DEPARTMENT OF INFORMATIONTECHNOLOGY

SVPM’SCOLLEGE OF ENGINEERING.
MALEGAON(BK)

A PROJECTPRESENTATION ON

SOCIAL DISTANCE MONITORING AND FACE MASK DETECTION

USING MACHINE LEARNING

Name Of the Students Guide Name

MISS. AMRUTA POPAT BHISE Prof. J.N. Mali
MISS.VISHAKHA SHANKAR KOKARE
MISS. SHARADA HANUMANT WAGH
Content
• Introduction
• Software and Hardware Requirements
• Proposed System
• Scope of project
• Existing System
• Literature survey
• System Architecture
• UML Diagrams
• Methodology
• Working
• Application
• conclusion
Introduction
• A novel virus has caused a world pandemic and huge life losses. Declared by
the World Health Organization (WHO), this coronavirus originated from Wuhan,
China in late December 2019. Upon thorough research, the virus has been observed
as pathogenic and transmissible by air or by coming in close contact with an infected
person.
• To avoid the spread of this virus, many measures have been suggested, such
as maintaining a social distance, that is, maintaining a proper physical distance
between people and lessening close contact with each other, and wearing a face
mask to avoid the droplets from transmitting through the air. Therefore, this
project focuses and aims its study towards implementing a Social Distancing and
Face Mask Detection System.
• This system will implement object detection and facial recognition in the video Pretrained models such as the
YOLO are used. People violating the minimum distance were detected as well as faces without face-masks
were detected. An overall results board is displayed in the output containing the number of people violating
or non-violating the respective measures. After implementing and deploying the models, this research project
achieved a confidence core of 100 percent Therefore, this research project concludes with proven facts that
social distancing and wearing face masks helps reduce the spread of the virus and thus builds a model to help
detect these measures.
Software Requirements
1. Python
2. Opencv
3. Numpy
Hardware Requirements

1. Processor-i3 Processor minimum

2. Ram-2 GB minimum
3. Free Space-1 GB Free Disk Space minimum
Technology Used
• OpenCV - Video capturing , Frame processing
• Python – Development
• YOLO- is extremely fast, easy to train, robust, stable and gives promising results even
for tiny objects, hence, we selected it as our object detector of choice. For an input
image/frame, it detects objects belonging to three classes unmasked faces, masked
faces and people
Proposed System
• The proposed system focuses on how to identify the person on LIVE camera whether the
social distancing is maintained or not with the help of computer vision and deep learning
algorithm by using the Open-CV, Tensor flow library.
• The Proposed system focuses on how to identify the person on LIVE camera whether the
face mask is wear or not with the help of deep learning algorithm.
Scope of the project
• The emerging trends and the availability of intelligent technologies make us to develop
new models that help to satisfy the needs of emerging world. This system works very
effectively and efficiently in identifying the social distancing between the people and
generating the alert that can be handled and monitored
Existing System

• Name of System:Real-time Face Mask and Social Distancing Violation Detection System using YOLO
Developer:Krisha Bhambani, Tanmay Jain, Dr. Kavita A. Sultanpure
Explanation: Focuses on a solution to help enforce proper social distancing and wearing masks in public
using YOLO object detection on video footage and images in real time.

• Name of System: Social Distancing and Face Mask Detection From CCTV Camera
Developer: Vishwesh M S, Nikhil D K, Savita S
Explanation: System will detect face masks in photos/images and in real-time videos.

