🗑️ Waste Classification using Deep Learning (VGG16 + Transfer Learning)

♻️ Organic vs Recyclable Waste Classifier

A deep learning project built with TensorFlow & Keras, using Transfer Learning (VGG16) to classify waste images into Recyclable and Organic categories.

This project was developed as the final capstone of my Deep Learning with Keras & TensorFlow course.

---

🚀 Project Overview

Waste management is a growing challenge in modern cities. Manual sorting is:

• ❌ Time-consuming
• ❌ Error-prone
• ❌ Resource-intensive

This project leverages AI-powered image recognition to automate waste classification, improving recycling efficiency and reducing contamination.

✅ Input: Waste image (e.g., food, bottles, paper)
✅ Output: Organic 🌱 or Recyclable ♻️

---

🔑 Key Features

• 📸 Image preprocessing & augmentation
• 🧠 Transfer Learning with VGG16
• 🎯 Fine-tuning for improved accuracy
• 📊 Training & validation performance visualization
• 🔍 Test image predictions

---

🛠️ Tech Stack

• Python 3.10+
• TensorFlow / Keras
• Matplotlib & Seaborn (visualization)
• NumPy & Pandas (data handling)
• Google Colab / Jupyter Notebook

---

📂 Project Structure

├── Final Proj-Classify Waste Products Using TL- FT-completed.ipynb         # Main notebook             
├── o-vs-r-split/                                                           # Dataset
│   └── train
│   │   └── O
│   │   └── R
│   └── test
│       └── O
│       └── R
├──  README.md                                                   # Project documentation
├──  requirements.txt                                            # Requirements 
└── .gitignore                                                   # Ignore big files & venv

---

🚀 How to Run

1. Clone the repo:

    git clone https://github.com/buildwithmehul/waste-classification-using-transfer-learning.git
    cd waste-classification-using-transfer-learning

2. Create a virtual environment & install dependencies

    python -m venv venv
    source venv/bin/activate   # On Mac/Linux
    venv\Scripts\activate      # On Windows

    pip install -r requirements.txt

3. Open the notebook:

    jupyter notebook notebooks/Final\ Proj-Classify\ Waste\ Products\ Using\ TL-FT-v1.ipynb

4. Run all cells to train/evaluate the model.

---

📊 Workflow

1. Load dataset and preprocess images.
2. Apply transfer learning using pre-trained CNNs (e.g., MobileNet, ResNet).
3. Fine-tune model layers for better feature extraction.
4. Train and evaluate model performance.
5. Save trained model for deployment.

---

📊 Results & Visuals

Accuracy / Loss Curves

Example Prediction

Input Image: 🍌 banana peel Model Output: Organic 🌱

Input Image: 🥤 plastic bottle Model Output: Recyclable ♻️

---

💡 Future Improvements

• Extend to more waste categories (metal, glass, paper, plastic)
• Deploy as a web app with Streamlit / Flask
• Optimize inference for edge devices (e.g., Raspberry Pi in smart bins)

---

🤝 Contributing

Contributions are welcome! Feel free to open issues or submit PRs.

---

✨ Acknowledgements

• VGG16 Pre-trained Model (ImageNet)
• Course: Deep Learning with Keras & TensorFlow
• Inspiration: Real-world need for sustainable waste management

---

🎓 Certificate

---

📜 License

This project is licensed under the MIT License - see the LICENSE file for details.