There is no people eating function or toolbox, so you'll have to come up with something on your own. Most likely it will involve deep learning where you'll have to train with videos exactly what frame epople are eating and when they are not.
It's very much a judgment call. Are people eating if their hand is near their mouth, with or without a utensil? What about if their hand is not close and their mouth is closed but moving (chewing). What about mouth open and moving (chewing or talking?)? What about if they are sitting at a table, either by themselves or with others? You see, it's no so easily defined.
