INTRODUCTION

To create our face mask detector, I trained a two-class model of people wearing masks and people not wearing masks.

I fine-tuned MobileNetV2 on our mask/no mask dataset and obtained a classifier that is ~99% accurate.

I then took this face mask classifier and applied it to both images and real-time video streams by:

• Detecting faces in images/video
• Extracting each individual face
• Applying our face mask classifier

This face mask detector is accurate, and since I used the MobileNetV2 architecture, it’s also computationally efficient, making it easier to deploy the model to embedded systems (Raspberry Pi, Google Coral, Jetosn, Nano, etc.).

Project Structure

$ tree --dirsfirst --filelimit 10
.
├── dataset
│   ├── with_mask [690 entries]
│   └── without_mask [686 entries]
├── examples
│   ├── example_01.png
│   ├── example_02.png
│   └── example_03.png
├── face_detector
│   ├── deploy.prototxt
│   └── res10_300x300_ssd_iter_140000.caffemodel
├── detect_mask_image.py
├── detect_mask_video.py
├── mask_detector.model
├── plot.png
└── train_mask_detector.py

5 directories, 10 files

The dataset/ directory contains the dataset i used.

Three image examples/ are there to test the static image face mask detector.

• train_mask_detector.py: Accepts the input dataset and fine-tunes MobileNetV2 upon it to create the mask_detector.model. A training history plot.png containing accuracy/loss curves is also produced
• detect_mask_image.py: Performs face mask detection in static images
• detect_mask_video.py: Using the webcam, this script applies face mask detection to every frame in the stream

RESOURCES

Pretrained Model Link:- https://drive.google.com/file/d/1BeR4ukiaS5ZU5Dq3TU9Yc51Ymo8zP_Vs/view?usp=sharing

Dataset Link:- https://drive.google.com/folderview?id=1XDte2DL2Mf_hw4NsmGst7QtYoU7sMBVG

Acknowledgment

https://www.youtube.com/watch?v=TNZAbVNTLhA&t=6172s