♻️ Waste Classification App

This Streamlit web application allows users to classify images of waste into two categories: Organic or Recyclable. It leverages a pre-trained TensorFlow/Keras deep learning model to provide real-time predictions directly in your browser, either through image upload or live camera feed.

---

Features

• Dual Input Modes: Seamlessly switch between uploading an image file (JPG, JPEG, PNG, WEBP) or capturing a live picture using your device's camera.
• Instant Classification: Get immediate predictions (Organic or Recyclable) with a confidence score.
• Dynamic UI: The prediction section automatically clears when you switch between input methods, ensuring a clean user experience.
• Clear Camera Image: A dedicated button to clear the captured camera image and reactivate the camera.
• Simple Interface: Built with Streamlit for an intuitive and user-friendly web interface.

---

How to Run Locally

Follow these steps to get the app running on your local machine.

Prerequisites

• Python 3.8+
• pip (Python package installer)

1. Clone the Repository

git clone https://github.com/Roshians/Waste-Classification.git
cd Waste-Classification

2. Create a Virtual Environment (Recommended)

python -m venv venv

# On Windows
.\venv\Scripts\activate

# On macOS/Linux
source venv/bin/activate

3. Install Dependencies

pip install -r requirements.txt

requirements.txt content:

tensorflow==2.19.0
streamlit==1.45.1
numpy==2.1.3
pillow==11.1.0

4. Place Your Model File

Ensure your pre-trained Keras model file (best_model.h5) is in the same directory as app.py.
If your model is elsewhere, update the MODEL_PATH variable in app.py accordingly.

5. Run the Streamlit App

streamlit run app.py

This command will open the Streamlit app in your default web browser.

---

Model Information

This application uses a deep learning model based on the MobileNetV2 architecture to perform waste classification.

• Base Model: MobileNetV2, a lightweight and efficient convolutional neural network designed for mobile and embedded vision applications.
• Model File: best_model.h5
• Image Size: Resized to (128, 128) pixels (IMAGE_SIZE)
• Normalization: Pixel values are divided by 255.0 (RESCALE_FACTOR) to scale them to the 0–1 range.
• Classes:
  • Recyclable (Assumed index 0)
  • Organic (Assumed index 1)

⚠️ Note: The prediction logic in app.py assumes predictions[0][0] is the probability for 'Recyclable'. If your model's output order differs, adjust the predicted_class_index logic accordingly.

---

Usage

1. Launch the app:

   streamlit run app.py

2. Choose Input Method:

   • "Upload Image" or
   • "Use Camera" (via radio buttons)

3. Provide Image:

   • Upload Image: Click "Choose an image..." and select a file.
   • Use Camera: Click "Take a picture" to capture via webcam.

4. View Prediction:

   • The app classifies the image and displays the predicted category with a confidence score.

5. Reset:

   • Switching input methods automatically clears the image and prediction.
   • Use "Clear image" in camera mode to retake a picture.

---

Contributing

Pull requests are welcome! If you find bugs or have improvements in mind:

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Open a pull request

---

License

This project is released into the public domain under The Unlicense.

---

Acknowledgements

• Built and Hosted with Streamlit
• Waste Classification Dataset on Kaggle