Download as pptx, pdf, or txt
Download as pptx, pdf, or txt
You are on page 1of 13

A Real-time face mask

detection system

ALMA ANN GEORGE -19BEE1012


VRISCHIKA VINAYAN NAIR -19BEE1054
FEBIN KOSHY JACOB -19BEE1057
ANTO JOY -19BEE1084
INTRODUCTION:

The COVID-19 pandemic has reshaped life as we know it. Many of us are staying home, avoid
people on the street and changing daily habits, like going to school or work, in ways we never
imagined.

While we are changing old behaviors, there are new routines we need to adopt. First and
foremost is the habit of wearing a mask or face covering whenever we are in a public space.

For example in an office it should be necessary for all the employs to wear a mask. So to
reduce the man power used in checking whether a person is wearing mask or not our project
will automate the process and will detect whether a person is wearing mask or not
Existing work with limitations

In the current scenario in all public place and offices uses a security guard to check whether people
are wearing a mask or not. The main problem that arises here are.

 Extra manpower

 Putting a person in a risky job


SOFTWARE AND HARDWARE REQUIREMENTS :

a) Software requirements
• An Operating system
• Python 3.7 or higher
• Flask 1.0 or higher
• Any browser

b) Hardware Requirements
• Processor - Intel i5 or higher
• Disk space – 512 MB or higher
• Ram – 8GB or higher
• Camera
METHODOLOGY:

1. Training:

A dataset with images categorized into with_mask and without_mask is used .For testing
25% of the dataset is used. For testing, the image are preprocessed . The preprocessed data is
used for making the model. The remaining images in the dataset and some augmented data that
was created are used to train the model. The trained model is saved and the accuracy, loss graph
is plotted.

2. Deployment:

The trained model is loaded, the faces are detected in video stream and then the region of
interest (ROI) is extracted, Finally the face mask detector model is applied and faces in video
streams are classified as with mask or without mask.
Overall system architecture diagram
Phase 1 : Train Face Mask Detector

Train face mask


Load face mask Serialize face mask
classifier with
dataset classifier to disk
Keas/TensorFlow

Phase 2 : Apply Face Mask Detector

Load face mask Detect faces in


Extract each face ROI
classifier form disk image/video stream

Apply face mask


Show result classifier to each face
RIO to determine
“mask” or “no mask”
Phase 1 - Train Mask Detector
Phase 2 - Apply Face Mask Detector
Screenshot Of Application

Without Mask
Screenshot Of Application

With Mask
CONCLUSION:

This corona time, we have to use masks and hand sanitizer everywhere and in almost all
places automatic hand sanitizer machines will be available.

This project can be easily be implemented in exiting surveillance systems and can used in
public places like hospitals, offices, airports, etc.
REFERENCE:

• Wang, Zhongyuan, et al. “face mask recognition dataset and application." arXiv preprint
arXiv:2003.09093 (2020).

• Lin, Shang-Hung. "An introduction to face recognition technology." Informing Sci. Int. J. an Emerg.
Transdiscipl. 3 (2000): 1-7.

• Jalled, Fares, and Ilia Voronkov. "Object detection using image processing." arXiv preprint
arXiv:1611.07791 (2016).

• https://opensource.com/article/18/4/flask -An introduction to the Flask Python web app framework


• https://www.pyimagesearch.com/2019/09/02/opencv-stream-video-to-web-browser-html-page/
Thank
You

You might also like