Social Distance Detector

Abstract
This project addresses the urgent need for effective social distancing enforcement in the wake of the
COVID-19 pandemic through the development of a real-time social distance detection system using
deep learning. Leveraging the YOLO (You Only Look Once) object detection algorithm, the system
accurately identifies and locates individuals within a video feed. By calculating the Euclidean
distance between the centroids of the detected bounding boxes, the system assesses whether
individuals are maintaining a safe distance from each other as per predefined thresholds. To ensure
precise real-world distance measurements, the system incorporates camera calibration techniques
that account for perspective distortions. The model is fine-tuned using a custom dataset to enhance
detection accuracy in specific environments, achieving high performance in both controlled and real-
world settings. The implementation includes optimizations for real-time processing, enabling
deployment on various hardware platforms, including edge devices. Visual and auditory alerts are
generated when social distancing violations are detected, providing an immediate response
mechanism for monitoring compliance in public spaces, workplaces, and other critical areas. This
project demonstrates the potential of combining deep learning and computer vision technologies to
create scalable and efficient solutions for public health challenges, with future enhancements aimed
at improving accuracy and adaptability in more complex environments.

Dept. Of Computer Science, St Philomena’s College pg. 1

 Social Distance Detector

TABLE OF CONTENT
CHAPTER TITLE PAGE
NO.

1 INTRODUCTION……………………………………… 03
1.1 Problem Statement

1.2 Aim

1.3 Existing System

1.4 Proposed System

1.5 Objectives

2 LITERATURE SURVEY…………………………………… 06
2.1 Object Detection

3 SYSTEM REQURIMENT
SPECIFICATION……………………………………………. 08
3.1 Purpose

3.2 Software Architecture

3.3 Acronyms and Abbreviation

 Python
 Flask
 Anaconda
 Numpy
 Pandas
3.3 Overview

4 SYSTEM DESIGN………………………………………………….. 17

4.1 System Architecture

4.2 Data Flow Diagram
4.3 Use Case Diagram

Dept. Of Computer Science, St Philomena’s College pg. 2

Social Distance Detector

5 IMPLEMENTATION…………………………… 24

5.1 System Requirements

6 TESTING………………………………………… 53
6.1Objectives of testing

6.2 Testing Methods

6.3 Test Reports

7 CONCLUSIONS……………………………………………… 60

8 FUTURE ENHANCEMENT……………………… 61

9 REFERENCE…………………………………………… 62

Dept. Of Computer Science, St Philomena’s College pg. 3

Social Distance Detector

SOCIAL DISTANCE DETECTOR

Dept. Of Computer Science, St Philomena’s College pg. 4

 Social Distance Detector

SOCIAL DISTANCE DETECTOR

CHAPTER 1:
INTRODUCTION

INTRODUCTION:

In the wake of the global COVID-19 pandemic, the necessity of effective social distancing measures
has become paramount in curbing the spread of the virus. As societies worldwide grapple with the
challenges of maintaining public health while reopening economies and resuming daily activities,
technology emerges as a crucial ally in enforcing social distancing protocols. This project introduces
a cutting-edge approach to address this pressing need through the utilization of deep learning and
computer vision techniques. By harnessing the power of the YOLO (You Only Look Once) object
detection algorithm, this system aims to provide real-time monitoring and enforcement of social
distancing guidelines in various settings. The introduction of such a system is particularly timely
and relevant, as governments, businesses, and communities seek innovative solutions to mitigate the
risks associated with gatherings and interactions in public spaces. This project endeavors to
contribute to the broader effort of safeguarding public health by empowering stakeholders with the
tools necessary to monitor and enforce social distancing effectively. Through a combination of
advanced technology and practical implementation, this project aims to offer a scalable and
adaptable solution that can be deployed across diverse environments, from retail stores and public
transit systems to schools and workplaces. As the world continues to navigate the complexities of
the ongoing pandemic, the development of robust and reliable social distance detection systems
represents a critical step towards ensuring the safety and well-being of individuals and communities
alike.

1.1 PROBLEM STATEMENT:

The COVID-19 pandemic has underscored the critical importance of maintaining social distancing
to mitigate the spread of infectious diseases. However, enforcing social distancing protocols in
public spaces poses significant challenges due to the need for constant monitoring and intervention.
Traditional methods of manual enforcement are labor-intensive, prone to human error, and often

Dept. Of Computer Science, St Philomena’s College pg. 5

 Social Distance Detector

ineffective in ensuring compliance on a large scale. Thus, there is a pressing need for automated
systems capable of accurately detecting and alerting authorities to instances of social distancing
violations in real-time.

1.2 AIM:

The aim of this project is to develop an automated social distance detection system using deep
learning and computer vision techniques. By leveraging state-of-the-art object detection algorithms
such as YOLO (You Only Look Once), the system aims to provide robust and reliable monitoring
of social distancing compliance in various environments. The ultimate goal is to empower
authorities, businesses, and individuals with a scalable and effective tool for enforcing social
distancing protocols and reducing the risk of disease transmission.

1.3 EXISTING SYSTEM:

The existing systems for monitoring social distancing often rely on manual enforcement or
rudimentary technologies that lack the accuracy and scalability required for widespread deployment.
Manual enforcement is resource-intensive and impractical for continuous monitoring of large
crowds or public spaces. Some existing automated systems utilize basic computer vision algorithms
for detecting people and measuring distances, but they often from limitations in accuracy, speed,
and robustness. There is a clear need for more advanced and reliable solutions capable of
overcoming these challenges.

1.4 PROPOSED SYSTEM:

The proposed system will employ the YOLO object detection algorithm to accurately detect and
localize individuals within a video feed in real-time. By analyzing the spatial relationships between
detected individuals, the system will calculate the distances between them and compare these
distances to predefined thresholds to determine social distancing compliance. The system will be
designed to provide visual and auditory alerts when violations are detected, enabling authorities to
take immediate corrective action. Additionally, the system will incorporate optimizations for real-
time performance and scalability, making it suitable for deployment in diverse environments and
settings.

Dept. Of Computer Science, St Philomena’s College pg. 6

 Social Distance Detector

1.5
Objectives:

 Develop a custom dataset for training the YOLO object detection model, including annotated
images of individuals in various social distancing scenarios.
 Fine-tune the YOLO model using the custom dataset to enhance detection accuracy and
robustness in real-world environments.
 Implement real-time social distance detection algorithms that accurately measure the distances
between detected individuals and identify violations of predefined social distancing thresholds.
 Integrate visual and auditory alert mechanisms into the system to notify authorities and
individuals of social distancing violations in real-time.
 Evaluate the performance of the system using metrics such as detection accuracy, false positive
rate, and real-time processing speed, and validate its effectiveness in diverse environments and
scenarios.

Dept. Of Computer Science, St Philomena’s College pg. 7

 Social Distance Detector

CHAPTER 2

LITERATURE REVIEW

This recently imposed restriction is widely, but imprecisely, referred to as ―social distancing‖ (SD)
since prevention of the virus diffusion does not require us to weaken our social bonds. The likely
reason of SD naming is that, from a cognitive point of view, physical and social aspects of distance
are deeply intertwined [47], a phenomenon that popular wisdom captures through a proverb that, in
slightly different versions, appears in different languages and cultures, namely ―far from eyes, far
from heart‖. Not surprisingly, the time spent in physical proximity with others, in opposition to the
time spent in individual activities, is a crucial factor in the ―social brain hypothesis‖, one of the most
successful theories of human evolution [26]. Similarly, Attachment Theory, probably the
development model most widely accepted in child psychiatry, revolves around the ability of children
and parents to establish and maintain physical proximity [13]. Finally, the different modulation of
interpersonal distances is known to be one of the main obstacles in intercultural communication
[37]. The above suggests that dealing with interpersonal distances means to deal with evolutionary,
developmental and cultural forces that shape, to a significant extent, our everyday life. As a
consequence, the role of technologies for the analysis of such distances becomes crucial during
pandemics, given that they must mediate between the forces above, responsible for the human
tendency to get too close to avoid contagion, and the pressure of prophylactic measures, artificially
designed to fight a pathogen inaccessible to our senses and cognition. One possible solution is to go
beyond simply measuring how far we are are from one another, as most of the applications on the
market are doing (see Sec. II-C) and try to make sense of what distances mean. In other words, to
inform technologies with principles and laws of Proxemics, the psychology area showing how
people convey social meaning through interpersonal distances and, ultimately, how social and
physical dimensions of space interplay with one another [74]. Proxemics is strictly linked to the
definition of people gath erings, namely groups, and as such, it depends on its spatial organization
and the number of people involved. In general, the surrounding space around a person is
characterized by interpersonal distance classes [38], namely: intimate, personal, peri-personal or
social, and public spaces (see Fig. 3), all associated to different SDs, in turn, also dependent by the
degree of kinship and familiarity between the subjects and by the geometrical configuration and size
of the environment in which an interplay occurs. A blind application of social distancing rules,

