Debbie,

ABBYY is a good starting point for several reasons. First is they are the primary engine in most data capture products anyway. Second they do have both OCR and ICR ( handprint ), and very good on degraded documents. However...

Before looking at ANY products you have work to do. When looking at capture software there are many things to consider that are actually more important than the product. Such as your sample set, your expectations etc., how will you deal with verification and exceptions. I would focus on these first. If you don't know these FIRST your project is likely to fail no matter what of the vendor products you choose. I see it all the time, companies swapping one product for another just becuase of poor planning. Check out my blog blog.livinganalytics.com, and the AIIM articles on capture, as well as the AIIM OCR Toolkit for a lot of good ideas ( i am biased as I'm the author of a lot of it ). Here are a few other vendors to consider.

Kofax - ABBYY, Nuance, ODT Engine
Kodak - Nuance Engine
OpenText Document Technologies
Artsyl - ODT Engine
DataCap - ODT and ABBYY Engine
BrainWare - ABBYY Engine

Debbie, it has been a while since I was a product manager for a document management product that was designed to process mortgages. At that time, Fanny used a product that was called Desktop Underwriter (DU) to capture data that was used to complete a 1003. If the applications are you are receiving have been submitted using DU the metadata you are capturing from images should already be in database at Fanny. If this is the case the data you are looking to extract from the 1003 may already be in another database at Fanny, you should be able to capture a single key field from the 1003 and use the mapping capabilities of a data capture product to extract this data from the other database.

The other option is to replace the PDF with an eForm that will automatically send the data to your risk analysis system. We do similar forms to yours such state and federal IRS forms in both English and Spanish. The best way to solve this problem is going to be to capture as much of this data as you can with electronic forms, then what you can't- use a product like Abbyy for the paper forms.

You can check out http://www.lincware.com/oc-mobile.html for a video and overview of this technology.

Hello Debbie,

The spanish and the types of forms are no problem for ABBYY Flexicapture products. The only challenge will be the hand writing on the form. I believe that most (all) products have a very hard time reading hand writing on a form. If the form is typed then it is really relatively simple and degrees of accuracy can be made by tweaking the Flexicapture product. For example, if you expect a zip code then you can modify the product to look for numeric characters with a length of 5. ABBYY in my opinion has the strongest data capture engine on the market. These kind of problems are exactly what I deal with daily. If you have any questions to my comment feel free to email me bzindler@ufcinc.com.