A SYNOPSIS ON

Facial Recognition System

Submitted in partial fulfillment of the requirement for the award of the

degree of

MASTER OF COMPUTER APPLICATION

Submitted by:

Janvi Yadav 1102495

Under the Guidance of

Dr. Varsha Mittal
Associate

Department of Computer Science and Engineering

Graphic Era (Deemed to be University)
Dehradun, Uttarakhand
May-2023
 CANDIDATE’S DECLARATION
I/We hereby certify that the work which is being presented in the Synopsis entitled “Facial
Recognition System ” in partial fulfillment of the requirements for the award of the Degree
of Master of Computer Application in the Department of Computer Science and Engineering
of the Graphic Era (Deemed to be University), Dehradun shall be carried out by the
undersigned under the supervision of Dr. Varsha Mittal, Associate, Department of Computer
Science and Engineering, Graphic Era (Deemed to be University), Dehradun.

Janvi Yadav 1102495 signature

The above mentioned students shall be working under the supervision of the undersigned on
the “Facial Recognition System”

Signature Signature
Associate Head of the Department

Internal Evaluation (By DPRC Committee)

Status of the Synopsis: Accepted / Rejected

Any Comments:

Name of the Committee Members: Signature with Date

1.
 Table of Contents

Chapter No. Description Page No.

Chapter 1 Introduction and Problem Statement 1-3
Chapter 2 Background/ Literature Survey 4-6
Chapter 3 Objectives 7-8
Chapter 4 Hardware and Software Requirements 9
Chapter 5 Possible Approach/ Algorithms 10-12
References 13
Chapter 1

Introduction and Problem Statement

In the following sections, a brief introduction and the problem statement for the work has
been included.

1.1 Introduction

In today's fast-paced world, the demand for reliable and efficient security systems has never been
greater. Traditional methods of identity verification and access control are often time-consuming and
prone to human error, prompting the need for more advanced technological solutions. Among these
innovative advancements, facial recognition technology has emerged as a highly promising tool for
enhancing security measures. With its ability to accurately identify and authenticate individuals based
on their unique facial features, a facial recognition logging system holds immense potential to
revolutionize various industries, ranging from law enforcement and airports to corporate offices and
public spaces.

Facial recognition technology employs sophisticated algorithms and deep learning techniques to
analyze and compare facial patterns from images or video feeds. By capturing and processing facial
data, the system can verify an individual's identity against a pre-existing database, granting or denying
access accordingly. The widespread adoption of this technology has been fueled by its numerous
advantages, including its speed, accuracy, and versatility.

Face recognition is a biometric technology that involves identifying or verifying a person's identity by
comparing their facial features with those stored in a database. This process is achieved through
pattern recognition algorithms that analyze and match various facial features, such as the eyes, nose,
mouth, and jawline.

The task of face recognition can be divided into two main categories: verification and identification.
In verification, the system compares a person's face to a single reference image to confirm their
identity. This is commonly used for tasks such as unlocking smartphones or accessing secure
facilities.

In identification, the system compares a person's face to a database of known individuals to determine
their identity. This is commonly used in law enforcement and security applications to identify
suspects or persons of interest.

Face recognition systems are designed to learn and adapt to new faces over time, using machine
learning algorithms to continually improve their accuracy and performance. This involves analyzing
new faces and comparing them to existing data to identify patterns and features that can be used to
improve the accuracy of the system.

In addition to its use in law enforcement and security, face recognition is increasingly being used in
commercial applications, such as marketing and advertising. For example, some retailers are using
face recognition technology to identify and analyze customer demographics, preferences, and
behaviors, in order to provide more personalized and targeted advertising.

