EcoScan

log in page
sign up page
invalid image of a tree/plant
About us page
valid image of a tree/plant
Analyze image
upload trees/plant page
Home Page
recycling chatbot
Verifying plant page
Tree Collection page
Upload plants that you plant page

Inspiration

EcoScan was inspired by the need for sustainable solutions in our everyday lives. We wanted to provide a platform where users could learn about recycling, plant trees, and make a positive environmental impact. Our goal is to encourage responsible actions for the environment and empower people to contribute to tree growth while earning rewards.

What it does

Sign Up/Login: Users can create personalized accounts to track their progress and rewards.
Recycling Tips: Upload images of items for AI-powered recycling tips based on how to recycle or dispose of them.
Tree Rewards: Users can upload images of trees they've planted and receive rewards for contributing to environmental sustainability.
Image Validation: The AI validates whether the uploaded image is of a plant or tree, and scores are given accordingly.
Delete Trees: Users can delete trees from their collection, which will lower their rewards.
Quiz: Take a quiz to discover which organizations align with your interests for donations or volunteer opportunities.

How we built it

EcoScan was developed using a variety of technologies:

Backend: Flask (Python) to build the web application and handle server-side logic.
Frontend: HTML, CSS, and JavaScript for the user interface and interaction.
Database: SQLAlchemy with SQLite to manage user accounts, uploaded images, and rewards.
AI/ML: Gemini AI models for image recognition, and also built models from scratch, receiving accuracy of 94%: classification between organic and recyclable.
Machine Learning Frameworks: PyTorch, TensorFlow, and FastAI for model building and training.
TinType: Used TinType for project management, keeping a track of tasks completed, and integrating with Trello to make workflow easier.

Challenges we ran into

Backend Integration: One of the main challenges was integrating the backend system to handle user accounts and manage the database efficiently using SQLAlchemy and SQLite.
Image Recognition: Ensuring that the AI model could accurately validate plant/tree images presented some difficulties, especially regarding the model’s ability to detect and classify images with high accuracy.
Database Management: Setting up the database schema to store user data and image records posed challenges, especially as we scaled the platform.

Accomplishments that we're proud of

Built an AI-powered platform that provides users with personalized recycling tips based on uploaded images.
Developed a tree rewards system, allowing users to earn points and rewards by planting trees and uploading images.
Created a quiz feature that helps users find suitable organizations for donations or volunteering.
Set the foundation for future collaborations with environmental organizations to promote sustainability further.
Learned A LOT of CSS for styling Flask website!

What we learned

Backend Development: Gained a solid understanding of how to manage user accounts and interact with databases using SQLAlchemy and SQLite.
AI & ML Integration: Learned how to integrate computer vision models for image validation and how to leverage HuggingFace models for providing actionable recycling tips.
Database Management: Developed skills in designing databases and handling large-scale data operations within Python.
Technical Growth: We enhanced our skills in machine learning, AI model training (PyTorch, TensorFlow), and the application of modern Python libraries.

What's next for EcoScan

Collaboration with Organizations: We're wanting to go on partnerships with eco-friendly organizations, offering users incentives such as gift cards and swag for planting trees and getting involved.
Expanded AI Capabilities: Plan to enhance the image recognition model to increase accuracy and handle more diverse plant/tree images. Use models from HuggingFace or train more advanced models!
Recycling Education: Extend the platform's capabilities to offer more detailed and contextual recycling tips using advanced AI-powered language models.
Wider Impact: We envision EcoScan inspiring global tree planting efforts, with more people participating in tree planting and sustainable living practices.

Future Plans

User Engagement: Offer rewards for users who plant multiple trees, such as gift cards for eco-friendly products or food items.
Machine Learning Advancements: Integrate more advanced computer vision models and AI techniques to improve the accuracy of image classification.
Sustainability Partnerships: Collaborate with environmental groups and organizations to further expand EcoScan’s impact on global sustainability efforts.