Recyclables, electronics, and food scraps are mixed with common trash, as a result, it is difficult to recycle and estimate carbon emissions.
Proper waste sorting can significantly reduce environmental impact but many people don’t know how to sort their trash correctly.
That’s why we created SmartSort, an AI-powered tool that helps individuals and companies to properly dispose of waste while tracking its carbon footprint.
SmartSort is a smart guide to sorting trash!
With just the snap of a photo, users can:
- Instantly identify what type of waste it is (recyclable, compost, landfill, etc.)
- Learn where to properly dispose of it
- See the estimated carbon emissions of each item
- Compete on a leaderboard that gamifies sustainable habits
- Connecting the Flask backend with the React frontend
- Training and optimizing the PyTorch model to accurately classify trash images and calculate CO2 emissions
- Frontend: React + Tailwind CSS
- Backend: Flask (Python)
- Database & Authentication: Firebase
- AI Model: PyTorch image classification model trained on 15k waste datasets
- The user uploads a photo from the frontend.
- The Flask backend receives it, processes it through the PyTorch model, and returns classification + emission data.
- Firebase stores user data.
- The frontend displays classification results and CO₂ estimates
- Utilized the MobileNetN2 deep learning model for computer vision tasks, particularly on mobile and embedded devices due to its speed