The advent of facial recognition systems represents a significant leap forward in security and
identification technology. By harnessing the power of advanced algorithms and deep learning, these
systems offer unparalleled speed, accuracy, and versatility in verifying individuals' identities. From
law enforcement and access control to personalized customer experiences, the potential applications
of facial recognition technology are vast and diverse. However, it is imperative to implement these
systems responsibly, addressing privacy concerns and ethical considerations, to ensure that they
contribute positively to society. With ongoing research, development, and public engagement, facial
recognition technology can be harnessed as a powerful tool for enhancing security and efficiency
across various sectors.
 Figure 1.1 Basic functioning of a face recognition system

1.2 Problem Statement

The problem statement for the present work can be stated as follows:

To design and develop a facial recognition attendance system that accurately and efficiently tracks the
attendance of students/employees using their facial features. The system should be user-friendly, cost-
effective, and capable of handling large volumes of data in real-time. The system should also be
secure, protecting the privacy of individuals' biometric data while ensuring that only authorized
individuals can access the data. Finally, the system should be able to handle various environmental
conditions such as different lighting conditions, occlusion, and pose variations to provide reliable
attendance tracking.
Chapter 2

Background/ Literature Survey

Facial recognition technology has seen a surge in development and implementation in recent

years due to its ability to provide a fast, secure, and reliable way of identifying individuals. A

significant number of research studies and publications have been conducted in this area to

improve the accuracy, efficiency, and reliability of facial recognition systems.

It has become increasingly important in various applications such as security systems, credit

card verification, and criminal identification. Much of the work in computer recognition of

faces has focused on detecting individual features like the eyes, mouth, and head outline and

defining the face model by the position, size, and relationships among these features.

Although there are three different approaches to face recognition, two basic methods give rise

to these approaches. The first method is based on information theory concepts and principal

component analysis, where the most relevant information that describes a face is derived

from the entire face image. The second method extracts feature vectors from the basic parts

of a face such as eyes, nose, mouth, and chin with the help of deformable templates and

extensive mathematics.

Face recognition is done in two steps: feature extraction and face matching. The ability to

model a particular face and distinguish it from a large number of stored face models is

desirable in many applications, including criminal identification.

Moreover, several studies have been conducted to address the challenges of facial recognition

systems in terms of privacy and security. For instance, Jain et al. (2016) discussed the privacy

concerns associated with the use of facial recognition systems and proposed a framework for
enhancing privacy in these systems. The authors suggested the use of secure multi-party

computation and privacy-enhancing technologies to protect sensitive biometric data.

Additionally, a study by Yang et al. (2018) proposed a secure facial recognition system based

on block chain technology. The authors suggested the use of a block chain-based

decentralized architecture to protect the privacy and security of biometric data.

Facial recognition algorithms often use landmarks or features on the face, such as the eyes,

nose, cheekbones, and jaw, to identify individuals in images. In a study by Klare et al. (2015),

the authors evaluated the performance of facial recognition systems in the unconstrained

environment. They analyzed the performance of several facial recognition algorithms and

found that the accuracy of facial recognition systems can be improved by using large datasets

for training and implementing deep learning techniques.

In another study by Bhattacharyya et al. (2019), the authors developed a facial recognition

system using a combination of machine learning techniques and computer vision algorithms.

They evaluated the system's accuracy and efficiency in different lighting conditions and

found that the system's accuracy can be improved by incorporating illumination

normalization

There are two main approaches to facial recognition: the geometric approach and the

photometric approach.

1. Geometric approach

The geometric approach uses the segments, perimeters, and areas of figures formed by the

facial points to compare images. However, this approach can be problematic if the image is

of low quality or if facial points are covered.

 2. Photogenic approach

The photometric approach uses statistical methods to compare values in an image to

templates, taking into account light and geometric angles. This approach requires multiple

registered images of the same person and cannot handle new images of a different person not

included in the gallery set.

Chapter 3

Objectives

● The objective of a facial recognition attendance system is to automate the

attendance process by accurately identifying and verifying an individual's

identity through facial recognition technology. This system uses face-

recognition library of python to detect and recognize facial features in

real-time, allowing for quick and accurate identification.

● The primary objective of this system is to improve the efficiency and

accuracy of attendance tracking by eliminating the need for manual

attendance tracking methods. The system can quickly and easily identify

individuals as they enter or leave a specific location, such as a classroom

or workplace, without the need for manual check-ins or sign-ins.

● Another objective of this system is to provide a secure and reliable

method of attendance tracking, helping to prevent instances of proxy

attendance or fraud.

● Overall, the objective of a facial recognition attendance is to provide an

efficient, secure, and reliable method of attendance tracking, with a wide

range of potential applications in both educational and professional

settings. However, it is important to ensure that the system is designed

and implemented in a way that protects individual privacy and civil

liberties.
Chapter 4

System Requirements

4.1 Hardware Requirements

● System : Pentium i3 Processor or above.

