Ok, sounds like most of you guys would prefer a pay per use approach then.
It will recognize letters on most backgrounds, except for really cluttered ones.
About recognizing different effects (like overlay and others) unfortunately I think that is something it wont be able to do it.
The challenge we are facing is recognizing letters printed on paper by normal inkjet printers, because they are extremely blurred when captured . But we are also trying to find a solution for thisâŚ
Sounds like something that could be really valuable, given how many designers use What The Font.
Question: Iâm assuming you donât have access to 40k fonts directly, so how do you recognize the fonts? Wouldnât you need some sort of database that actually contains the fonts?
Also, given the fact that What The Font is free, I think itâs likely that people will try that first, and if it doesnât work, turn to your application, in which case you could probably easily charge ninety cents or so for a successful recognition. On the other hand, if itâs just 10 cents, people may skip What The Font altogether.
It would be nice if this application could remain independent of Monotype, by the way!
@alsg - Sounds like a neat idea, but not the best application for neural networks if you looking for monetization. I think it will be very hard to get any traction, though the costs of running the servers is not that high (I know some guys who do it as a business - neural network applications).
All in all, people will welcome the tool, but donât expect to make any money or even make ends meet.
The important thing in this kind of ventures is not to expect too much. I feel he was concentrating on the wrong stuff (e.g. the stuff he cannot control versus the actual product)
Just putting my 2 cents in, regarding the business model I would suggest selling credits (rather than singular micro-transactions, which are hard to track for the purchaser and problematic to bill in general).
Something like $5 would work (equating to 100 searches, at 5¢ per search). That way you also get more $ upfront to fund development too. Also searches that donât result in any hits should not cost any credits.
2- Take a screenshot of a letter.
IMPORTANT: Has to be an âAâ capital letter (I have a database of just 4k âAâ on the system to keep server costs down.
3- Click on âchoose fileâ and locate your screenshot
4- Click on "search a local file"
5- System gives the best match. Magic! ; )
Please notes that as the system:
Only works with A (capital letter). To keep server costs downâŚ
Currently has 4K fonts on the database. So it will locate the font that best matches. I have more than 40k fonts, but would be super costly to host all glyphs.
Would love to know what you guys think! feedbacks (good or bad) are highly appreciated!
talks for the feedback. But as mentioned the system is currently working with 5K fonts. So it will give the best match with the current database we have (5k).
Could you please post a screenshot with the results?
As I said, it was way too early to think about pricing and business side of things. Maybe you can do a kickstarter campaign to see if you can finance your server costs.
I am very interested in your work. I tried to test it out but it was loading forever. Is there any chance that you can get it working properly again? We see a strong business potential for your work.