C-AR | Devpost

Inspiration

We wanted to create an application that would allow users to seamlessly connect with regional dealership's inventory databases.

What it does

Using 360.Agency's extensive database, we built an application that utilizes an Image Recognition API from SightHound to identify the make and model of any vehicle. Using this information, C-AR queries the 360.Agency database for a list of cars that are available for purchase from nearby dealerships. Our application then displays the results and helps customers make trim modifications using a built-in Augmented Reality experience. We can store every customer interaction as a lead in 360.Agency's DB.

How we built it

The Image Recognition component, along with the core application structure was build using Python. We wrote a number of dynamic SQL queries to display the correct information given the photo taken. We used Unity 3D and Vuforia to build Augmented Reality models of vehicles. We created 4 unique 3D models, with the stated intent to build out a complete library for all cars. We used Xcode to help build out the application and render it compatible with our iOS devices.

Challenges we ran into

In developing the AR, we found it difficult to find the correct plane on which to project as well as properly rotate the 3D objects. To fix this, we changed the plane to a specific image target, upon which our image projection is more stable. Down the road, we expect to be able to project onto any flat surface, such as a street or floor.

For the Python script, we found it challenging to integrate the results of the image recognition with the 360.Agency Database. There were issues connecting and interpreting, which we solved by iteratively trying different approaches.