Inspiration
All of the group members for this project got together because all of us wanted to learn something about augmented reality. Interests also spread into image recognition with Google's Cloud Vision API, and so we found a way to implement that as well in a way that would be interesting.
What it does
In it's current state, Volhacks Query is capable of using Google Cloud Vision API to make a guess on some of the aspects of an object on camera, such as if it is a chair, or a person, laptop, or general electronic device, to name a few examples. If it also recognizes a company logo, it will display the name of that company as well. The project is also currently capable of basic augmented reality. When it sees an image that it recognizes, it will display a 3D model on top of it. Right now, it only recognizes two images, for which it displays either a sphere or a cube depending on the image.
What we really want it to do
We wanted Volhacks Query to use Google Cloud Vision API to make a guess on some of the aspects of an object, such as a chair or a person or a laptop. Then, based on what Google thinks it is, Volhacks Query will prepare a 3D object for that generic aspect of the object to be displayed in augmented reality.
Plans were to also have Volhacks Query to perform a search based on what the object was thought to be, or perhaps to find the price of the object, or a similar one online.
How we built it
The program was created in Unity, which by default uses C# as the primary programming language. Image recognition was implemented with Google's Cloud Vision API. Augmented reality was implemented with Vuforia, a built-in augmented reality library for Unity.
Challenges we ran into
One problem that was faced was implementing the Cloud Vision API. This was the first time that any of the group members had done work with API's, and so implementation was somewhat challenging at first.
The greatest challenge (so far) for us was getting Image Recognition and Augmented reality to work at the same time. With the way that they were both initially implemented, only one of them could have access to the physical camera. To fix this, the way that the Google Cloud Vision API was implemented was altered.
Accomplishments that we're proud of
We're proud of the whole thing. We faced many problems with this project, and every group member worked his hardest.
What we learned
We learned about using Unity as a development platform, as well as how to implement API's, Image Recognition, and Augmented Reality.
What's next for Volhacks Query
We don't know.
Log in or sign up for Devpost to join the conversation.