EcoVision

Real-time analysis of items
Scanning of object

Title: EcoVision: Smart Recycling Detection App

🚀 Inspiration: In a world where recycling rules vary significantly and confusion about recyclable materials persists, EcoVision aims to tackle several key environmental challenges.

Project Overview: EcoVision is an intelligent recycling assistant that leverages advanced computer vision and AI to help users accurately identify recyclable items, learn proper disposal methods, and track their environmental impact, all in real-time.

Our solution, EcoVision, integrates AMD Ryzen™ AI technology to enhance our YOLOv8 model, allowing for efficient real-time detection of recyclables using the onboard Neural Processing Unit (NPU). This setup ensures minimal latency and maximized energy efficiency.

🔥 Key Innovations:

Real-Time Object Detection: Utilizes the device's camera to instantly identify recyclable items. AI-Enhanced Learning: Provides users with detailed information about each item, powered by the ChatGPT API. Environmental Impact Tracking: Users earn points for proper recycling and can view their overall environmental contribution. Gamification and Community: Includes a leaderboard to encourage competitive sustainability.

💻 Technical Stack:

Backend: Framework: FastAPI (Python), Node.JS Authentication: Firebase Authentication Database: Firebase AI Model: YOLOv8 with NPU acceleration, GPT 04-mini Frontend: Next.js, React, TypeScript, TailwindCSS

🌎 Environmental Impact: By facilitating better recycling habits, EcoVision helps reduce contamination in recycling streams, boosts recycling rates, and educates users about sustainable practices through interactive feedback. Designed to be environmentally friendly, the app minimizes power consumption and data transfer, aligning with broader ecological goals.

Future Prospects: Looking ahead, we plan to expand EcoVision’s capabilities to include more granular object recognition, broader material categorization, and multi-lingual support to increase accessibility and user engagement worldwide.