As we’ve seen in the previous exercises, humans are slow. They take a long time to think about what they want to do, and they don’t always get it right. Also, filling out the same form repeatedly is tedious and time consuming, and people get tired of doing the same thing over and over.
We saw that we can reduce some of this time by reducing the number of forms, but still, there is a form that requires human interaction.
In this final exercise we will see how we can automate an entire process by removing all human interaction and still achieving a result.
In order for this portion of the workshop to work, we are going to change the overall task a little bit. We could continue with the costume evaluation process, but in order to do so we would first have to train a Visual AI model to detect the things we’re looking for, and that is well outside of the scope of this workshop.
If training an AI model is something you’re interested in doing, please feel free to contribute that model to this github repo so that others can use it.
Since we don’t have a trained AI model for costumes and ages, we’re going to use a pre-trained Google Vision AI model to evaluate the emotional state of a person and dispense candy depending on their state.