interesting topic..

 

[ffox]

I would take care with this.
1. If the software misnames a document filename - how would you find it?
2. OCR scanned documents are not admissible in many legal and contractual arenas (non OCR PDF and TIFF are).

A recent Delphi survey showed that we spend 20% of each working day searching for documents in our own file systems. That can be vastly reduced by utilisation of a good EDMS (electronic document management system). Office 356 is a good example. The meta data (including the file name and doc type - invoice, statement etc.) is fundamental to the index/find process. I'd look for a means of improving document location rather than reducing scanning time and effort.

 

[MrTomFarrow]

I concur, don't just OCR a file and hope for the best.

Keep two filesystems, (of course, keep backups of both, I can advise on backup solutions if needed). One containing original images, and another with OCR documents.

Organize things into directories. My file system looks like this (of course, there's more, but this is an example)

Originals
    HMRC
      Corporation Tax
            Jan
            Feb
            Mar
            ...
      Web Access
            Jan
            Feb
            Mar
            ...
    HSBC
      Online Banking
            Jan
            Feb
            Mar
            ...
      Agreements
            Jan
            Feb
            Mar
            ...
      Correspondence
            Jan
            Feb
            Mar
            ...

Of course, I have an OCR clone of that.

My document traffic is relatively low, so I just manually OCR everything, META keyword it and file it.

 

[Karen Bennett]

Hi Jon,

OCRex has developed an OCR data extraction software solution that is specifically designed to extract transaction data from bank, credit card and online statements. You simply scan the hard copy of the statement, upload the file to AutoRec and it is converted into a spreadsheet for you. You can also upload digital files. So basically AutoRec save you time compared to manually typing the data into excel. You can then export to xml, excel or csv. We are have exports for a number of accounting and bookkeeping software packages such as Sage, IRIS, VT, CCH, Digita, Xero and much more.

What is unique about AutoRec is that we have developed an OCR template for each bank statement mapping out how the statement is formatted. This allows AutoRec then to recognise where exactly the date is located and where the debit and credit is located for example, etc.

Regarding invoices, receipts, PO’s, packing slips, etc, OCRex is launching a data extraction software solution that extracts data from business documents also. The key data is extracted including the name of the supplier, the date, the invoice number, the currency, tax, total value, PO number and more! It is customisable so to extract any relevant data that you want or don’t want to have extracted.

What’s different about DocuRec is that it uses OCR technology plus the latest artificial intelligence technology. The software has a self learning feature so that when you process a document, it will remember the information you require from each document type. Therefore, it can then extract just the information you want from given supplier invoices/receipts etc automatically going forward.

DocuRec is currently in beta phase at the moment but will be available very soon.

Thanks, Karen