This is an individual IoT project was done by me as part of a module at my university. Those days were the outbreak of the COVID pandemic. So I got an idea to develop a website platform that records students who are entering the university by checking these things.
- Whether the student is wearing a mask or not
- Temperature status of the student
From this, the platform gets the idea of the state of the student. If the student is not wearing a mask, then the student will be advised, or if the student has a high temperature, then the student will be inspected more and will be sent to get the required medical treatment. Here I used an ESP8266 as my controller because it’s cheap and well-suited in my case because of its Wi-Fi capability. As the temperature sensor, I used an NTC Temperature Sensor Module. At that time, the challenging part for me was detecting whether the student was wearing a face mask or not. For that, I used pre pre-trained model to get the boundary box of the face. I created a labeled dataset of 1,200 images categorized by whether individuals were wearing a face mask or not. From that, I trained a model and used that in combination with the previous pre-trained model for face boundary detection to detect whether the person is wearing a mask or not. That model had high accuracy in bright environments. But on the other hand, it faced accuracy issues in dark light environments.
At that time MERN stack was trending. So I developed the database management website using the MERN stack. After all, my Extended face masking system performed very well and I got good marks for that module evaluation :)