OCR and VFP

[Scott24x7]

So proceeding onward in my quest for ease of data input...
Has anyone implemented OCR with integrated scanner? I was reading the EZTWAIN and Transom sites, and while EZTWAIN mentions easy integration with Transom, none of the materials mention integrating with VFP (VB, C++, Delphi, few others) but not VFP.

Anyone ever try this? Or is there an alternative product for OCR with VFP that would be recommended (yes, I know I can Google OCR VFP, but that doesn't tell me if it's any good... hoping to save some time on not using the trial and error method).
Thanks.


Best Regards,
Scott
ATS, CDCE, CTIA, CTDC

"Everything should be made as simple as possible, and no simpler."

 

[Mike Lewis]

Try googling OCR Activex Control. I'll bet you see a few likely candidates. Chances are that none of them will explicitly mention VFP, but that's no reason to rule them out.

Mike

__________________________________
Mike Lewis (Edinburgh, Scotland)

Visual FoxPro articles, tips and downloads

 

[Scott24x7]

Thanks Mike.
Yeah, I went ahead and downloaded Transom just to see if I could get it to work as well, since the .DLL calls are all there for EZTWAIN, which has Transom integration.
Apparently there is 32bit and 64bit of Transom, and older versions of Delphi only work with 32bit, so my expectation as well is, best bet, try the 32bit of Transom and see if there is enough in the VB examples (which I don't really know either) to make it work in VFP... if it does, maybe I'll write an FAQ on it.


Best Regards,
Scott
ATS, CDCE, CTIA, CTDC

"Everything should be made as simple as possible, and no simpler."

 

[Mike Lewis]

Yes, let us know how you get on with it, Scott.

Another one I noticed is SimpleOCR. I don't know anything about it, except that it seems to be free, which suggests it is very basic (or maybe not). The same company has more sophisticated paid-for products.

Mike

__________________________________
Mike Lewis (Edinburgh, Scotland)

Visual FoxPro articles, tips and downloads

 

[Mike Lewis]

Presumably, the OCR does not have to be integrated with the scanner. The scanner component will let you save an image in a file, probably as a TIFF or JPG. Any OCR component should then be able to read that file.

For that matter, the OCR doesn't need to be an ActiveX control, a DLL or an SDK. A free-standing utility should do the job, provided it has a command-line interface.

Just a thought.

Mike

__________________________________
Mike Lewis (Edinburgh, Scotland)

Visual FoxPro articles, tips and downloads

 

[Scott24x7]

Yeah, it's very sketchy actually...
Having some difficulty coming up with anything other than "whole page text" scanning.
The tricky part with OCR and a business card is figuring out what components belong to what fields you want to map them to. Image bits get in the way, and it's a very messy process from what I can expect.
I think I'm going to drop EZTWAIN another note to their support desk and ask if they have any examples using VFP, since they flog so hard the Transym product. I've spent several hours now not making great progress in that area.


Best Regards,
Scott
ATS, CDCE, CTIA, CTDC

"Everything should be made as simple as possible, and no simpler."

 

[Mike Lewis]

I just did a quick Google search for dedicated business card scanners / OCR tools. Scott, I know this is probably not what you want, but I was curious to see what's available. The only products it turned up were smartphone apps that claim to scan a business card and intelligently extract the various fields and store them in a file or database.

One example is WorldCard Mobile. Among other formats, it can store the card details in a CSV file (on the mobile device). So, it occurred to me that, at a pinch, you could keep a smartphone permanently tethered to a PC (or connected to it wirelessly); use the phone to scan the card; then, in your VFP app, pick up the CSV file from the phone's storage and import it into your database.

It wouldn't be as good as a fully-integrated dedicated card reader. But I thought I'd mention it anyway, just in case it's of interest to anyone else.

