I’m working on an App that recognizes fonts, using Artificial Intelligence (Deep Neural Networks).
The app already recognizes 40k fonts with extreme accuracy, and our library is growing every day.
A typical use case would be…
You are walking on the streets and you found this amazing bauhaus poster. You take a shot with your phone, the app automatically recognizes the Font Family, Subfamily, Author and other useful info. The search is saved on your history. Magic! : )
What you guys think? Would it be useful on your stack?
WhatTheFont is pretty old now, so I’m sure you’re right about the tech.
Have you seen this recent talk from TYPO Labs 17? Sounds like they are investigating similar ideas at Monotype, although aimed at helping customers find fonts using various criteria.
It will recognize fonts from all possible foundries we can find. Currently we have around 900 foundries mapped, about 40k fonts and we are daily adding others to the system. Foundries will not pay for the app, only end users will.
about business model, how about something like 5 or 10 cents by recognition? No charge for duplicate font recognition nor for “can’t find the font” result.
Mark, I just watched the video, and their method is for categorizing and finding similar fonts. But they don’t actually discover what font it is. Anyway, brilliant work! ; )
Wow, sounds amazing. I would love to have such an app.
Would it only recognize black letters on white background? (That’s what WhatTheFont actually is able to do, as long as the fonts are in their catalogue anyway). It would get really interesting if your app had the ability to recognise, say red letters with a color border and whatever effect the designer had put on it.
As of the business model… subscription is really uncool in my opinion, everybody hates that model, except people who might use it many times every day for professional purposes.
Like Jan said, 5-10 cents per recognition sounds pretty reasonable to me. That’s something I could stomach.
I think this app would be awesome for designer who need it and at the same time be something quite niche in regard to the general market… So my guess would be a subscription model would further reduce the user base. But if you get it out to every designer as a handy and cheap everyday tool this might become a success. Many times a few cents is always better than a few times many bucks imho.