Dumpster Duck

• 

  The pop-up UI

• 

  Added "Upload Image" button

Inspiration

At a young age, we are taught to reduce, reuse, and to recycle. We wanted to make a product that would provide a nice quality of life change that can encourage others to throw away their trash properly. We also love ducks.

What it does

We allow users to upload pictures of waste, and using AI, we used object detection to classify the waste in order to provide information on how to dispose of it safely!

How we built it

Instead of making a whole new website which costs energy to host, we reused TMU's existing Waste Wizard and added more functionality. Using javascript, CSS, and html, we created a chrome extension that sent API calls to Google Gemini and Google Vision.

Challenges we ran into

This hackathon was a first for us, so we never made a chrome extension or used Google's AI systems before. We at first used Node.js but ran into many problems with our modules. We managed to create a working prototype by sending API calls!

Accomplishments that we're proud of

We finished and managed to build an actual working prototype. Our solution is simple but is actually quite handy and we believe it is very scalable and accessible.

What we learned

How to make API calls to Google Vision and Gemini, what's needed to make a chrome extension, and that ducks are the best.

What's next for Dumpster Duck

We currently use a pre-trained model due to time restrictions, but we would love to expand and learn how to train our own models like in Tensorflow. By using react, we can also speed up the loading time. It was very simple to add the button to TMU's website, maybe they can add it in the future so people on mobile have easier access to it!

Built With

• css3
• google-cloud
• google-gemini
• google-vision
• html5
• javascript