Credit to @JanelleCShane for the idea of putting the original image through Cloud Vision, and @FlyingOctopus0 in the replies for noting that the results vary on rotation!https://twitter.com/JanelleCShane/status/1103420287519866880 …
You can add location information to your Tweets, such as your city or precise location, from the web and via third-party applications. You always have the option to delete your Tweet location history. Learn more
Credit to @JanelleCShane for the idea of putting the original image through Cloud Vision, and @FlyingOctopus0 in the replies for noting that the results vary on rotation!https://twitter.com/JanelleCShane/status/1103420287519866880 …
How biased. Google Cloud vision's training set clearly has more images of ducks facing down than ducks facing up!
Avoids a duck up whenever it can
You should try this on ambigrams. Here the correct answer should be philosophy.pic.twitter.com/yJiO9xOLoD
Looks like google cloud needs some google glasses
Isn’t that ... correct?
A model could probably generalize, but my perception certainly isn't rotation invariant. At certain orientations I'm much more likely to see a duck and at others a rabbit. Our "training data" is hugely biased to faces at certain orientations, so maybe not surprising.
Twitter may be over capacity or experiencing a momentary hiccup. Try again or visit Twitter Status for more information.