What do you think of Kofax' "Point of Origination" strategy ?

Scanner vendors did attempt to make personal scanners more accessible. The real case demands a radical approach: the use of ubiquitous smart-phone or tablet computers as a means to capture business-ready documents from paper.


Opened by Dimitri Schmitz, Solution Manager, bpost
May 11, 2012.

recommanded this debate

Participate in the debate




Antoine Fournier Head of ECM, Input and Output management, Zurich Insurance
May 13, 2012

recommanded this answer

I think this strategy in already mainstream for all SW vendors in this area.
Mobile technology is actually the engine of a strong "come back" of our good old client/server architecture.

This is obvious that, as soon as users have a scanner in their pocket, they will be able to be the origin of capture (or inbound) processes.
Is this anyway enough? Can we do more?

Bringing unmaterialization of inbound documents at client side is, in my opinion, half of the path that we, software vendor, have to do.

Categorization and data extraction can and must be done at the same level. Some of the business processes that involve capture need a data validation step and an identification of the document type. These steps also may be done at the "Point of origination".

You can see what was built with ISIS Papyrus for Raiffeisen online banking:

Comment on this response




Riccardo Rosalli CEO, ICADO S.r.l
Aug 20, 2012

recommanded this answer

In principal the strategy is fascinating and may extend the flexibility of any organization managing inbound business processes.

Nevertheless from a pure technical point of view I see some weakness using a tablet or a smart-phone built-in camera as "point of Origination" vs. MFP or flatbed scanners which are obviously not so portable.
Well-Known devices are currently offering the following image resolution:
- tablet:2048-by-1536-pixel
- smart-phone: 3264-by-2448-pixel
Which end respectively to 175-by-185 dpi and 279-by-296 dpi on a A4 paper size.The tablet's resolution is definitely poor for real mission critical A4 document text recognition and the smart-phone's one it's just sufficient.
In addition there are other distortion factors like not uniform flash illumination / handy text paper focusing / lens angle distortion, etc.

A well-designed document together with its capture application can reduce the above distortions but from my experience you cannot expect high level OCR accuracy (99%) in such scenario. For general purpose barcode recognition or photo and video acquisition the resolution should be enough.

So then if your document is quite small <=A5 and its size fit the minimal focal length you can expect decent recognition percentage otherwise it's better to enhance your "Point of Origination" which is the first step to assure a good recognition quality.
Comment on this response

Antoine Fournier 67 Antoine Fournier Head of ECM, Input and Output management, Zurich Insurance

Aug 24, 2012
Don't you think this actual limitations will be shortly covered by technology ?
I believe a suitable redesign of documents to scan, together with increased resolution of tablets and smart phone cameras, will most probably allow consumer doing capture step themsleves, no?