Dept. Of Computer Science, St Philomena’s College pg. 8

 Social Distance Detector

encouraging to stay further than 1-2 meters, will eliminate an entire interpersonal distance class and
all of the social interactions which take play within it, including for example those between children
and relatives. As can be noticed, behavior, social interactions, and space arrangements are tightly
coupled, and affect each other. This is why it is important to take into considerations all these aspects
when constraints in this respect are to be imposed, in particular when people health is in play.

2.1 OBJECT DETECTION?

Object Detection is the way toward discovering true item cases like vehicle, bicycle, TV, blossoms,
and people in still pictures or Videos. It takes into consideration the acknowledgment, confinement,
and identification of numerous articles inside a picture which gives us a greatly improved
comprehension of a picture overall. It is regularly utilized in applications, for example, picture
recovery, security, observation, and progressed driver help frameworks

Dept. Of Computer Science, St Philomena’s College pg. 9

 Social Distance Detector

CHAPTER-3
SOFTWARE REQUIREMENT SPECIFICATION:

The presentation of the Software Requirements Specification (SRS) gives a review of the whole
SRS with reason, scope, definitions, abbreviations, contractions, references and diagram of the SRS.
The point of this report is to assemble, dissect, and give a top to bottom knowledge of the total
"Social distance detection" by characterizing the difficult articulation in detail. The point by point
necessities of the Indian car purchasing conduct – client related capacities are given in this archive.

3.1 PURPOSE

The Purpose of the Software Requirements Specification is to give the specialized, Functional and
non-useful highlights, needed to build up a web application App. The whole application intended to
give client adaptability to finding the briefest as well as efficient way. To put it plainly, the
motivation behind this SRS record is to give an itemized outline of our product item, its boundaries
and objectives. This archive depicts the task's intended interest group and its UI, equipment and
programming prerequisites. It characterizes how our customer, group and crowd see the item and its
usefulness.

Scope

The Scope of this framework is to presents a survey on information digging strategies utilized for
the expectation of Social distance detection. It is obvious from the framework that information
mining strategy, similar to grouping, is profoundly productive in expectation of Indian car.

Dept. Of Computer Science, St Philomena’s College pg. 10

 Social Distance Detector

3.2 SOFTWARE ARCHITECTURE:

3.3 ACRONYMS AND ABBREVIATION:

PYTHON:

Python is a deciphered, significant level, broadly useful programming language. Made by Guido van
Rossum and first delivered in 1991, Python's plan reasoning accentuates code meaningfulness with
its prominent utilization of critical whitespace. Its language develops and object-arranged
methodology plan to assist software engineers with composing clear, consistent code for little and
huge scope ventures.

Python is progressively composed and trash gathered. It underpins numerous programming

standards, including procedural, object-arranged, and practical programming. Python is frequently

Dept. Of Computer Science, St Philomena’s College pg. 11

 Social Distance Detector

portrayed as a "batteries included" language because of its thorough standard library. Python is a
multi-worldview programming language. Article arranged programming and organized writing
computer programs are completely upheld, and a significant number of highlights uphold useful
programming and angle situated programming (counting by metaprogramming and metaobjects
(enchantment methods)).] Many different standards are upheld by means of expansions, including
plan by agreement and rationale programming.

FLASK:

Flask is a miniature web system written in Python. It is delegated a microframework in light of the
fact that it doesn't need specific apparatuses or libraries.[3] It has no information base deliberation
layer, structure approval, or whatever other segments where prior outsider libraries give normal
capacities. In any case, Flask upholds augmentations that can include application includes as though
they were executed in Flask itself. Augmentations exist for object-social mappers, structure
approval, transfer dealing with, different open confirmation advancements and a few basic system
related devices. Augmentations are refreshed unmistakably more as often as possible than the center
Flask program.

ANACONDA:

Anaconda is a free and open-source circulation of the programming dialects Python and R . The
dissemination accompanies the Python translator and different bundles identified with AI and
information science.

Essentially, the thought behind Anaconda is to make it simple for individuals inspired by those fields
to introduce all (or a large portion) of the bundles required with a solitary establishment.

An open-source bundle and condition the executives framework called Conda, which makes it
simple to introduce/update bundles and make/load situations.

AI libraries like TensorFlow, scikit-learn and Theano.

Information science libraries like pandas, NumPy and Dask.

Dept. Of Computer Science, St Philomena’s College pg. 12

 Social Distance Detector

Perception libraries like Bokeh, Datashader, matplotlib and Holoviews.

Jupyter Notebook, a shareable note pad that joins live code, representations and text.

NUMPY:

NumPy is the principal bundle for logical registering with Python. It contains in addition to other
things:

• a amazing N-dimensional cluster object

• sophisticated (broadcasting) capacities

• tools for incorporating C/C++ and Fortran code

• useful straight polynomial math, Fourier change, and arbitrary number abilities

PANDAS:

pandas is an open source, BSD-authorized library giving elite, simple to-utilize information
structures and information investigation apparatuses for the Python programming language.

pandas is a Num FOCUS supported undertaking. This will help guarantee the achievement of
improvement of pandas as an a-list open-source venture, and makes it conceivable to give to the
task.

YOLO

With regards to deep learning-based item discovery, there are three essential article locators you'll
experience:

• R-CNN and their variations, including the first R-CNN, Fast R-CNN, and Faster R- CNN

• Single Shot Detector (SSDs)

• YOLO

Dept. Of Computer Science, St Philomena’s College pg. 13

 Social Distance Detector

R-CNNs are one of the primary deep learning-based article locators and are a case of a two-phase
identifier.

1. In the principal R-CNN distribution, Rich element progressive systems for exact item
recognition and semantic division, (2013) Girshick et al. proposed an item finder that necessary a
calculation, for example, Selective Search (or proportionate) to propose applicant jumping boxes
that could contain objects.

2. These districts were then passed into a CNN for characterization, at last prompting one of the
main deep learning-based item locators.

The issue with the standard R-CNN strategy was that it was horrendously moderate and not a total
start to finish object finder.

Girshick et al. distributed a second paper in 2015, entitled Fast R-CNN. The Fast R-CNN calculation
made significant enhancements to the first R-CNN, specifically expanding exactness and decreasing
the time it took to play out a forward pass; notwithstanding, the model actually depended on an outer
area proposition calculation.

It wasn't until Girshick et al's. subsequent 2015 paper, Faster R-CNN: Towards Real-Time Object
Detection with Region Proposal Networks, that R-CNNs turned into a genuine start to finish deep
learning object locator by eliminating the Selective Search prerequisite and rather depending on a
Region Proposal Network (RPN) that is (1) completely convolutional and (2) can foresee the article
bouncing boxes and "objectness" scores (i.e., a score measuring how likely it is a locale of a picture
may contain a picture). The yields of the RPNs are then passed into the R-CNN part for conclusive
grouping and naming.

While R-CNNs watch out for extremely exact, the most serious issue with the R-CNN group of
organizations is their speed — they were unimaginably moderate, getting just 5 FPS on a GPU.

To help speed up deep learning-based article finders, both Single Shot Detectors (SSDs) and YOLO
utilize a one-phase identifier procedure.

