Software

Requirements
Specification

GROUP - 4

-AMAN SINGH (IT)

-ANSH YADAV (IT)

-SURABHI NIGAM(IT)

-MUSKAN RAIKWAR (CS)

Table of Contents

1. Introduction
1.1 Purpose
1.2 Document Conventions
1.3 Intended Audience and Reading
1.4 Product Scope
2. Overall Description
2.1 Product Perspective
2.2 Product Functions
2.3 Operating Environment
2.4 Design and Implementation Constraints
3.External Interface Requirements
3.1 User Interfaces
3.2 Software Interfaces
4 Graph Visualization
5.Other nonfunctional Requirements
5.1 Performance Requirements
5.2 Software Quality Attributes
References
1.Introduction
1.1 Purpose
Face mask detection refers to detect whether a person is wearing a mask or not. In fact, the
problem is reverse engineering of face detection where the face is detected using different
machine learning algorithms for the purpose of security, authentication and surveillance.

1.2 Document Conventions

Effective strategies to restrain COVID-19 pandemic need high attention to mitigate negatively impacted
communal health and global economy, with the brim-full horizon yet to unfold. In the absence of
effective antiviral and limited medical resources, many measures are recommended by WHO to control
the infection rate and avoid exhausting the limited medical resources. Wearing a mask is among the non-
pharmaceutical intervention measures that can be used to cut the primary source of SARS-CoV2 droplets
expelled by an infected individual. Regardless of discourse on medical resources and diversities in masks,
all countries are mandating coverings over the nose and mouth in public. To contribute towards
communal health, this paper aims to devise a highly accurate and real-time technique that can efficiently
detect non-mask faces in public and thus, enforcing to wear mask. The proposed technique is ensemble
of one-stage and two-stage detectors to achieve low inference time and high accuracy. We start with
ResNet50 as a baseline and applied the concept of transfer learning to fuse high-level semantic
information in multiple feature maps.

1.3 Intended Audience and Reading

Suggestions
The 209th report of the world health organization (WHO) published on 16th August 2020 reported that
coronavirus disease (COVID-19) caused by acute respiratory syndrome (SARS-CoV2) has globally infected
more than 6 Million people and caused over 379,941 deaths worldwide [1]. According to Carissa F.
Etienne, Director, Pan American Health Organization (PAHO), the key to control COVID-19 pandemic is to
maintain social distancing, improving surveillance and strengthening health systems [2]. Recently, a study
on understanding measures to tackle COVID-19 pandemic carried by the researchers at the University of
Edinburgh reveals that wearing a face mask or other covering over the nose and mouth cuts the risk of
Coronavirus spread by avoiding forward distance travelled by a person’s exhaled breath by more than
90% [3]. Steffen et al. also carried an exhaustive study to compute the community-wide impact of mask
use in general public, 

1.4 Product Scope

Our technique achieves outstanding accuracy of 98.2%. The major contribution of the proposed
work is given below:
Develop a novel object detection method that combines one-stage and two-stage detectors for
accurately detecting the object in real-time from video streams with transfer learning at the back
end.
mproved 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.
Creation of unbiased facemask dataset with imbalance ratio equals to nearly one.
The proposed model requires less memory, making it easily deployable for embedded devices
used for surveillance purposes.

2. Overall Description
The proposed model is based on the object recognition benchmark .According to this benchmark,
all the tasks related to an object recognition problem can be ensembled under three main
components: Backbone, Neck and Head as depicted in .Here, the backbone corresponds to a
baseline convolutional neural network capable of extracting information from images and
converting them to a feature map. In the proposed architecture, the concept of transfer learning is
applied on the backbone to utilize already learned attributes of a powerful pre-trained
convolutional neural network in extracting new features for the model.

