The world is fighting with Covid19 pandemic. There are so many essential equipments needed to fight against Corona virus. One of such most essential is Face Mask. Firstly face mask was not mandatory for everyone but as the day progresses scientist and Doctors have recommended everyone to wear face mask. Now To detect whether a person is wearing Face Mask or not, we will use Face Mask Detection Technique. Face Mask Detection Platform utilizes Artificial Network to perceive if a person does/doesn’t wear a mask.,also The temperature will be measured by thermal camera AMG8833

Step 1: Motivation

In the present scenario due to Covid-19, there is no efficient face mask detection applications which are now in high demand for transportation means, densely populated areas, residential districts, large-scale manufacturers and other enterprises to ensure safety.

Step 2: TechStack/framework Used

Step 3: Features

Our face mask detector didn't use any morphed masked images dataset. The model is accurate, and since we used the MobileNetV2 architecture, it’s also computationally efficient and thus making it easier to deploy the model to embedded systems (Raspberry Pi, Google Coral, etc.).

This system can therefore be used in real-time applications which require face-mask detection for safety purposes due to the outbreak of Covid-19. This project can be integrated with embedded systems for application in airports, railway stations, offices, schools, and public places to ensure that public safety guidelines are followed.

Panasonic's AMG8833 Grid-EYE is an 8x8 array of IR thermal sensors. When connected to your microcontroller (or raspberry Pi) it will return an array of 64 individual IR temperature readings over I2C.

Step 4: Dataset

The dataset used can be downloaded here - Click to Download

This dataset consists of 3835 images belonging to two classes: with_mask: 1916 imageswithout_mask: 1919 imagesThe images used were real images of faces wearing masks. The images were collected from the following sources:

Bing Search API (See Python script)

Kaggle datasetsRMFD

dataset (See here)

Raspberry Pi Thermal Camera (AMG8833)

Step 5: Prerequisites

All the dependencies and required libraries are included in the file requirements.txt See here

Step 6: Installation

1.Clone the repo

$ git clone

2.Change your directory to the cloned repo and create a Python virtual environment named 'test'

$ mkdir test

3.Now, run the following command in your Terminal/Command Prompt to install the libraries required

$ pip3 install -r requirements.txt

Step 7: Working

1.Open terminal. Go into the cloned project directory and type the following command:

$ python3 --dataset dataset

2.To detect face masks in an image type the following command:

$ python3 --image images/pic1.jpeg

3.To detect face masks in real-time video streams type the following command:

$ python3

'for Thermal Camera AMG8833'

4.Install Python Software

sudo apt-get install -y python-scipy python-pygame
sudo pip3 install colour

5.Wiring Up Sensor

With the Pi powered off, we can wire up the sensor to the Pi Cobbler like this:

  • Connect Vin to the 3V or 5V power supply (either is fine)
  • Connect GND to the ground pin on the Cobbler
  • Connect SDA to SDA on the Cobbler
  • Connect SCL to SCL on the Cobbler

You can also use direct wires, we happen to have a Cobbler ready. remember you can plug the cobbler into the bottom of the PiTFT to get access to all the pins!

6.Now you should be able to verify that the sensor is wired up correctly by asking the Pi to detect what addresses it can see on the I2C bus:

sudo i2cdetect -y 1

7.Run example code (AMG8833)

"""This example is for Raspberry Pi (Linux) only!
   It will not work on microcontrollers running CircuitPython!"""

import os
import math
import time

import busio
import board

import numpy as np
import pygame
from scipy.interpolate import griddata

from colour import Color

import adafruit_amg88xx

i2c_bus = busio.I2C(board.SCL, board.SDA)

#low range of the sensor (this will be blue on the screen)

#high range of the sensor (this will be red on the screen)

#how many color values we can have

os.putenv('SDL_FBDEV', '/dev/fb1')

#initialize the sensor
sensor = adafruit_amg88xx.AMG88XX(i2c_bus)

# pylint: disable=invalid-slice-index
points = [(math.floor(ix / 8), (ix % 8)) for ix in range(0, 64)]
grid_x, grid_y = np.mgrid[0:7:32j, 0:7:32j]
# pylint: enable=invalid-slice-index

#sensor is an 8x8 grid so lets do a square
height = 240
width = 240

#the list of colors we can choose from
blue = Color("indigo")
colors = list(blue.range_to(Color("red"), COLORDEPTH))

#create the array of colors
colors = [(int( * 255), int( * 255), int( * 255)) for c in colors]

displayPixelWidth = width / 30
displayPixelHeight = height / 30

lcd = pygame.display.set_mode((width, height))

lcd.fill((255, 0, 0))


lcd.fill((0, 0, 0))

#some utility functions
def constrain(val, min_val, max_val):
    return min(max_val, max(min_val, val))

def map_value(x, in_min, in_max, out_min, out_max):
    return (x - in_min) * (out_max - out_min) / (in_max - in_min) + out_min

#let the sensor initialize

while True:

    #read the pixels
    pixels = []
    for row in sensor.pixels:
        pixels = pixels + row
    pixels = [map_value(p, MINTEMP, MAXTEMP, 0, COLORDEPTH - 1) for p in pixels]

    #perform interpolation
    bicubic = griddata(points, pixels, (grid_x, grid_y), method='cubic')

    #draw everything
    for ix, row in enumerate(bicubic):
        for jx, pixel in enumerate(row):
            pygame.draw.rect(lcd, colors[constrain(int(pixel), 0, COLORDEPTH- 1)],
                             (displayPixelHeight * ix, displayPixelWidth * jx,
                              displayPixelHeight, displayPixelWidth))


Step 8: Results

  • Our model gave 93% accuracy for Face Mask Detection after training via tensorflow-gpu==2.0.0
  • We got the following accuracy/loss training curve plot

Step 9: Graduation Project Report