You are on page 1of 21

Automated Face

Mask Detection

A project by Nishant Goel


Under the Guidance of Dr. Anil Kumar
Mishra and Mr. Vivek Birla
Need For this Project
• To automate the unorganized offline System by
providing them user-friendly low cost and secured
online automated face mask detection system.
• To make face mask detection easier.
• To help small businesses adapt to new covid rules and
regulations.
• To upgrade nearby local shops facilitating easy and
convenient use of new and trending technologies.
• To make a system that restricts the growth of covid-19
by finding out who is not wearing a facial mask.
• Manual Monitoring is very difficult for officers to check
whether the peoples are wearing mask or not. So, in
our technique, We are using web cam to detect
people’s faces and to prevent from virus transmission.
Table Of Contents

1. Key Features
2. Technologies Used
3. Packages Used
4. Methodology
5. Learnings & Future Scope
Key Features
• An automated face mask detection system.
• It has fast and high accuracy.
• This system can be implemented in ATMs, Banks, etc.
• Interactive UI to attract various types of customers.
• We can keep people safe using this system.
• It requires less memory, making it easily deployable for
embedded devices used for surveillance purposes.
• Improved affine transformation is developed to crop the facial
areas from uncontrolled real-time images having differences in
face size, orientation and background.
• This step helps in better localizing the person who is violating
the facemask norms in public areas/ offices.
Technologies Used
Keras
Keras isis aa powerful
powerful and
and easy-to-use
easy-to-use
free
free open-source
open-source Python
Python library
library for
Keras developing
developing and and evaluating
evaluating deep
for
deep
learning
learning models.
models.

TensorFlow is an end-to-end open-


source platform for machine learning.
TensorFlow It has a comprehensive, flexible
ecosystem of tools, libraries and
community resources that lets
researchers push the state-of-the-art in
ML and developers easily build and
deploy ML powered applications.
Technologies Used

Python3.7 OpenCV-Python

Python is broadly utilized universally and is OpenCV-Python is a huge open-source


a high-level programming language. It was library for computer vision, machine
primarily introduced for prominence on learning, and image processing. It can
code, and its language structure enables process images and videos to identify
software engineers to express ideas in objects, faces, or even the handwriting
fewer lines of code. of a human.
Technologies Used
MobileNet Numpy

MobileNet is a streamlined NumPy is a general-purpose


architecture that uses depth array-processing package. It
wise separable convolutions provides a high-performance
to construct lightweight deep multidimensional array object,
convolutional neural and tools for working with
networks and provides an these arrays.
efficient model for mobile and It is the fundamental package
embedded vision for scientific computing with
applications. Python.
Technologies Used
Matplotlib

Matplotlib is an amazing
visualization library in Python
for 2D plots of arrays.
Matplotlib is a multi-platform
data visualization library built
on NumPy arrays and
designed to work with the
broader SciPy stack.
System Requirements
Processor:
Processor: 1.3GHZ
1.3GHZ
At
At least
least 500
500 Mb
Mb
of
of free
free space
space

Windows
Windows 77 or
or
Minimum
Minimum higher
higher versions
versions
8GB
8GB of
of RAM
RAM
Methodology
Working Model Of the Project
Dataset Used

Dataset with images without


mask

Dataset with images with mask


Results:-
1. Experimental Results
All the experimental trials have been conducted on a
laptop equipped by an Intel i5-8750H processor, 8 GB of
RAM with 1050ti max-Q. The Jupyter Notebook software
equipped with Python 3.7 kernel was selected in this
research for the development and implementation of
the different experimental trails.
The metrics selected for evaluation of SSDMNV2 model
is explained below.
Accuracy=Tp+Tn(Tp+Fp+Fn+Tn)
Precision=Tp(Tp+Fn)
Recall=Tp(Tp+Fn)
f1 score=2*Recall*Precision(Recall+Precision)
 
Where Tp = True positive,
Tn = True negative,
Fp = False positive,
Fn = False negative
Model Accuracy
2. Visualization of Results

Image without
mask
2. Visualization of Results

Image with
mask

Image with
partial wearing
of mask
Learnings & Future Scope
Learnings from the Project
We proposed an automated smart framework for screening persons
who are not using a face mask in this paper

The technique of OpenCV deep neural networks used in this model


generated fruitful results. Classification of images was done accurately
using the MobilenetV2 image classifier, which is one of the uniqueness
of the proposed approach. The ensemble approach not only helps in
achieving high accuracy but also improves detection speed
considerably. Furthermore, the application of transfer learning on pre-
trained models with extensive experimentation over an unbiased
dataset resulted in a highly robust and low-cost system. The identity
detection of faces, violating the mask norms further, increases the utility
of the system for public benefits.
Future Scope

Firstly, the proposed technique can be


integrated into any high-resolution video
surveillance devices and not limited to
mask detection only. Secondly, the model
can be extended to detect facial
landmarks with a facemask for biometric
purposes.

The above-mentioned points are the


enhancements which can be done to
increase the applicability and usage of
this project.
Acknowledgements

Dr.Anil Kumar Mishra,


Assistant Professor,
Amity University
Gurugram
Mr. Vivek Birla,
Assistant Professor,
Amity University
Gurugram
Thank You

You might also like