Mike

__________________________________
Mike Lewis (Edinburgh, Scotland)

Visual FoxPro articles, tips and downloads

 

[Scott24x7]

Mike,
Yeah, that notion actually crossed my mind when I first started this. To be fair, getting the front/back image scanned was a real triumph, and the process with the Canon P-208II which duplex scans, and has an "adjustable" feeder which makes it very easy to feed cards through it is just brilliant. In fact, you can even stack about 10 cards on it and it will do them all at once... now, if I had OCR, I could check to see if their company already existed, if not, add the company, then add them as a contact. I know there are commercial software out there that do this for their "own" systems, and mobiles, as you mention. It's actually where I got the idea to integrate it into our system, mostly because I want an interface (including card scanning) that makes doing the entry easier than not doing the entry.
I'm sure some will criticize me for not buying some package that does this already, and even more foolish for starting modern enterprise system development in VFP 9... the "Dead Language". Well, I happen to have 20+ years prior experience in this dead language (going back to Fox for Dos FoxBase+, and even further back to dBase III and Paradox before it went pear shaped. I just haven't used it so much in the last 10 years. That said, my "learn curve" has been exceedingly short, and I have the framework I put together years ago. In some ways, I'm stronger than I used to be, as it's caused me to look at "old stuff" that I no longer understood, and simplify it.
The business I started only 4 months ago has generated near half a million $USD with a staff of 3. (Not bad for a startup that began with $40.) What we do is pure service, so we have no product overhead, save for a few licensing costs for some of our engineering system software, but we pass hat on to the clients too.
So, what I need is a system that doesn't exist. What we are doing (and I've been working with consultancies for more than 15 years in this field alone), there is no "software" for. There are even few drawing the benefit of knowledge management systems, which would put them a long way down the track. A few try to "kludge" SharePoint into some system, but it's disconnected from too many other things. The IP we are building up as a research company, (one branch of what we do) doesn't exist in any other organization that I've seen (and I've worked for Dell, Uptime Institute, global consultancies...) what we are creating is unique.
So all that said, it may seem crazy to try to replicate some functionality in VFP for our "Enterprise Application" (and by the way, SalesForce, SAP, ERP, CRMs) none of them do what we are doing (though they could be "made to do so", the cost would be significant, we'd have no control over it, and at the end of the day, I've been able to build a large base of our needs in only a few days. Largely thanks as well to the Tek-Tips community, and leveraging "what we know" which in reality is probably what we do best. So for an "internal enterprise IP management system", VFP is fine. I always laugh when I tell people about what I'm building and their first question is "What language are you using"? (99% of these people have never written a line of code in their life, and the ones that have, have never gotten much beyond report writers.) So my reply is always "English". Though I recently started saying .NET, just because I don't feel like explaining myself. Those who don't program are still surprisingly aware that VFP was canceled by MS in 2007.
The OCR thing is monumental, in the task of parsing even a few small bits of data from front/back of card. You have complexities like some cards (that I have) have 2 addresses on them... Cards in English, Chinese, Spanish, Japanese, Bahasa Indonesia... all going to be a significant challenge, and I expect the "OCR" challenge could even take me a year to tackle, but I will tackle it sooner or later.
For the moment, because there wasn't a reasonable example for me to work from, I've decided to focus on completing some of the other areas. With the front/back card scanning accomplished now, I feel that is more than 1/2 the battle in most cases. And in fairness, it's not really a bad thing for people to enter that data... I'm making getting large amounts of other data into the system exceedingly easy and automated. (450+ data points imported from Excel spreadsheets, which are populated in many cases from exports from the system as well). Templates for "data collection" are established in the database using "code tables" and then thrown at an Excel file write. The users complete the excel sheets, which are tightly controlled for entry, and then it's all sucked back into the system. Decoupling the data collection in this way means we don't need users to have access to our proprietary IP system. And we can give them outputs to Office files (like Word, Excel, PowerPoint) generated from VFP instead of PDFs, as they want to use/edit them, but we can do that by giving them the "raw numbers" but without the IP embedded in the docs. (None of our competitors do this...)
If we become ridiculously successful and a $100million dollar company, then we'll use this as the prototype, and hire someone else to write it, but as a startup... this is awesome. Leveraging the skill set we already have... priceless.
I guess I make this point because I don't want to kludge a mobile phone and mucking with the get data out, get data in... I think it would wind up being more type than just typing a few bits of data off a business card into the system until we can integrate OCR. But I like your out of the box thinking... that's the kind of stuff that leads to real innovation. And I think our bold choice of VFP at this late stage of it's life is actually an innovation in itself.


Best Regards,
Scott
ATS, CDCE, CTIA, CTDC

"Everything should be made as simple as possible, and no simpler."