The subject is a bit misleading: this is a new computer vision library. There is a separate augmented reality library which lives on the same website, but the two don't appear to be connected.
FastCV has the potential to be interesting, but the website is pretty off-putting at the moment. They need to clarify the license, as others have mentioned. It also appears that the SDK might be Windows-only, judging by their installation guide (I didn't want to register, just to find out that was the case). Having to register just to see the release notes is pretty bad too. Hopefully these are just teething problems and will be sorted out in due course.
It's great to see new computer vision resources like this appearing. Although I'm not willing to register for the download at the moment, I have been browsing the API docs a bit and would be very interested to see a comparison between this and OpenCV from someone who's used both.
Skimming through the API docs, it seems to be a very small subset of OpenCV at the moment.
However, a lot of vision is how you handle the details (both for speed and for accuracy/performance). Without using it extensively, it's hard to say if this is better.
OpenCV has at least a long history and being battle-tested going for it (despite its warts).
Qualcomm's augmented reality project, from the same site, has license information available here: https://ar.qualcomm.at/legal/license/ I'd guess FastCV will have a similar license.
That license for Android seems to suggest it includes at least the GPL. That means commercial apps built with this library must make the source available, right?
Where do you get that from? The only reference to the GPL that I noticed was in section 2.2, which expressly forbids you from doing anything which could make the code become subject to the GPL.
I hate how this (and to some extend OpenCV) expect you to understand computer vision to use the library.
I sympathize with your frustration, but being in Vision myself, I have an idea why this is the case: vision only barely works!
What this means is that most things you would want to do are only barely possible, if at all, and it is highly application- and input-dependent. To get it to work the way you want, you have to understand some vision, so that you can tweak things the right way.
I think once vision matures more and things become more "off-the-shelf", you'll find more libraries aimed at non-experts.
FastCV has the potential to be interesting, but the website is pretty off-putting at the moment. They need to clarify the license, as others have mentioned. It also appears that the SDK might be Windows-only, judging by their installation guide (I didn't want to register, just to find out that was the case). Having to register just to see the release notes is pretty bad too. Hopefully these are just teething problems and will be sorted out in due course.
It's great to see new computer vision resources like this appearing. Although I'm not willing to register for the download at the moment, I have been browsing the API docs a bit and would be very interested to see a comparison between this and OpenCV from someone who's used both.