Dept. Of Computer Science, St Philomena’s College pg. 14

 Social Distance Detector

These calculations treat object location as a relapse issue, taking a given info picture and all the
while picking up jumping box arranges and comparing class name probabilities.

When all is said in done, single-stage indicators will in general be less exact than two-phase locators
however are fundamentally quicker.

Feasibility Study

The feasibility study helps to find solutions to the problems of the project. The solution is given how
looks like a new system look like.

Technical Feasibility

The project entitled ―social distance detection‖ is technically feasible because of the below-
mentioned features. The project is developed in Python. The web server is used to develop ―social
distance detection‖ is local serve. The local server very neatly coordinates between the design and
coding parts. It provides a Graphical User Interface to design an application while the coding is done
in python. At the same time, it provides high-level reliability, availability, and compatibility.

Economic Feasibility

In economic feasibility, cost-benefit analysis is done in which costs and benefits are evaluated.
Economic analysis is used for the effectiveness of the proposed system. In economic feasibility, the
main task is cost-benefit analysis. The system ―social distance detection‖ is feasible because it does
not exceed the estimated cost and the estimated benefits are equal.

Operational Feasibility

The project entitled ―social distance detection‖ is technically feasible because of the below-
mentioned features. The system predicts the social distance detection based on the historical data,
further the details of the patient are added to the DataBase. The performance of the Data mining
techniques are compared based on their execution time and displayed it through a graph.

Behavior Feasibility

The project entitled ―social distance detection‖ is beneficial because it satisfies the objectives when
developed and installed.

Dept. Of Computer Science, St Philomena’s College pg. 15

 Social Distance Detector

3.4 OVERVIEW
Following a section of this document will focus on describing the system in terms of product
functions In the next section, we will address specific requirements of the system, which will
enclose functional requirements and non-functional requirements.

Dept. Of Computer Science, St Philomena’s College pg. 16

 Social Distance Detector

Product Functions

 Pre-Processing

 video has been live detected.

 Frames are extracted from video.

 Each frames are converted from RGB to Gray conversion

 Objects are detected in images and feature extraction

 The extracted and google trained datasets are compared using TensorFlow

 The compared data is predicted using CNN

General Constraints

The results generated have to be entered in to the system and any error or any value entered out of
the boundary will not be understood by the system. In any case if the database crashes, the whole
information collected and the results generated will be of no use.

Specific Requirements

External Interface Requirements

This section provides a detailed description of all inputs into and outputs from the system. It also
gives a description of the hardware, software and communication interfaces and provides basic
prototypes of the user interface

Dept. Of Computer Science, St Philomena’s College pg. 17

 Social Distance Detector

Hardware Requirements

Processor : Intel i3 3.30 GHz.

Hard Disk : 40 GB (min)

Ram : 4GB

Software Requirements

Operating system : Windows 7 and above.

Coding Language : Python.

Framework: Flask

Dept. Of Computer Science, St Philomena’s College pg. 18

 Social Distance Detector

CHAPTER 4
SYSTEM DESIGN

Introduction to Design document

The Software Design will be used to aid in software development for android application by
providing the details for how the application should be built. Within the Software Design,
specifications are narrative and graphical documentation of the software design for the project
includes use case models, sequence diagrams, and other supporting requirement information.

Scope
The Design Document is for a primary level system, which will work as a basement for building a
system that provides a base level of functionality to show feasibility for large-scale production use.
The software Design Document, the focus placed on the generation and modification of the
documents. The system will be used in conjunction with other pre-existing systems and will consist
largely of a document interaction faced that abstracts document interactions and handling of the
document objects. This Document provides the Design specifications of ―Video object detection‖.
4.1 SYSTEM ARCHIETECTURE:

Dept. Of Computer Science, St Philomena’s College pg. 19

 Social Distance Detector

Dept. Of Computer Science, St Philomena’s College pg. 20

 Social Distance Detector

4.2 DATA FLOW DIAGRAM

LEVEL 0 DFD:

Fig 4.2.1: LEVEL 0 DFD

LEVEL 1 DFD:

Fig 4.2.2: LEVEL 1 DFD

Dept. Of Computer Science, St Philomena’s College pg. 21

 Social Distance Detector

4.3 ACTIVITY DIAGRAM

An activivity diagram outwardly presents a progression of activities or stream of

control in a framework like a flowchart or an information stream chart. Action graphs
are regularly utilized in business measure demonstrating. They can likewise depict the
means in an utilization case chart. Exercises demonstrated can be consecutive and
simultaneous. In the two cases, an action outline will have a start (an underlying state)
and an end (a last state).

initial state) and an end (a final state).

Dept. Of Computer Science, St Philomena’s College pg. 22

 Social Distance Detector

Fig 4.3.1: ACTIVITY DIAGRAM

Dept. Of Computer Science, St Philomena’s College pg. 23

 Social Distance Detector

4.4 SEQUENCE DIAGRAM

sequence diagram depict cooperations among classes as far as a trade of messages after some time.
They're likewise called occasion charts. A grouping chart is a decent method to envision and approve
different runtime situations. These can assist with anticipating how a framework will act and to find
duties a class may need to have during the time spent demonstrating another framework.

Fig 4.4.1: SEQUENTIAL DIAGRAM

Dept. Of Computer Science, St Philomena’s College pg. 24

 Social Distance Detector

4.5 USE CASE DIAGRAM

The motivation behind use case diagram is to catch the dynamic part of a framework. In any case,
this definition is too nonexclusive to even think about describing the reason, as other four outlines
(action, grouping, cooperation, and Statechart) likewise have a similar reason. We will investigate
some particular reason, which will recognize it from other four charts.

Use case graphs are utilized to accumulate the prerequisites of a framework including inside and
outside impacts. These prerequisites are generally plan necessities. Consequently, when a
framework is investigated to accumulate its functionalities, use cases are readied and entertainers
are distinguished.

Dept. Of Computer Science, St Philomena’s College pg. 25

 Social Distance Detector

CHAPTER-5
IMPLEMENTATION

Introduction

The project is implemented using python which is an object oriented programming language
and procedure oriented programming language. Object oriented programming is an approach
that provides a way of modularizing program by creating partitioned memory area of both data
and function that can be used as a template for creating copies of such module on demand.

This project is implemented using python programming language. Python is dynamically

typed and garbage-collected. It supports multiple programming paradigms,
including procedural, object-oriented, and functional programming. Python is often described
as a "batteries included" language due to its comprehensive standard library. The machine
Learning techniques are used in this project.

Implementation of software refers to the final installation of the package in its real environment,
to the satisfaction of the intended users and the operation of the system. The people are not sure that the
software is meant to make their job easier.

 The active user must be aware of the benefits of using the system
 Their confidence in the software built up
 Proper guidance is impaired to the user so that he is comfortable in using the application

Before going ahead and viewing the system, the user must know that for viewing the result, the server
program should be running in the server. If the server object is not running on the server, the actual processes
will not take place.

User Training

To achieve the objectives and benefits expected from the proposed system it is essential for the people
who will be involved to be confident of their role in the new system. As system becomes more complex, the need
for education and training is more and more important. Education is complementary to training. It brings life to
formal training by explaining the background to the

Dept. Of Computer Science, St Philomena’s College pg. 26

 Social Distance Detector

resources for them. Education involves creating the right atmosphere and motivating user staff.
Education information can make training more interesting and more understandable.

Training on the Application Software

After providing the necessary basic training on the computer awareness, the users will have to be
trained on the new application software. This will give the underlying philosophy of the use of the new system
such as the screen flow, screen design, type of help on the screen, type of errors while entering the data, the
corresponding validation check at each entry and the ways to correct the data entered. This training may be
different across different user groups and across different levels of hierarchy.

Operational Documentation

Once the implementation plan is decided, it is essential that the user of the system is made familiar and
comfortable with the environment. A documentation providing the whole operations of the system is being
developed. Useful tips and guidance is given inside the application itself to the user. The system is developed
user friendly so that the user can work the system from the tips given in the application itself.