• Name of System: COVID-19 Face Mask and Social Distancing Detector using Machine Learning
Developer: Bhavik Limbasiya, Chinmay Raut
Explanation: Focuses on detect people without Mask and violence of social distancing.
 Literature survey
Author
Krisha Bham-bani, Tanmay Jain,
Dr. KavitaA. Sultanpure
Mohammad Marufur Rah-man,
Jong-Hoon Kim, Saifuddin
Mahmud,Md.Milon Islam,Md.
MotalebHossen Manik
Vishwesh M S, Nikhil D K, Savita Social Distancing and Face Mask
S Detection from CCTV Camera
Bhavik Lim- basiya, Chinmay
Rau
Rujula Singh R, Nikhil Nayak,
Sahana Srini-vasan, Ruchita
R Biradar
Title
Real-time Face Mask andSocial
Distancing Violation Detection System
using YOLO
An Automated System to limit COVID-
19 Using Facial Mask Detection inSmart monitoring the social distancing and safety
City Network
COVID-19 Face Mask and Social
Distancing Detector using
MachineLearning
COVID-19 Monitoring System using
Social Distancing and Face Mask
Detection on Surveillance video
datasets
Summary
Focuses on a solution to help enforce proper
social distancing and wearing masks in public
using YOLO object detection on video footage
and images in real time.
Focuses on human detection framework, for
misuse in the pandemic situation
System will detect face masks in photos/images
and in realtime videos
Focuses on detect people without Mask and
violence of social distancing
Focuses on person detection, social distancing
violation detection, face detection and face mask
classification using object detection, clustering
and Convolution Neural Network based
binary classifier.
 Methodology
In our study, we propose a solution which performs real time detection of individuals to track social distancing norms being
followed and real-time face detection to track usage of face-masks, in several setups, including complex setups which are
crowded and poorly lit. The techniques we used to formulate this solution have been described in this section
A. Dataset
The bounding box coordinates and labels were then extracted from xml files for each image and normalized with respect to
the height and width of the image
B. YOLO
resulting in significant improvements in both speed and accuracy.
YOLO is extremely fast, easy to train, robust, stable and gives promising results even for tiny objects, hence, we selected it
as our object detector of choice.
For an input image/frame, it detects objects belonging to three classes unmasked faces, masked faces and people
C. Tracking Social Distancing
The advantage of this is that authorities can actually select the social distance they want maintained, according to the
specific guidelines.
The social distance between 2 people is solely judged relative to the initial calibration, and the absolute distance need not
be provided to the model.
This Social Distancing model uses the principle that a camera lens is essentially a convex lens, where the image is essentially
captured on a screen.
SYSTEM ARCHITECTURE
Working
1. Capture the video.
2. Read the video by dividing it into a number of frames.
3. Else, detect persons in each frame and get the bounding boxes around them with the
help of YOLO.
4. Further, get the positions of the people to detect where the clusters are forming.
5. While detecting persons, detect their faces to detect whether they have masks on or
not.
6. With the help of bounding boxes on the person and their faces, measure the distance
between them and detect masks on them
7. Create a results board on top of the video and display the results.
8. Create an output stream and then show the results.
9. Do this for every frame till it reaches to end of file.
 User Scenarios
Face Detection and Mask Detection
• detect persons in each frame and get the bounding boxes around them with the help of
YOLO.
• While detecting persons, detect their faces to detect whether they have masks on or
not.

• With the help of bounding boxes on the person and their faces, measure the distance
between them and detect masks on them.
Create a results board on top of the video and display the results. Create an output
stream and then show the results
SOCIAL DISTANCE DETECTION
A. Person Detection

YOLO model was used for person detection t acts as a powerful feature extractor
and an additional 53 layers for detection giving a total of 106 layered fully convolutional neural network depicts the YOLOv3
architecture. Anchor box with3 scales: 13x13, 26x26 and 52x52 are used. These three anchor boxes are used to predict the presence of
a person as shown in figure. The output of this model after
prediction is a list of bounding boxes along with the confidence of the person class
detected.

B. Distance Computing

Euclidean distance algorithm was used to check if social distancing is maintained between the persons detected. The model will take
the video frame as the input and output a list of coordinates in a bounding box in a rectangular shape across each and every person
detected in the frame. The rectangular bounding box is
represented. Each and every person in the video frame will have a centroid for the resulting bounding box. By calculating the distance
between two centroids, the model is going to calculate the distance between two people. The calculation of distance between two
centroids is done using the Euclidean Distance formula. Euclidean distance is generally used to calculate the distance between any two
real valued vectors
System FLOW
Application
• This Project can help in monitoring the social distance and also it detects face mask
•People violating social distancing or not wearing masks were detected.
•This Project can help to reduce spread of Corona Virus.
Disadvantages
• YOLO tend to be less accurate than two-stage detectors (R-CNN) but are significantly
faster
OUTPUT:-
Output 2:-
conclusion
• In this project we have used a recent techniques in the field of computer vision.
• The proposed system will correctly detect the presence of face mask and person is in
the safe the distance.
• The system is accurate, since we have used the architecture for detecting face mask and
for distance computing we used Euclidean distance formula. Thus, it makes easier to
deploy our model. We believe that this approach will enlarge the safety of the
individuals during the pandemic.
References
[1] https://keras.io/api/applications/mobilenet/mobilenetv2-function
[2]https://www.geeksforgeeks.org/face-mask-detection-and-thermal scanner for covid-
care-python-project/
[3] https://www.ijert.org/social-distancing and-face mask detection from cctv camera
[4] World Health Organization (WHO) ”Modes of transmission of virus
causing COVID-19: implications for IPC precaution recommendations
”https://www.who.int/news-room/commentaries/detail/modes-
of-transmissionof-virus-causing-covid-19-implications-for precaution
recommendations.