EcoScan

main page
scan result page

Inspiration

In our campus, we have many recycle bins, some of them come with three (green one, blue one and black one). We got confused what recycle bin we should place our unused item in. Some recycle bin has photo descriptions on it to describe certain types of items, but most it is time-consuming to identify it. That is the work that AI tools can help us. ![alt text] https://photos.google.com/album/AF1QipNbl9grr3LziUgsYs_wtuCuFxwpJpr8-9T44OLo

What it does

The app allows user to take a photo of a ready to toss item, sort it to four categories: compost (green recycle bin), recycle (blue recycle bin), landfill (black recycle bin) and other (cannot be recycled).

How we built it

We develop a web that can open front camera(default) and back camera that can snap photo.
We use node.js to develop the server that can call FireLLaVA-13B API and give response.
We integrate the front-end website, back-end server and API fetch.
We design the UI of the frond-end website.

Challenges we ran into

First time Hackathon experience
Some emergency happens, so two of your original teammates cannot attend. So, our group has only two members.
First time we work with AI API, so it takes some time to figure it out.

Accomplishments that we're proud of

Finish the main part of our project (from idea to prototype to optimism) within 24 hours!
First time to use AI API!
Proud of our UI design!

What we learned

Don't give up.
Ask for help and suggestions.
Teamwork.

What's next for EcoScan

Train a more specialized model to recognize and sort multiple items in one photo and increase accuracy.
Develop a scanner, build on the recycle bin so that user can scan the item and then put it correct recycle bin.
Deploy the app, and test in mobile devices.