I envision a world where people are traveling around and find objects of interest that they're interested in purchasing, but do not know of themselves.
What it does
QuikPricer quickly identifies unique objects and prices them accordingly, storing your most recent QuikItems in your shopping cart. If inclined, you may visit your shopping cart and QuikPricer conveniently directs you to a URL to purchase your item of interest.
How I built it
Upon launch, users encounter an ARSCNView. They have the option of clicking on a button which calls a function to capture a screenshot of the ARSCNView. The ML for image classification uses CreateML to identify what an object is based off a pre-trained dataset that I trained myself (uses a CNN). These snapshots are loaded into another view, the PopUpViewController, which animates in the foreground on top of the main view controller when users click on the Shop icon. A CollectionView is embedded in this PopUpViewController and users can scroll through and see what their most recent item IDs were. They can click on an item in the collection view and they will be prompted by an actionSheet which asks if they wish to purchase said item. If they say they want to purchase the item, they will be directed to an affiliate link (if applicable) to purchase said item. Their most recent item IDs are persisted in the app using UserDefaults for local persistence.
Challenges I ran into
Setting up collection views. Learning ARKit.
Accomplishments that I'm proud of
Learned how to properly transmit data between views seamlessly with persistence.
What I learned
More iOS tips and tricks for customizing Collection Views and ARKit in general.