A machine learning system that classifies different types of waste materials into categories. It's built on TensorFlow and leverages the VGG16 model architecture for image classification.
-
VGG16 Model Architecture: The project utilizes the power of the pre-trained VGG16 model for image classification.
-
12 Waste Categories: Classify waste into 12 distinct categories including cardboard, metal, paper, and more.
-
Image Augmentation: Uses
ImageDataGenerator
for real-time data augmentation. -
Visual Predictions: Provides a visual representation of predictions using Matplotlib.
-
Clone the repository:
git clone https://github.com/supriya811106/Garbage-Classification-System.git
-
Navigate to the cloned repository:
cd Garbage-Classification-System
-
Install the required dependencies:
pip install -r requirements.txt
-
Running the Notebook:
-
Ensure you have Jupyter Notebook or Jupyter Lab installed. If not, install it:
pip install jupyterlab
-
Launch Jupyter:
jupyter lab
-
Navigate to the project notebook and run the cells to either train the model or make predictions.
-
-
Classifying Waste:
- Use the
waste_prediction
function within the notebook to classify waste by providing the path to your image.
- Use the
The trained model is saved as predictWaste12.h5
and can be loaded using TensorFlow/Keras for further predictions or improvements.