Note: For a nicer rendering of the notebook, click on this link to NBViewer.
In this project, I used a Convolutional neural-network architecture using the PyTorch framework on image data to create a model that can detect whether or not an image has: a fully covered mask,
a partially covered mask
a non covered face or
not a face at all
Below is a sample of the images used in this project, two samples from each category.
| Fully covered | Partially Covered | Not covered | Not Face |
|---|---|---|---|
 |
 |
 |
 |
 |
 |
 |
 |
- Convolutional Neural Networks
- Matplotlib
- Setting a learning rate decay
- decreasing the learning rate