Try it yourself:-> marks 30 points on your face (landmark) and distance between any 2 points is an input each on which model decided the state of the person
^
||
High level on how the model works
This works on images and video is contionus frames of images; optimized it that way
"Model architecture and trained model with parameters are 2 different files above"