FASK: A face mask detector implementation

a project with tensorflow and opencv.

This is the 3rd of three individual projects

1. First Project: A face dataset maker, using a state of the art SSD trained on face detection we created a dataset of faces with mask and without mask.
2. Second Project: Creating a classifier of mask or nomask classes. We use transfer learning on a trained mobilenet.
3. Third Project (this one): Using the classifier of the 2nd project and the SSD of the first project to detect masks on real time.

Examples:

Many faces:

or one...

Uses:

In a command line type

python fask.py

by default uses the first webcam, but you can give it a video by typing

python fask.py -i "path_to_my_video/my_video.mp4"

By default uses a confident value of 0.5, you can set to other desired value between 0 and 1 by typing

python fask.py -c X

X=desired confidence value