PET Perplexity ♻️

Winner of the Agglomeration 2.0 Hackathon🏆 (Team internship_needed - AG41)

PET Perplexity is an intelligent, automated polymer segregation system designed to revolutionize plastic waste management. It utilizes a cross-platform mobile application powered by advanced Computer Vision and Machine Learning to detect, classify, and analyze PET bottles in real-time.

The system addresses the challenge of segregating plastic waste by identifying key attributes such as bottle presence, size, brand, and material properties.

🚀 Key Features

• Real-time Object Detection: Instantly detects PET bottles within a video feed or captured image.
• Bottle Size Classification: Automatically categorizes bottles into standard sizes (e.g., small, medium, large) to aid in sorting logistics.
• Brand Recognition: Identifies the brand of the bottle using custom-trained deep learning models.
• Material Analysis: Utilizes XGBoost algorithms to analyze polymer characteristics for precise segregation.
• Batch Scanning: Capability to process multiple items in a batch for high-throughput environments.
• User Dashboard: A comprehensive mobile interface for tracking scan history and segregation statistics.

🛠️ Tech Stack

Frontend (Mobile App)

• Framework: Flutter (Dart)
• Platforms: Android, iOS, Web
• State Management: Provider / Riverpod (Inferred)
• Architecture: Feature-first architecture (lib/features/)

Backend & Machine Learning

• Languages: Python, Node.js
• Frameworks: Flask, FastAPI / Uvicorn, Gradio
• Computer Vision:
  • YOLOv5 / YOLOv8: For robust object detection and bounding box regression.
  • OpenCV (cv2): For image preprocessing and frame manipulation.
• Deep Learning Models:
  • EfficientNet-B2: Finetuned for high-accuracy bottle size classification.
  • Custom CNNs (PyTorch): For brand logo detection and classification.
• Machine Learning:
  • XGBoost: For tabular data analysis and material property prediction.

🧠 ML Pipeline Architecture

The system operates on a microservices-based architecture where the Flutter app communicates with specialized ML services:

1. Detection Layer: The input image is passed through a YOLO model to detect the presence and location of a bottle.
2. Dimension Layer: Cropped regions of interest are sent to the Dim Predictor (EfficientNet-B2) to estimate physical dimensions and volume.
3. Brand Layer: The Brand Predictor analyzes visual features to classify the bottle's brand, aiding in source separation.
4. Analysis Layer: The XGBoost model aggregates these features to make a final segregation decision.

📂 Project Structure

agglomeration-2.0/
├── internshipneeded/
│   ├── agglomeration-2.0-33bf.../  # Main Flutter Application
│   │   ├── lib/
│   │   │   ├── features/           # UI Screens (Home, Scan, History)
│   │   │   ├── services/           # API Integration (Auth, ScanService)
│   │   │   └── main.dart           # App Entry Point
│   │   └── pubspec.yaml            # Dart Dependencies
│   │
│   ├── brand_predictor/            # Brand Recognition Service
│   │   ├── app.py                  # API Entry Point
│   │   ├── pipeline.py             # Inference Pipeline
│   │   └── model_arch.py           # PyTorch Model Architecture
│   │
│   ├── dim_predictor/              # Dimension/Size Service
│   │   ├── app.py                  # Flask App for Size Classification
│   │   └── README.md
│   │
│   ├── Agglomeration-2.0-bottlesize/ # Size Classification Model Training
│   │   └── model.py                # EfficientNet-B2 Implementation
│   │
│   └── Agglomeration-2.0-ML/       # Core ML & XGBoost Logic
│       ├── app.py                  # Gradio/Python App Interface
│       └── xgboost_main.py         # XGBoost Logic

---

🤝 Contributing

Contributions are always welcome!

1. Fork the repository.
2. Create a feature branch (git checkout -b feature/AmazingFeature).
3. Commit your changes.
4. Open a Pull Request.

---

👥 Development Team

• Aditya (Mobile App & UI/UX Development)
• Samarth Agarwal (Backend Development)
• Apurva Arya, Arnav Tripathi, Suryansh Kulshreshtha (AI & ML/DL Model Development)