We wanted to create an intuitive and direct method of translating object names.
What it does
After starting up the app, you simply need to point the camera at the object you want to translate and tap the screen. The name of the object (along with its translated name) will then show up above the object in real-time. You do not need to know the name of the object to translate it.
How we built it
We used a Tensorflow (with Python) machine learning algorithm to identify objects in the camera and the Google Cloud platform to translate the names. The names are then displayed above the object in AR using ARKit. The app was built in XCode using Swift.
Challenges we ran into
Our XCode version was out of date and we had to wait three hours for it to update. We also had to install a Mac software update that took over two hours to complete. We also only had one Mac that the team members were willing to use.
Accomplishments that we're proud of
Our XCode is now up to date.
What we learned
Make sure your software is up to date.
What's next for TranslateAR
More language compatibility and increases in accuracy (maybe pronunciation of translation).