System Maintenance

The maintenance phase of the software cycle is the time in which software performs useful work. After
a system is successfully implemented, it should be maintained in a proper manner. System maintenance is an
important aspect in the software development life cycle. The need for system maintenance is to make adaptable
to the changes in the system environment. There may be social, technical and other environmental changes,
which affect a system which is being implemented. Software product enhancements may involve providing new
functional capabilities, improving user displays and mode of interaction, upgrading the performance
characteristics of the system. So only thru proper system maintenance procedures, the system can be adapted
to cope up with these changes. Software maintenance is of course, far more than “finding mistakes”.

Corrective Maintenance

The first maintenance activity occurs because it is unreasonable to assume that software testing will
uncover all latent errors in a large software system. During the use of any large program, errors will occur and
be reported to the developer. The process that includes the diagnosis and correction of one or more errors
is called Corrective Maintenance.

Adaptive Maintenance

The second activity that contributes to a definition of maintenance occurs because of the rapid change

Dept. Of Computer Science, St Philomena’s College pg. 27

 Social Distance Detector

that
is encountered in every aspect of computing. Therefore Adaptive maintenance termed as an activity that modifies
software to properly interfere with a changing environment is both necessary and commonplace.

Perceptive Maintenance

The third activity that may be applied to a definition of maintenance occurs when a software package is
successful. As the software is used, recommendations for new capabilities, modifications to existing functions,
and general enhancement are received from users. To satisfy requests in this category, Perceptive maintenance
is performed. This activity accounts for the majority of all efforts expended on software maintenance.

Preventive Maintenance

The fourth maintenance activity occurs when software is changed to improve future maintainability or
reliability, or to provide a better basis for future enhancements. Often called preventive maintenance, this
activity is characterized by reverse engineering and re-engineering techniques

Machine Learning vs Deep Learning

Experts in machine learning and deep learning have not yet reached consensus on these concepts. in
this context, almost every day new ideas are being discussed. Machine Learning is an older concept
than Deep Learning. Deep learning can also be called a technique that performs machine learning.
The differences are listed below;

1) In deep learning, too much data is needed to bring the algorithm structure to the ideal.In machine
learning, the problem can be solved with much less data because the person gives specific features
to the algorithm.
2) Deep learning algorithms try to extract features from data. In machine learning, the features are
determined by the expert.
3) While Deep Learning algorithms work on high performance machines, Machine Learning
algorithms can work on ordinary CPUs.
4) In machine learning, the problem is usually divided into pieces, these parts are solved one by one
and then the solutions are formed as a result of the solutions. In deep learning, the problem is
solved end-to-end.

Dept. Of Computer Science, St Philomena’s College pg. 28

 Social Distance Detector

5) It takes a long time to train deep learning algorithms.

DeepLearning overview

The term Deep Learning or Deep Neural Network refers to Artificial Neural Networks (ANN)
with multi layers . Over the last few decades, it has been considered to be one of the most
powerful tools, and has become very popular in the literature as it is able to handle a huge
amount of data. The interest in having deeper hidden layers has recently begun to surpass
classical methods performance in different fields; especially in pattern recognition. One of the
most popular deep neural networks is the Convolutional Neural Network (CNN). It take this
name from mathematical linear operation between matrixes called convolution. CNN have
multiple layers; including convolutional layer, non-linearity layer, pooling layer and
fullyconnected layer. The convolutional and fully- connected layers have parameters but
pooling and non-linearity layers don't have parameters. The CNN has an excellent performance
in machine learning problems. Specially the applications that deal with image data, such as
largest image classification data set (Image Net), computer vision, and in natural language
processing (NLP) and the results achieved were very amazing . In this paper we will explain
and define all the elements and important issues related to CNN, and how these elements work.
In addition, we will also state the parameters that effect CNN efficiency. This paper assumes
that the readers have adequate knowledge about both machine learning and artificial neural
network.

Convolutional Neural Network has had ground breaking results over the past decade in a
variety of fields related to pattern recognition; from image processing to voice recognition.
The most beneficial aspect of CNNs is reducing the number of parameters in ANN . This
achievement has prompted both researchers and developers to approach larger models in order
to solve complex tasks, which was not possible with classic ANNs; . The most important
assumption about problems that are solved by CNN should not have features which are
spatially dependent. In other words, for example, in a face detection application, we do not

Dept. Of Computer Science, St Philomena’s College pg. 29

 Social Distance Detector

need to pay attention to where the faces are located in the images. The only concern is to detect
them regardless of their position in the given images . Another important aspect of CNN, is to
obtain abstract features when input propagates toward the deeper layers.

CHALLENGES IN IMPLEMENTING DEEP LEARNING:

Lots and lots of data

Deep learning algorithms are trained to learn progressively using data. Large data sets are needed to
make sure that the machine delivers desired results. As human brain needs a lot of experiences to learn
and deduce information, the analogous artificial neural network requires copious amount of data. The
more powerful abstraction you want, the more parameters need to be tuned and more parameters require
more data.

Overfitting in neural networks

At times, the there is a sharp difference in error occurred in training data set and the error encountered
in a new unseen data set. It occurs in complex models, such as having too many parameters relative to
the number of observations. The efficacy of a model is judged by its ability to perform well on an
unseen data set and not by its performance on the training data fed to it.

Hyperparameter Optimization

Hyperparameters are the parameters whose value is defined prior to the commencement of the learning
process. Changing the value of such parameters by a small amount can invoke a large change in the
performance of your model.

Dept. Of Computer Science, St Philomena’s College pg. 30

 Social Distance Detector

Requires high-performance hardware

Training a data set for a Deep Learning solution requires a lot of data. To perform a task to solve real
world problems, the machine needs to be equipped with adequate processing power. To ensure better
efficiency and less time consumption, data scientists switch to multi-core high performing GPUs and
similar processing units. These processing units are costly and consume a lot of power.

Lack of Flexibility and Multitasking

Deep Learning models, once trained, can deliver tremendously efficient and accurate solution to a
specific problem. However, in the current landscape, the neural network architectures are highly
specialized to specific domains of application

Introduction
This section will investigate the distinctive viewpoints concerned about the
implementation of the developed system. This task was concerned about the
improvement and usage the Video Object Detection.

Dept. Of Computer Science, St Philomena’s College pg. 31

 Social Distance Detector

Methodology

Cnn model

 Module 1: Region Proposal. Generate and extract category independent region proposals, e.g.
candidate bounding boxes.
 Module 2: Feature Extractor. Extract feature from each candidate region, e.g. using a deep
convolutional neural network.
 Module 3: Classifier. Classify features as one of the known class, e.g. linear SVM classifier model.

Convolutional Neural Networks have a different architecture than regular Neural Networks. Regular
Neural Networks transform an input by putting it through a series of hidden layers. Every layer is made
up of a set of neurons, where each layer is fully connected to all neurons in the layer before. Finally,
there is a last fully-connected layer — the output layer — that represent the predictions.

Convolutional Neural Networks are a bit different. First of all, the layers are organised in 3 dimensions:
width, height and depth. Further, the neurons in one layer do not connect to all the neurons in the next
layer but only to a small region of it. Lastly, the final output will be reduced to a single vector of
probability scores, organized along the depth dimension.

Dept. Of Computer Science, St Philomena’s College pg. 32

 Social Distance Detector

Feature Extraction:Convolution

Convolution in CNN is performed on an input image using a filter or a kernel. To understand filtering
and convolution you will have to scan the screen starting from top left to right and moving down a bit
after covering the width of the screen and repeating the same process until you are done scanning the
whole screen.

Feature Extraction: Pooling

After a convolution layer once you get the feature maps, it is common to add a pooling or a sub-sampling
layer in CNN layers. Similar to the Convolutional Layer, the Pooling layer is responsible for reducing
the spatial size of the Convolved Feature. This is to decrease the computational power required to
process the data through dimensionality reduction. Furthermore, it is useful for extracting dominant
features which are rotational and positional invariant, thus maintaining the process of effectively training
of the model. Pooling shortens the training time and controls over-fitting.

