Developed by Caryn Tran
This is a project done at UC Berkeley under the supervision of the Berkeley Institute of Data Science. This is an android app demo created to explore the usage of image recognition and AR in the product displays.
The app was created using Vulforia technology.
I spent 3-4 hours just doing the initial set up. Setting up Android Studio, Android SDK/NDK, Unity, and Vulforia. I spent a lot of time looking at tutorials and looking through documentation to understand Vulforia. The easiest way to work with Vulforia is through Unity in my opinion. But there are provided sample apps that cover nearly all of the functionality one would need for an AR app. But when loading the sample apps, I had to spend another hour or so just debugging my set up. In the end I got the sample app to work on my phone which I won't push to gitHub since I didn't write any of the code there. I also followed the tutorial through Unity, but that's not really Android specific and generates an APK.
Ideally, I would like to make modifications to the sample app and branch off from it. I need to look more at the samples and their use of the Java API. I don't understand the code well enough.
- VuMark generation process is not trivial (requires Adobe Illustrator)
- Haven't delved into Unity enough to be comfortable, Unity has limitations for Native Android
- I don't know any C++
Picked apart the demo app. Two versions. One uses Image Target and an imported device database. The other is the Cloud one, I haven't gotten that working yet.
Change the 3D-image according to the tracker. Create OpenGL text mesh to overlay.
- Lots of reading code. Not really any good guides on specific things.
###Past Week Got api calls to work but not correct calls.
Change the 3D-image according to the tracker.
- Need to refactor code (remove unnecessary code)
- Get product id
- Fix image loads
- Polished the UI
- Finished MVP and assisted with iPhone development