This project is done in two parts. In the first part I utilized the functionality of Keras and OpenCV to combine the combination of pre-built face detection models using OpenCV and then use a mask/non-mask classification, using Convolutional Neural Networks (CNN). The notebook to this approach is available here
In the second part of this project, I used the power of TensorFlow Object Detection API to train my own customized Face Mask Classifier, using around 170 images of people either wearing face mask or not and then labeling them. You can check the notebook of the end result here and the report file is available here