Classification — Fully Connected Layer (FC Layer):

Adding a Fully-Connected layer is a (usually) cheap way of learning non-linear combinations of the
high-level features as represented by the output of the convolutional layer. The Fully-Connected layer
is learning a possibly non-linear function in that space. Example of CNN network:

Dept. Of Computer Science, St Philomena’s College pg. 33

 Social Distance Detector

Fully Connected model

Now that we have converted our input image into a suitable form, we shall flatten the image into a
column vector. The flattened output is fed to a feed-forward neural network and backpropagation applied
to every iteration of training. Over a series of epochs, the model is able to distinguish between
dominating and certain low-level features in images and classify them using the Softmax Classification
technique.

So now we have all the pieces required to build a CNN. Convolution, ReLU and Pooling. The output of
max pooling is fed into the classifier we discussed initially which is usually a multi-layer perceptron
layer. Usually in CNNs these layers are used more than once i.e. Convolution ->ReLU -> Max-Pool ->
Convolution ->ReLU -> Max-Pool and so on. We won’t discuss the fully connected layer right now.

Dept. Of Computer Science, St Philomena’s College pg. 34

 Social Distance Detector

Initialize Parameters and Load Data

In this step we initialize the parameters of the convolutional neural network. You will be using 10 filters
of dimension 9x9, and a non-overlapping, contiguous 2x2 pooling region.

Implement CNN Objective

Implement the CNN cost and gradient computation in this step. Your network will have two layers.
The first layer is a convolutional layer followed by mean pooling and the second layer is a densely
connected layer into softmax regression.

Learn Parameters

Using a batch method such as L-BFGS to train a convolutional network of this size even on object
datasets, a relatively small dataset, can be computationally slow. A single iteration of calculating the
cost and gradient for the full training set can take several minutes or more. Thus you will use stochastic
gradient descent (SGD) to learn the parameters of the network.

Test

With the convolution network and SGD optimizer in hand, you are now ready to test the performance
of the model. We’ve provided code at the end of cnnTrain.py to test the accuracy of your networks
predictions on the object test set.

In neural networks, Convolutional neural network (ConvNets or CNNs) is one of the main categories to

do images recognition, images classifications. Objects detections, recognition faces etc., are some of the

areas where CNNs are widely used.

CNN image classifications takes an input image, process it and classify it under certain categories (Eg.,
Dog, Cat, Tiger, Lion). Computers sees an input image as array of pixels and it depends on the image

Dept. Of Computer Science, St Philomena’s College pg. 35

 Social Distance Detector

resolution. Based on the image resolution, it will see h x w x d( h = Height, w = Width, d = Dimension

). Eg., An image of 6 x 6 x 3 array of matrix of RGB (3 refers to RGB values) and an image of 4 x 4 x

1 array of matrix of grayscale image.

Technically, deep learning CNN models to train and test, each input image will pass it through a series

of convolution layers with filters (Kernals), Pooling, fully connected layers (FC) and apply Softmax

function to classify an object with probabilistic values between 0 and 1. The below figure is a

complete flow of CNN to process an input image and classifies the objects based on values.

Convolution Layer

Convolution is the first layer to extract features from an input image. Convolution preserves the
relationship between pixels by learning image features using small squares of input data. It is a

mathematical operation that takes two inputs such as image matrix and a filter or kernel.

Strides

Stride is the number of pixels shifts over the input matrix. When the stride is 1 then we move the filters

to 1 pixel at a time. When the stride is 2 then we move the filters to 2 pixels at a time and so on. The
below figure shows convolution would work with a stride of 2.

Padding

Sometimes filter does not fit perfectly fit the input image. We have two options:

 Pad the picture with zeros (zero-padding) so that it fits

Dept. Of Computer Science, St Philomena’s College pg. 36

 Social Distance Detector

 Drop the part of the image where the filter did not fit. This is called valid padding which keeps only
valid part of the image.

Non Linearity (ReLU)

ReLU stands for Rectified Linear Unit for a non-linear operation. The output is ƒ(x) = max(0,x).Why

ReLU is important : ReLU’s purpose is to introduce non-linearity in our ConvNet. Since, the real
world data would want our ConvNet to learn would be non-negative linear values.There are other

non linear functions such as tanh or sigmoid that can also be used instead of ReLU. Most of the data

scientists use ReLU since performance wise ReLU is better than the other two.

Pooling Layer
Pooling layers section would reduce the number of parameters when the images are too large. Spatial

pooling also called subsampling or downsampling which reduces the dimensionality of each map but
retains important information. Spatial pooling can be of different types:

 Max Pooling

 Average Pooling

 Sum Pooling

Max pooling takes the largest element from the rectified feature map. Taking the largest element
could also take the average pooling. Sum of all elements in the feature map call as sum pooling.

Fully Connected Layer

Dept. Of Computer Science, St Philomena’s College pg. 37

 Social Distance Detector

The
layer we call as FC layer, we flattened our matrix into vector and feed it into a fully connected layer
like a neural network.

 Provide input image into convolution layer

 Choose parameters, apply filters with strides, padding if requires. Perform convolution on the image
and apply ReLU activation to the matrix.

 Perform pooling to reduce dimensionality size

 Add as many convolutional layers until satisfied

 Flatten the output and feed into a fully connected layer (FC Layer)

 Output the class using an activation function (Logistic Regression with cost functions) and classifies
images.

In the next post, I would like to talk about some popular CNN architectures such as AlexNet,
VGGNet, GoogLeNet, and ResNet.

Image classification!

The convolutional neural network (CNN) is a class of deep learning neural networks. CNNs

represent a huge breakthrough in image recognition. They’re most commonly used to analyze visual
imagery and are frequently working behind the scenes in image classification. They can be found at

the core of everything from Facebook’s photo tagging to self-driving cars. They’re working hard

behind the scenes in everything from healthcare to security.

They’re fast and they’re efficient. But how do they work?

Dept. Of Computer Science, St Philomena’s College pg. 38

 Social Distance Detector

Image classification is the process of taking an input (like a picture) and outputting a class (like ―cat‖)

or a probability that the input is a particular class (―there’s a 90% probability that this input is a cat‖).

You can look at a picture and know that you’re looking at a terrible shot of your own face, but how

can a computer learn to do that?

With a convolutional neural network!

A CNN has

 Convolutional layers

 ReLU layers

 Pooling layers

 a Fully connected layer

A classic CNN architecture would look something like this:

Input ->Convolution ->ReLU ->Convolution ->ReLU ->Pooling ->

ReLU ->Convolution ->ReLU ->Pooling ->Fully Connected

A CNN convolves (not convolutes…) learned features with input data and uses 2D convolutional layers.

This means that this type of network is ideal for processing 2D images. Compared to other image

classification algorithms, CNNs actually use very little preprocessing. This means that they

can learn the filters that have to be hand-made in other algorithms. CNNs can be used in tons of

applications from image and video recognition, image classification, and recommender systems to
natural language processing and medical image analysis.

Dept. Of Computer Science, St Philomena’s College pg. 39

 Social Distance Detector

CNNs are inspired by biological processes. They’re based on some cool research done by Hubel and

Wiesel in the 60s regarding vision in cats and monkeys. The pattern of connectivity in a CNN comes

from their research regarding the organization of the visual cortex. In a mammal’s eye, individual
neurons respond to visual stimuli only in the receptive field, which is a restricted region. The receptive

fields of different regions partially overlap so that the entire field of vision is covered. This is the way

that a CNN works!

CNNs have an input layer, and output layer, and hidden layers. The hidden layers usually consist of

convolutional layers, ReLU layers, pooling layers, and fully connected layers.

 Convolutional layers apply a convolution operation to the input. This passes the information on to
the next layer.

 Pooling combines the outputs of clusters of neurons into a single neuron in the next layer.

 Fully connected layers connect every neuron in one layer to every neuron in the next layer.

In a convolutional layer, neurons only receive input from a subarea of the previous layer. In a fully
connected layer, each neuron receives input from every element of the previous layer.

