1) The document describes a face mask detection system created by students to help enforce mask mandates during the COVID-19 pandemic.
2) The system uses OpenCV, Keras/Tensorflow and deep learning to train a model to detect faces and determine if they are wearing a mask or not wearing a mask.
3) The model was trained on a dataset of images labeled as 'mask' and 'no mask' and achieved 99% accuracy on the test set.