Recently, I realize that my 2 yr-old son had figured out how to unlock my iPhone and taking random photos. Occasionally, he'd show the pictures to me, point at the phone and ask me what's in the photo.
That's when we decided to make a simple app that teaches young toddlers like my son new words with the photos they take!
We plan to launch Chinese (as well as other languages like Spanish, French, German, etc) version in the next 1-2 weeks so that my son can brush up his Mandarin too :)
To be frank, the accuracy of the image recognition (we're using Google Cloud Vision API) is still not that great (although it's already pretty mind-blowing to me given that this is probably still strictly sci-fi stuff a few yrs ago).
So we've also built a "treasure hunt" option where the kids can say "hunt" for "products" (eg. spoon, fork, etc) in the kitchen (sorry can't resist the PH pun). This helps to limit the scope a fair bit so that we can reduce the "false positive" and "false negative" results somewhat.
We've also built in a child-safe feature such that MoMo won't display any word suggestion if we detect any potential adult/violent content in the photo.
Finally, this is built on React Native. We're still having some performance issues but hope to start optimizing things in future versions.
Ask us anything!
Wicked idea! Computer vision matched with kids should work quite smooth, since the two are around the mental same age. Any particular reason why you went with ReactNative?
@tudorizer LOL u'r right the CV and kids have similar mental age (and it shows in their recognition rate, though I suspect CV is gonna improve leaps and bounds in near-term). In any case, we thought it's a simple enough project to try out ReactNative (our first time messing with it), with the obvious benefit of publishing to both iOS and Android.
Product Hunt
MoMo
MoMo
MoMo