A CNN works by extracting features from images. This eliminates the need for manual feature
extraction. The features are not trained! They’re learned while the network trains on a set of images.

This makes deep learning models extremely accurate for computer vision tasks. CNNs learn feature

detection through tens or hundreds of hidden layers. Each layer increases the complexity of the
learned features.

Dept. Of Computer Science, St Philomena’s College pg. 40

 Social Distance Detector

A
CNN

 starts with an input image

 applies many different filters to it to create a feature map

 applies a ReLU function to increase non-linearity

 applies a pooling layer to each feature map

 flattens the pooled images into one long vector.

 inputs the vector into a fully connected artificial neural network.

 processes the features through the network. The final fully connected layer provides the ―voting‖

of the classes that we’re after.

 trains through forward propagation and backpropagation for many, many epochs. This repeats
until we have a well-defined neural network with trained weights and feature detectors.

 For a black and white image, those pixels are interpreted as a 2D array (for example, 2x2 pixels).

Every pixel has a value between 0 and 255. (Zero is completely black and 255 is completely
white. The greyscale exists between those numbers.) Based on that information, the computer can

begin to work on the data.

For a color image, this is a 3D array with a blue layer, a green layer, and a red layer. Each one of those
colors has its own value between 0 and 255. The color can be found by combining the values in each

of the three layers.

Dept. Of Computer Science, St Philomena’s College pg. 41

 Social Distance Detector

ReLU layer

The ReLU (rectified linear unit) layer is another step to our convolution layer. You’re applying

an activation function onto your feature maps to increase non-linearity in the network. This is because

images themselves are highly non-linear! It removes negative values from an activation map by setting
them to zero.

Convolution is a linear operation with things like element wise matrix multiplication and addition. The
real-world data we want our CNN to learn will be non-linear. We can account for that with an operation

like ReLU. You can use other operations like tanh or sigmoid. ReLU, however, is a popular choice

because it can train the network faster without any major penalty to generalization accuracy.

WORKING OF YOLO ALGORITHM

YOLO ("you only look once") is one of the popular algorithm because it achieves high accuracy
along with being able to run in real-time. The algorithm "only looks once" at the image, i.e. it
requires only one forward propagation pass through the network so that it can make predictions.
After non-max suppression, it gives the name of the recognized object along with the bounding
boxes around them. The diagrams for explaining YOLO are from Andrew Ng’s video explanation
of the same

Anchor box
By using Bounding boxes for object detection, only one object can be identified by a grid. So, for
detecting more than one object we go for Anchor box.

Dept. Of Computer Science, St Philomena’s College pg. 42

 Social Distance Detector

Fig. 2- Detection of object using anchor box

Consider the above picture, in that both the human and the car’s midpoint come under the same grid
cell. For this case, we use the anchor box method. The purple color grid cells denote the two anchor
boxes for those objects. Any number of anchor boxes can be used for a single image to detect multiple
objects. In our case, we have taken two anchor boxes.

Fig. 3- Anchor boxes

The above figure shows the anchor box of the image we considered. The vertical anchor box is for the
human and the horizontal one is the anchor box of the car.

Model Details:
The model details are as follows:

•The input is batch of images with shape (m, 608, 608, 3)

•The output is a list of bounding boxes with the recognized classes. Each bounding box is denoted
by 6 numbers (p_c, b_x, b_y, b_h, b_w, c). If you expand c i.e. classes we get an 80-dimensional vector,
each bounding box is then represented by 85 numbers.

Dept. Of Computer Science, St Philomena’s College pg. 43

 Social Distance Detector

Fig. 4 -Architecture of YOLO So, the architecture can be summarized as:

IMAGE (m, 608, 608, 3) -> DEEP CNN -> ENCODING (m, 19, 19, 5, 85).

If the center or the midpoint of an object falls into a grid cell, then that grid cell is responsible for
detecting that object.

Since in the model we are using 5 anchor boxes and each of the 19 x19 cells thus encodes
information about 5 boxes. Anchor boxes are defined by their width and height. For simplicity, the
image is first flattened that is the last two last dimensions of the shape (19, 19, 5, 85) encoding. so
the output of the Deep CNN is in form: (19, 19, 425). Fig 3 shows the flattening.

Fig. 5- Flatten last two dimensions

Now for each grid that is for each box of the cell compute the following elementwise product as
well as the probability that the box contains a particular class.

Dept. Of Computer Science, St Philomena’s College pg. 44

 Social Distance Detector

Fig. 6- Determining the probability

After plotting only the boxes that the algorithm had given of higher probability, there are too many
boxes and hence filtering these boxes is very important for accuracy.

Fig. 7-Output without Filtering algorithms

Each cell has 5 anchor boxes. So in total if we calculate, the model predicts: 19x19x5 = 1805
boxes. In the figure different colors denote different classes. So we filter the algorithm's output down
to a less number of boxes i.e. much smaller number of detected objects. To do this we carry out two
important steps:

• Get rid of boxes with a low score that is to remove the box which are not very confident
aboutdetecting a class

• Select only one box that overlaps many other boxes with each other and which detects the same
object.

After the filtering based on the score of the classes, the second filter which is applied on the left boxes
is the Non maximum Suppression (NMS).

Dept. Of Computer Science, St Philomena’s College pg. 45

 Social Distance Detector

Fig. 8 -Non Max Suppression

It uses the concept of Intersection Over Union (IoU). IoU is the ratio of intersection of two boxes to
the union of the boxes. This is shown in Fig 7.

Yolo Working Procedure:

Here are the fundamental concepts of how YOLO object detection can able to detect an object.

The YOLO detector can predict the class of object, its bounding box, and the probability of the class
of object in the bounding box. Each bounding box is having the following parameters.

 The center position of the bounding box in the image (bx, by)
 The width of the box( bw )
 The height of the box ( bh )
 The class of object ( c )
Codes:
import the necessary packages
from .social_distancing_config import NMS_THRESH

from .social_distancing_config import MIN_CONF

import numpy as np
import cv2

def detect_people(frame, net, ln, personIdx=0):

# grab the dimensions of the frame and initialize the list of

# results

Dept. Of Computer Science, St Philomena’s College pg. 46

 Social Distance Detector

(H, W) = frame.shape[:2]
results = []

# construct a blob from the input frame and then perform a forward

# pass of the YOLO object detector, giving us our bounding boxes

# and associated probabilities
blob = cv2.dnn.blobFromImage(frame, 1 / 255.0, (416, 416),
swapRB=True, crop=False)

net.setInput(blob)
layerOutputs = net.forward(ln)

# initialize our lists of detected bounding boxes, centroids, and

# confidences, respectively

boxes = []
centroids = []
confidences = []

# loop over each of the layer outputs

for output in layerOutputs:
# loop over each of the detections
for detection in output:

# extract the class ID and confidence (i.e., probability)

# of the current object detection

scores = detection[5:]
classID = np.argmax(scores)
confidence = scores[classID]

Dept. Of Computer Science, St Philomena’s College pg. 47

 Social Distance Detector

# filter detections by (1) ensuring that the object

# detected was a person and (2) that the minimum
# confidence is met

if classID == personIdx and confidence > MIN_CONF:

# scale the bounding box coordinates back relative to
# the size of the image, keeping in mind that YOLO
# actually returns the center (x, y)-coordinates of

# the bounding box followed by the boxes' width and

# height
box = detection[0:4] * np.array([W, H, W, H])
(centerX, centerY, width, height) = box.astype("int")

# use the center (x, y)-coordinates to derive the top

# and and left corner of the bounding box

x = int(centerX - (width / 2))

y = int(centerY - (height / 2))

# update our list of bounding box coordinates,

# centroids, and confidences
boxes.append([x, y, int(width), int(height)])
centroids.append((centerX, centerY))
confidences.append(float(confidence))

# apply non-maxima suppression to suppress weak, overlapping

# bounding boxes

idxs = cv2.dnn.NMSBoxes(boxes, confidences, MIN_CONF, NMS_THRESH)

