Winner of 1st Place General Category, iRobot Best Home Robot, MakeMIT 2021
We created a device that automatically detects if a person is walking through a doorway without a mask, and if no mask is found, drops a mask onto them from overhead. The device stores masks on a corkscrew, which a stepper motor turns to drop the masks. The device detects maskless faces using a webcam mounted at eye-level, which is plugged into a laptop. The laptop runs a computer vision (CV) algorithm, and sends signals to a companion computer that performs the actuation. The system also features an infrared (IR) sensor, which will measure temperature and send you a text if your temperature is too high.
The mask detection code is taken from here, which uses YOLO finetuned on the WIDER FACE and MAFA datasets. Both the CV and embedded code is written in Python, with the Tinkerforge Python APIs used to drive the stepper motor and read from the IR sensor. The text messages are sent over wifi using Twilio.
This project won first place overall at MakeMIT 2021, which we were really excited about. It also received the Best Home Robot sponsor award from iRobot.
For more information, check out our Devpost Page and read through our GitHub repository!
