Garbage Classification

Introduction

This project is a simple image classification project that classifies images of garbage into different classes of household waste.

Dataset

The dataset used for this project is the Garbage Classification dataset from Kaggle. The dataset contains 12 classes of garbage images. The classes are:

1. battery
2. biological
3. brown-glass
4. cardboard
5. clothes
6. green-glass
7. metal
8. paper
9. plastic
10. shoes
11. trash
12. white-glass

Model

The model used for this project has EfficientNet as the backbone, following by a fully connected layer.

Usage

To use this project, you need to have Python 3 installed. A virtual environment is recommended to manage your Python packages.

Setup

Install the required packages

To install the required packages, run the following command:

pip install -r requirements.txt

If you also want to use the jupyter notebook, you need to install the corresponding requirements with the following command:

pip install -r notebook_requirements.txt

Download the dataset

Download the dataset using the script below: python -m scripts.download_dataset. The script will ask you to provide your Kaggle username and Kaggle API key. You can get your Kaggle API key by following the instructions here.

Running the scripts

To use the repository, you can either train the model or use a pre-trained model. You can find on this link. (Note: the pretrained should be placed in the 'checkpoints' folder.)

Training

To train the model, run the following command:

python -m scripts.train

Testing

To test the model, run the following command:

python -m scripts.eval

Predicting

To predict the class of an image, run the following command:

python -m scripts.predict --image_path path/to/image

Visualizing the results

You can observe some results for the following images:

paper	plastic-bottle	wash-clot	white-glass

For that, just run the following command:

python -m scripts.visualize_results

You should see the following output:

Image 1/4
         Image: paper
         Prediction: paper
         Confidence: 0.999967634677887
         Class probabilities: [0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.00 0.00 0.00 0.00 0.00]
Image 2/4
         Image: wash-cloth
         Prediction: clothes
         Confidence: 1.0
         Class probabilities: [0.00 0.00 0.00 0.00 1.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00]
Image 3/4
         Image: white-glass
         Prediction: white-glass
         Confidence: 1.0
         Class probabilities: [0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.00]
Image 4/4
         Image: plastic-bottle
         Prediction: plastic
         Confidence: 1.0
         Class probabilities: [0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 1.00 0.00 0.00 0.00]

License

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