Dept. Of Computer Science, St Philomena’s College pg. 48

 Social Distance Detector

# ensure at least one detection exists

if len(idxs) > 0:
# loop over the indexes we are keeping
for i in idxs.flatten():
# extract the bounding box coordinates
(x, y) = (boxes[i][0], boxes[i][1])
(w, h) = (boxes[i][2], boxes[i][3])

# update our results list to consist of the person

# prediction probability, bounding box coordinates,

# and the centroid

r = (confidences[i], (x, y, x + w, y + h), centroids[i])
results.append(r)

# return the list of results

return results

Social Distance Detector:

# USAGE
# python social_distance_detector.py --input pedestrians.mp4
# python social_distance_detector.py --input pedestrians.mp4 --output output.avi

# import the necessary packages

from social_detect import social_distancing_config as config

from social_detect.detection import detect_people
from scipy.spatial import distance as dist

Dept. Of Computer Science, St Philomena’s College pg. 49

 Social Distance Detector

import numpy as np
import argparse
import imutils
import cv2
import os

# construct the argument parse and parse the arguments

ap = argparse.ArgumentParser()

ap.add_argument("-i", "--input", type=str, default="",

help="path to (optional) input video file")
ap.add_argument("-o", "--output", type=str, default="",
help="path to (optional) output video file")
ap.add_argument("-d", "--display", type=int, default=1,
help="whether or not output frame should be displayed")
args = vars(ap.parse_args())

# load the COCO class labels our YOLO model was trained on
labelsPath = os.path.sep.join([config.MODEL_PATH, "coco.names"])
LABELS = open(labelsPath).read().strip().split("\n")

# derive the paths to the YOLO weights and model configuration

weightsPath = os.path.sep.join([config.MODEL_PATH, "yolov3.weights"])
configPath = os.path.sep.join([config.MODEL_PATH, "yolov3.cfg"])

# load our YOLO object detector trained on COCO dataset (80 classes)
print("[INFO] loading YOLO from disk...")
net = cv2.dnn.readNetFromDarknet(configPath, weightsPath)

Dept. Of Computer Science, St Philomena’s College pg. 50

 Social Distance Detector

# check if we are going to use GPU

if config.USE_GPU:

# set CUDA as the preferable backend and target

print("[INFO] setting preferable backend and target to CUDA...")
net.setPreferableBackend(cv2.dnn.DNN_BACKEND_CUDA)
net.setPreferableTarget(cv2.dnn.DNN_TARGET_CUDA)

# determine only the *output* layer names that we need from YOLO
ln = net.getLayerNames()
ln = [ln[i[0] - 1] for i in net.getUnconnectedOutLayers()]

# initialize the video stream and pointer to output video file

print("[INFO] accessing video stream...")
vs = cv2.VideoCapture(args["input"] if args["input"] else 0)
writer = None

# loop over the frames from the video stream

while True:
# read the next frame from the file
(grabbed, frame) = vs.read()

# if the frame was not grabbed, then we have reached the end
# of the stream
if not grabbed:
break

Dept. Of Computer Science, St Philomena’s College pg. 51

 Social Distance Detector

# resize the frame and then detect people (and only people) in it
frame = imutils.resize(frame, width=700)
results = detect_people(frame, net, ln,
personIdx=LABELS.index("person"))

# initialize the set of indexes that violate the minimum social

# distance

violate = set()

# ensure there are *at least* two people detections (required in

# order to compute our pairwise distance maps)
if len(results) >= 2:
# extract all centroids from the results and compute the
# Euclidean distances between all pairs of the centroids
centroids = np.array([r[2] for r in results])

D = dist.cdist(centroids, centroids, metric="euclidean")

# loop over the upper triangular of the distance matrix

for i in range(0, D.shape[0]):
for j in range(i + 1, D.shape[1]):

# check to see if the distance between any two

# centroid pairs is less than the configured number
# of pixels

if D[i, j] < config.MIN_DISTANCE:

# update our violation set with the indexes of

# the centroid pairs
violate.add(i)

Dept. Of Computer Science, St Philomena’s College pg. 52

 Social Distance Detector

violate.add(j)

# loop over the results

for (i, (prob, bbox, centroid)) in enumerate(results):

# extract the bounding box and centroid coordinates, then
# initialize the color of the annotation
(startX, startY, endX, endY) = bbox
(cX, cY) = centroid

color = (0, 255, 0)

# if the index pair exists within the violation set, then

# update the color

if i in violate:
color = (0, 0, 255)

# draw (1) a bounding box around the person and (2) the

# centroid coordinates of the person,

cv2.rectangle(frame, (startX, startY), (endX, endY), color, 2)

cv2.circle(frame, (cX, cY), 5, color, 1)

# draw the total number of social distancing violations on the

# output frame
text = "Social Distancing Violations: {}".format(len(violate))

cv2.putText(frame, text, (10, frame.shape[0] - 25),

cv2.FONT_HERSHEY_SIMPLEX, 0.85, (0, 0, 255), 3)

# check to see if the output frame should be displayed to our

Dept. Of Computer Science, St Philomena’s College pg. 53

 Social Distance Detector

# screen
if args["display"] > 0:

# show the output frame

cv2.imshow("Frame", frame)
key = cv2.waitKey(1) & 0xFF

# if the `q` key was pressed, break from the loop

if key == ord("q"):

break

# if an output video file path has been supplied and the video
# writer has not been initialized, do so now
if args["output"] != "" and writer is None:
# initialize our video writer
fourcc = cv2.VideoWriter_fourcc(*"MJPG")
writer = cv2.VideoWriter(args["output"], fourcc, 25,
(frame.shape[1], frame.shape[0]), True)

# if the video writer is not None, write the frame to the output
# video file
if writer is not None:
writer.write(frame)

Dept. Of Computer Science, St Philomena’s College pg. 54

 Social Distance Detector

CHAPTER-6
TESTING

INTRODUCTION

Testing is the way toward running a framework with the expectation of discovering blunders. Testing
upgrades the uprightness of the framework by distinguishing the deviations in plans and blunders in
the framework. Testing targets distinguishing blunders – prom zones. This aides in the avoidance of
mistakes in the framework. Testing additionally adds esteems to the item by affirming the client's
necessity.

The primary intention is to distinguish blunders and mistake get-prom zones in a framework. Testing
must be intensive and all around arranged. A somewhat tried framework is as terrible as an untested
framework. Furthermore, the cost of an untested and under-tried framework is high. The execution
is the last and significant stage. It includes client preparation, framework testing so as to guarantee
the effective running of the proposed framework. The client tests the framework and changes are
made by their requirements. The testing includes the testing of the created framework utilizing
different sorts of information. While testing, blunders are noted and rightness is the mode.

6.1 OBJECTIVES OF TESTING

• Testing in a cycle of executing a program with the expectation of discovering mistakes.

• A effective experiment is one that reveals an up 'til now unfamiliar blunder.

Framework testing is a phase of usage, which is pointed toward guaranteeing that the framework
works accurately and productively according to the client's need before the live activity initiates. As
expressed previously, testing is indispensable to the achievement of a framework. Framework testing
makes the coherent presumption that if all the framework is right, the objective will be effectively

Dept. Of Computer Science, St Philomena’s College pg. 55

 Social Distance Detector

accomplished. A progression of tests are performed before the framework is prepared for the client
acknowledgment test.

6.2 TESTING METHODS

System testing is a stage of implementation. This helps the weather system works accurately and
efficiently before live operation commences. Testing is vital to the success of the system. The
candidate system is subject to a variety of tests: online response, volume, stress, recovery, security,
and usability tests series of tests are performed for the proposed system are ready for user acceptance
testing.

White Box Testing

The test is conducted during the code generation phase itself. All the errors were rectified at
the moment of its discovery. During this testing, it is ensured that

• All independent module have been exercised at least one

• Exercise all logical decisions on their true or false side.

• Execute all loops at their boundaries.

Black Box Testing

It is focused around the practical necessities of the product. It's anything but a choice to white box
testing; rather, it is a reciprocal methodology that is probably going to reveal an alternate class of
blunders than White Box strategies. It is endeavored to discover mistakes in the accompanying
classes.