2.1 Product Perspective

After detecting faces with masks and non-mask in the search proposal, the non-mask faces are passed
separately into a neural network for further exploration of a person’s identity for being violating the
facemask norm. The step requires a fixed-sized input. One possible way of getting a fixed-size input is to
reshape the face in the bounding box to 96 × 96 pixels. The potential issue with this solution is that the
face could be looking in a different direction. Affine transformation can handle this issue very easily

2.2 Product Functions

Effective strategies to restrain COVID-19 pandemic need high attention to mitigate negatively impacted
communal health and global economy, with the brim-full horizon yet to unfold. In the absence of
effective antiviral and limited medical resources, many measures are recommended by WHO to control
the infection rate and avoid exhausting the limited medical resources. Wearing a mask is among the non-
pharmaceutical intervention measures that can be used to cut the primary source of SARS-CoV2 droplets
expelled by an infected individual. Regardless of discourse on medical resources and diversities in masks,
all countries are mandating coverings over the nose and mouth in public. To contribute towards
communal health, this paper aims to devise a highly accurate and real-time technique that can efficiently
detect non-mask faces in public and thus, enforcing to wear mask. The proposed technique is ensemble
of one-stage and two-stage detectors to achieve low inference time and high accuracy. We start with
ResNet50 as a baseline and applied the concept of transfer learning to fuse high-level semantic
information in multiple feature maps. In addition, we also propose a bounding box transformation to
improve localization performance during mask detection.

2.3 Operating Environment

We can operate this function in command prompt at intial level and update this into any website or in
any app .to use it for any useful purpose .

2.4 Design and Implementation

Constraints
Face Recognition systems are an identification procedure in which a person is verified based on human
traits. This paper describes a fast face detection algorithm with accurate result. Lip Tracking is one of the
biometric systems based on which a genuine system can be developed. Since the uttering characteristics
of an individual are unique and difficult to imitate, lip tracking holds an advantage of making the system
secure. We use pre-recorded visual utterance of speakers has been generated and stored in the database
for future verification. The entire project occurs in four different stages in which the first stage includes
obtaining face region from the original image, the second stage includes mouth region extraction by
background subtraction, the third stage includes key points extraction by considering the lip centroid as
origin of co-ordinates and the fourth stage includes storing the obtained feature vector in the database.
The user who wants to be identified by the system provides the new live information, which is then
compared with the existing template in the database. The feedback provided by the system will be ‗a
match or a miss-match'. This project will increase the accuracy level of biometric systems.

3. External Interface Requirements

3.1 User Interfaces
Face Mask Detection Platform uses Artificial Network to recognize if a user is not wearing a mask. The
app can be connected to any existing or new IP mask detection cameras to detect people without a
mask. App users can also add faces and phone numbers to send them an alert in case they are not
wearing a mask.

3.2 Software Interfaces

The face mask recognition system uses AI technology to detect the person with or without a mask. It can
be connected with any surveillance system installed at your premise. The authorities or admin can check
the person through the system to confirm their identity

4 Graph Visualization

5. Other Nonfunctional
Requirements
5.1Performance Requirements
It should scan facial image and diagoned that is is a face and diagoned valley
point .that gives a information of other parts of face if it is a face
then it will search for mouth and gives us a percentage ratio of finding mouth or
not and if percentage is less then 70 percent it will give a red signal and above
that iyt will give a green and verify that a pearson is wearing a mask .
5.2Software Quality Attributes
Function is connected through a well trained data set of both masked and
unmasked photos that gives you a information that the subject is wearing mask or
not.
References
1."Coronavirus disease 2019 (covid-19): situation report 205", 2020.

2."Coronavirus Disease 2019 (COVID-19) – Symptoms", Centers for Disease Control and

Prevention, 2020, [online] Available: https://www.cdc.gov/coronavirus/2019-ncov/symptoms-
testing/symptoms.html

.3."Coronavirus — Human Coronavirus Types — CDC", 2020, [online] Available: Cdc.gov.

4 M. Jiang, X. Fan and H. Yan, "RetinaMask: A Face Mask detector", 2020, [online]
Available: arXiv.org.

5"Guide to the Sequential model - Keras Documentation", 2020, [online] Available:

Faroit.com.

6"Keras documentation: MaxPooling2D layer", 2020, [online] Available: Keras.io

7"OpenCV", 2020, [online] Available: Opencv.org.