Skip to content

Latest commit

 

History

History
65 lines (51 loc) · 2.11 KB

README.md

File metadata and controls

65 lines (51 loc) · 2.11 KB

Smart Waste Classification & Recycling Suggestion System

Overview

This project implements a deep learning model using MobileNet for smart waste classification. It categorizes waste into different classes using the TrashNet dataset and provides recycling suggestions to promote sustainable waste management.

Features

  • Real-time Image Classification: Upload an image, and the model predicts the waste category.
  • Recycling Suggestions: Get recommendations on how to dispose of or recycle the classified waste.
  • Streamlit Web Application: A user-friendly interface for easy interaction.

Dataset

  • Source: TrashNet Dataset
  • Classes: Plastic, Metal, Glass, Paper, Cardboard, and Organic Waste
  • Preprocessing: Image resizing, normalization, and augmentation for better model generalization.

Model Architecture

  • Base Model: MobileNet (Pretrained on ImageNet)
  • Fine-tuning: Last few layers trained on the TrashNet dataset
  • Optimizer: Adam
  • Loss Function: Categorical Cross-Entropy
  • Metrics: Accuracy

Installation

  1. Clone the repository: 
    git clone https://github.com/arpanpramanik2003/smart-waste-classification.git
cd smart-waste-classification
  2. Install dependencies: 
    pip install -r requirements.txt
  3. Run the Streamlit app: 
    streamlit run app.py

Usage

  1. Open the Streamlit web app.
  2. Upload an image of waste.
  3. View the predicted category and recycling suggestions.

Customization

  • Image Size Adjustment: Ensures that uploaded images appear correctly in the app.
  • Expanded Recycling Information: Provides more details on how to recycle different waste types.

Results

  • Training Accuracy: ~92%
  • Validation Accuracy: ~88%
  • Loss: Optimized for minimal classification error

Deployment

  • The model can be deployed on platforms like Render, AWS, or Hugging Face Spaces for online access.

Contributing

Feel free to open issues and contribute to improving the project.

License

Apache-2.0 Licence

Author

Arpan Pramanik

For any queries, reach out via GitHub or LinkedIn.