• Incorrect or missing capacities

• Interface blunders

• Errors in an information structure or outside information base access

Dept. Of Computer Science, St Philomena’s College pg. 56

 Social Distance Detector

Unit Testing
Unit testing chiefly centers around the littlest unit of programming plan. This is known as module
testing. The modules are tried independently. The test is done during the programming stage itself.
In this progression, every module is discovered to be working acceptably as respects the normal
yield from the module.

Integration Testing

Mix testing is an efficient methodology for developing the program structure, while simultaneously
leading tests to reveal blunders related with the interface. The goal is to take unit tried modules and
manufacture a program structure. All the modules are joined and tried in general.

Output Testing

Subsequent to performing approval testing, the following stage is yield trying of the proposed
framework, since no framework could be valuable on the off chance that it doesn't create the
necessary yield in a particular configuration. The yield design on the screen is discovered to be right.
The organization was planned in the framework configuration time as indicated by the client needs.
For the printed copy likewise, the yield comes according to the predefined prerequisites by the client.
Subsequently yield testing didn't bring about any amendment for the framework.

User Acceptance Testing

Client acknowledgment of a framework is the vital factor for the achievement of any framework.
The framework viable is tried for client acknowledgment by continually staying in contact with the
imminent framework clients at the hour of creating and making changes at whatever point required.

Dept. Of Computer Science, St Philomena’s College pg. 57

 Social Distance Detector

VALIDATION

Toward the consummation of the reconciliation testing, the product is totally amassed as bundle
interfacing blunders have been revealed and adjusted and a last arrangement of programming tests
starts in approval testing. Approval testing can be characterized from multiple points of view,
however a straightforward definition is that the approval succeeds when the product work in a way
that is normal by the client. After approval test has been directed as follows:

• The capacity or execution qualities adjust to detail and are acknowledged.

• A deviation from the particular is revealed and a lack list is made.

• Proposed framework viable has been tried by utilizing an approval test and discovered to be
working acceptably.

6.3 TEST REPORTS

The users test the developed system when changes are made according to the needs. The
testing phase involves the testing of the developed system using various kinds of data. An elaborate
testing of data is prepared and system is tested using the test data. Test cases are used to check for
outputs with different set of inputs.

5 Test Cases

Test Case Test Test Expected Actual Pass

Purpose condition outcome result or
Fail

Dept. Of Computer Science, St Philomena’s College pg. 58

 Social Distance Detector

Load Load Check with Load Videos is Pass

Video Video vide format Video. loaded
data in support Successfully.
Mp4 video or
format. send error
message.

Convert Video data Convert Frames are As Pass

videos into video into converted Expected.
frames 14 frames suucessfully
per second

Yolo Pre- Frames Training of As Pass

object processed areTrained data is Expected.
detection data for each complete

Algorithm available
data

Detecting Frame Frames Testing of As Pass

object based passed input data is Expected.
person object to yolo for complete
detection object
detection

Identifying Result Person with Predicted As Pass

social obtained maintain result Expected.
detectin from yolo scial
frm vide detection
and violated
are
identified

Dept. Of Computer Science, St Philomena’s College pg. 59

 Social Distance Detector

Failed test Cases

Test Case Test Test Expected Actual Pass

Purpose condition outcome result or
Fail

Load Load Check with Load File format fail

Video Video vide format Video. Is not
data in support
Mp4 video or Supprted.
format. send error
message.

Convert Video data Convert Frames are Video Pass

videos into video into 14 converted difficult t
frames frames per suucessfully cnert into
second frames .

Yolo Pre- Frames Training of Version Pass

object processed areTrained data is error,
detection data for each complete Or gpu is
Algorithm available required
data

Detecting Frame based Frames Testing of Not Pass

object object passed input data is detecting
person detection to yolo for complete objects.

Dept. Of Computer Science, St Philomena’s College pg. 60

 Social Distance Detector

object
detection

Identifying Result Person with Predicted Not Pass

social obtained maintain result identified
detectin from yolo scial object.
frm vide detection
and violated
are identified

Dept. Of Computer Science, St Philomena’s College pg. 61

 Social Distance Detector

CHAPTER-7
CONCLUSION:
This work proposed an AI and monocular camera based real-time system to monitor the social
distancing. The article proposes an efficient real-time deep learning based framework to automate
the process of monitoring the social distancing via object detection and tracking approaches, where
each individual is identified in the real-time with the help of bounding boxes. The generated
bounding boxes aid in identifying the clusters or groups of people satisfying the closeness property
computed with the help of pairwise vectorized approach. The number of violations are confirmed
by computing the number of groups formed and violation index term computed as the ratio of the
number of people to the number of groups. The extensive trials were conducted with popular state-
of-the-art object detection models: Faster RCNN, SSD, and YOLO v3, where YOLO v3 illustrated
for the efficient performance. Since this approach is highly sensitive to the spatial location of the
camera, the same approach can be fine tuned to better adjust with the corresponding field of view.

Dept. Of Computer Science, St Philomena’s College pg. 62

 Social Distance Detector

CHAPTER-8
FUTURE ENHANCEMENT:

In this project we applied yolo Object detection techniques for identifying person and The generated
bounding boxes aid in identifying the clusters or groups of people satisfying the closeness property
computed with the help of pairwise vectorised approach.in future we can implement for mask detection
in public areas. we can apply different algorithms for better accuracy which works in less time
complexity. As for future scope we can make use of tensor flow library for object detection.

Dept. Of Computer Science, St Philomena’s College pg. 63

 Social Distance Detector

CHAPTER-9
REFERENCES

[1] S. A. Abbas and A. Zisserman. A geometric approach to obtain a bird’s eye view from an image.
In 2019 IEEE/CVF International Conference on Computer Vision Workshop (ICCVW), pages
4095–4104, 2019.

[2] P. Ball. Critical Mass. Arrow Books, 2004.

[3] C. Bartneck, D. Kulic, E. Croft, and S. Zoghbi. Measurement ´ instruments for the
anthropomorphism, animacy, likeability, perceived intelligence, and perceived safety of robots.
International Journal of Social Robotics, 1(1):71–81, 2009.

[4] J.-C. Bazin and M. Pollefeys. 3-line ransac for orthogonal vanishing point detection. In 2012
IEEE/RSJ International Conference on Intelligent Robots and Systems, pages 4282–4287. IEEE,
2012.

[5] J.-C. Bazin, Y. Seo, C. Demonceaux, P. Vasseur, K. Ikeuchi, I. Kweon, and M. Pollefeys. Globally
optimal line clustering and vanishing point estimation in manhattan world. In 2012 IEEE Conference
on Computer Vision and Pattern Recognition, pages 638–645. IEEE, 2012.

[6] L. Bazzani, M. Cristani, D. Tosato, M. Farenzena, G. Paggetti, G. Menegaz, and V. Murino.

Social interactions by visual focus of attention in a three-dimensional environment. Expert Systems,
30(2):115–127, 2013.

[7] A. N. Belbachir. Smart cameras, volume 2. Springer, 2010.

[8] C. BenAbdelkader and Y. Yacoob. Statistical body height estimation from a single image. In
2008 8th IEEE International Conference on Automatic Face & Gesture Recognition, pages 1–7.
IEEE, 2008. [9] R. Benenson, M. Omran, J. Hosang, and B. Schiele. Ten years of pedestrian

Dept. Of Computer Science, St Philomena’s College pg. 64

 Social Distance Detector

detection, what have we learned? In European Conference on Computer Vision, pages 613–627.
Springer, 2014.

[10] H. Blumer. Collective behavior. Bobbs-Merrill Company Incorporated, 1957.

Dept. Of Computer Science, St Philomena’s College pg. 65

 Social Distance Detector

Dept. Of Computer Science, St Philomena’s College pg. 66

 Social Distance Detector

Dept. Of Computer Science, St Philomena’s College pg. 67

 Social Distance Detector

Dept. Of Computer Science, St Philomena’s College pg. 68