My model is an inverted image detector and can help detect if images are inverted with 99% accuracy.
I used a dataset containing people with and without masks. I trained my model on ~ 300 images of people without masks and tested it on ~ 60 of the same images distribution:
author = {Prasoon Kottarathil},
title = {Face Mask Lite Dataset},
year = {2020},
publisher = {kaggle},
journal = {Kaggle Data},
how published = {\url{
https://www.kaggle.com/prasoonkottarathil/face-mask-lite-dataset}} \
