When we were playing around on the Clarifai website, we noticed how out of place some tags were (e.g. a picture of american football players playing the game did not return a tag of American Football!) So we decided to make a game based around the tags and how inaccurate they can be at times (sorry Clarifai!).
What it does
From a bank of pictures we have collected from the internet, we randomly pick one and it displays with the top 10 tags Clarifai can come up with. The aim of the game is then to pick the 3 highest rated tags that Clarifai comes up with for that image. 1 point is given for each right answer and the aim is to get the highest score you can across 5 pictures.
How I built it
Challenges I ran into
1 main challenge that was encountered was that we had originally planned for a random picture API to be used to select the images. However, it was not compatible with the Clarifai API, and hence we switched to the idea of using a bank of picture URLs we had collected and selecting from them randomly.
Accomplishments that I'm proud of
The fact that we have managed to get it working completely with no major faults, and that it actually resulted in a fun game to play!
What I learned
What's next for TagIT!
We are looking to develop it further as the concept is one that is really unique, and fun! Look out for us in the future!