RecyclEye

Inspiration We were inspired by the growing waste crisis and the disconnect between recycling systems and community engagement. We saw an opportunity to empower citizens with technology—combining cutting‐edge computer vision with gamification—to drive real behavioral change in waste management.

What It Does RecyclEye is an AI-powered platform that instantly classifies trash and recyclables from images taken on your mobile device. Users simply snap or upload a photo, and our YOLOv8 model provides real-time feedback. Every unique classification earns points, helping communities track their impact and identify local litter hotspots.

How We Built It Backend: Built with Flask, our server integrates YOLOv8 for image classification. It handles file uploads, converts images to JPEG (to handle various formats like HEIC), and prevents duplicate submissions using MD5 hashing. A SQLite database stores user data and points.

Frontend: Developed in React with Vite, our mobile-friendly interface lets users register, capture live photos or select from their gallery, and view their results along with a dynamic leaderboard.

Integration: We connected the frontend and backend via RESTful endpoints, using CORS and, in production, a proxy to ensure seamless communication.

Challenges We Ran Into Dependency Setup: Configuring the right versions of Python, PyTorch, and YOLOv8 was a challenge. File Format Compatibility: Handling and converting different image formats (like HEIC from iPhones) required extensive debugging. Cross-Origin Integration: Connecting our React frontend with the Flask backend across different ports was tricky. Git & Repository Management: Managing large files and cleaning up training artifacts in our repo posed unexpected hurdles.

Accomplishments That We're Proud Of Successful YOLO Integration: We seamlessly integrated YOLOv8 to deliver accurate, real-time trash detection. Robust File Handling: Our image conversion and duplicate detection system ensures reliable performance. Engaging User Interface: We built a mobile-first UI that makes environmental cleanup fun and interactive. End-to-End System: Overcoming integration challenges, we delivered a complete solution that brings AI and community engagement together.

What We Learned The critical role of robust error handling in full-stack development. How to integrate advanced computer vision models into practical, real-world applications. The importance of user-friendly design and gamification in driving community participation. Best practices for dependency management and repository organization in collaborative projects.

What's Next for RecyclEye Augmented Reality Integration: Develop AR overlays that provide live, visual feedback on waste classification. Smart Bin Prototypes: Create automated bins or robots equipped with our technology for real-time sorting. Enhanced Data Analytics: Expand our mapping and visualization tools to pinpoint litter hotspots and guide local cleanups. Partnerships & Incentives: Collaborate with local governments and eco-friendly brands to reward community efforts and promote sustainable practices.

Built With

• flask
• googlecollab
• python
• react
• sqlite
• tacodataset
• tailwindcss
• typescript
• ultralytics
• yolov8