I Wonder?

If such an app worked well enough (~98%+ on either mint or used) I might be willing to pay as much as $20 or even $30 for the smartphone app, but likely not more unless it was really "killer" (nearly 100% accurate, blazing fast, etc). Other collectors might only want to pay $5 or $10 or not at all. I have a hard time seeing where there's enough potential buyers of it to pay for development and licensing fees to Scott et. al. for use of their numbering systems. To sell such an app globally, it would also need to be able to find by Michel, Gibbons, Yvert, etc. A system using Scott would likely see it's market limited to primarily North America.

Now...if we really want to pipe dream, the ideal developers of such an app would be the catalog publishers themselves, and it not only would identify them for you, but pull up all relevant catalog information, provide links to similar stamps, etc. This may well in fact happen someday, but I just don't see it for quite awhile yet.

Most of the product ID tools out there are using publicly accessible databases for their search results. A barcode scanning app is a piece of cake, relatively speaking, because it just identifies the product's sku and then bouces that sku off Google. That barcode will always lead to an item number that is specific to that product and that product only. Even image recognition just bounces things off Google. Usable results in the stamp world will be much more mixed and contain a huge amount of "noise" in the search results if one is using Google.

What someone should really do is build a complete, comprehensive database of global stamp images in specific high resolution formats, identified by the numbers of all the major catalog publishers. The owners of that database could then license access to the images to any app developer, catalog publisher, dealer, or anyone who wanted to use them. Once that global image repository is complete, no one else will need to "reinvent the wheel" for a particular app as they could just point their search toward that database only. As a result, much of the e-tools we all seem to want will become easier to make reality. Maybe such an image repository already exists somewhere and I'm just not aware of it, but I think it would be very helpful in developing good electronic reference tools for stamps.

No doubt it would be a daunting task. I personally was thinking how they do those bugs you see all the time in papers and the such. You know the ones that you can scan and then it takes you to a website Granted it is in black and white. A few years back I knew a person who worked for the company symbol technologies. He was telling me that they could get a lot of information into one of those bugs or bulls eyes.

Though I have not looked at any software if I was looking I would probably also check out fingerprint software. I think it does not look at the whole image but points on that image. I think if something is out there that could work past the cancels it would be of this nature. Say on a stamp scan 80% of the points says its a certain stamp its a good bet that it would be. I don't think there is anything available yet that could get us to 100%. But if there is something that could get us close electronically a human could take the ID the rest of the way. As I said before we would probably still need to have a stamp expertized to its authenticity.

If such an app ever came out, I think it would be a wonderful tool that really would be a timesaver. I buy a lot of mixtures and kiloware and the like. I think it would be wonderful if I could just scan them in with my phone and have them instantly cataloged and valued. There are other similarly wonderful things that technology could enable us to do in this hobby, but we'll just, unfortunately, have to wait for or do without. Good thing this hobby teaches patience!

I started a open source SQL based website which could store stamp images, links, etc. I populated it with images of every US stamp and it was indexed by Scott numbers. It attracted little attention and support so rather than keep paying for the hosting I just shut it down.
It is a good idea but as I found out you have to have the backing of certain entities (like APS) to make something like that work or have very deep pockets. Grassroots efforts, for whatever reasons, are very difficult to get going in philately.
Don

51studebaker...i applaud you for trying my friend...that's what I was afraid I would end doing...closing it down for good...Still think Face Recognition software that can be adapted to stamp recognition is still a good idea...Robert