This is a code sample repository for using Convolutional Neural Network (CNN) to build an image classification model to recognition people are correctly wearing the face mask or not.
The source image dataset is coming from MaskedFace-Net. MaskedFace-Net is a dataset of human faces with a correctly or incorrectly worn mask (133,783 images) based on the Flickr-Faces-HQ (FFHQ) dataset.
- correctly_face_masked_recognition_basic_neural.ipynb <-- Python notebook for CNN model training with basic CNN and scoring.
- /images/Correctly Masked.zip <-- Correctly Face Masked dataset, contain 950 samples.
- /images/Incorrectly Masked.zip <-- Incorrectly Face Masked dataset, contain 928 samples.
- /images/No Face Detected.zip <-- No Face Detected dataset, contain 60 samples.
- /test_images/Correctly Masked/ <-- Contain 3 Correctly Face Masked sample images for testing purpose.
- /test_images/Incorrectly Masked/ <-- Contain 3 Incorrectly Face Masked sample images for testing purpose.
- /test_images/No Face Detected/ <-- Contain 1 No Face Detected sample images for testing purpose.
Enjoy.
Below is a screenshot of sample images.