● Hard Disk : 40 GB.

● Ram : 4 GB

● Display: : 1024 x 768 High color-64-bit

4.2 Software Requirements

● Operating system : Windows 10 or above.

● Language : Python 3.11

● IDE : Visual Studio Code

Chapter 5

Possible Approach/ Algorithms

1. Principal Component Analysis Using Eigenfaces:

This algorithm is a popular approach to face recognition, where it uses the Eigenvectors of
the image dataset to recognize new faces. Eigenfaces are created by taking the most
important facial features from a set of images and converting them into a set of vectors that
can be used for analysis. The algorithm works by projecting an image onto this Eigenface
space and finding the best match between the projected image and a set of training images.
This method is computationally efficient and works well with large datasets.

2. Linear Discriminate Analysis:

This approach is also known as Fisherfaces, where it aims to find the optimal projection that
separates different classes of images by maximizing the between-class distance and
minimizing the within-class distance. The algorithm works by first computing the mean face
for each class, and then finding the directions in the Eigenspace that maximize the ratio of the
between-class variance to the within-class variance. It is commonly used for face recognition,
and it works well when there is a significant difference between different classes.

3. Elastic Bunch Graph Matching Using the Fisher Face Algorithm:

This algorithm is an extension of the Fisher Face algorithm, where it uses an elastic bunch
graph matching technique to improve the accuracy of face recognition. The algorithm creates
a graph of facial features and finds the best matching graph for a new face image. The elastic
matching technique allows for deformations and distortions in the facial features, which
improves the algorithm's ability to recognize faces under different conditions.

4. The Hidden Markov Model:

This algorithm is a statistical model that is commonly used for speech recognition, but it can
also be used for face recognition. The model uses a sequence of observations to estimate the
probability of a hidden state. In the case of face recognition, the hidden state could represent
the identity of the person in the image. The algorithm can work with partial observations,
making it robust to occlusions and variations in lighting conditions.

5. The Multilinear Subspace Learning Using Tensor Representation:

This algorithm uses a multilinear subspace learning approach to learn a tensor representation
of the image data. The algorithm works by representing each image as a high-dimensional
tensor and then finding a low-dimensional subspace that captures the most important features
of the images. The algorithm can work with images that have missing or incomplete data,
making it robust to occlusions and partial observations.

6. The Neuronal Motivated Dynamic Link Matching:

This algorithm is inspired by the structure and function of the human brain. It uses a dynamic
link matching technique to match the input image with a set of training images. The
algorithm works by first creating a set of neurons that represent the facial features, and then
linking these neurons together in a dynamic network. The network learns to recognize faces
by adjusting the weights of the links between the neurons, making it robust to variations in
the facial features.
Begin
1. def fisherfaces (X ,y , num_components =0) :
2. y = np . asarray (y)
3. [n , d] = X . shape
4. c = len ( np . unique (y ))
5. [ eigenvalues_pca , eigenvectors_pca , mu_pca ] = pca (X , y , (n -c ))
6. [ eigenvalues_lda , eigenvectors_lda ] = lda ( project ( eigenvectors_pca , X ,
mu_pca ) , y ,
7. num_components )
8. eigenvectors = np . dot ( eigenvectors_pca , eigenvectors_lda )
9. return [ eigenvalues_lda
End

Table 5.1 Pseudo code of the FisherFace algorithm

References

[1] Introna LD, Nissenbaum H. “Facial recognition technology: a survey of policy and
implementation issues”. Report of the Center for Catastrophe Preparedness and Response. New
York: New York University, 2009.
[2] Agrawal AK, Singh YN. “Evaluation of face recognition methods in unconstrained
environments.” Procedia Comput Sci 2015
[3] Fontaine X, Achanta R, Süsstrunk S. “Face recognition in real-world images”. IEEE
International Conference on Acoustics, Speech and Signal Processing (ICASSP), New Orleans,
LA, 5–9 March 2017, pp.1482–1486. Piscataway, NJ: IEEE.
[4] Matthew T, Pentland A. “Eigenfaces for recognition”. J Cogn Neurosci 1991; 3: 71–89.

Website
[1] www.hindawi.com/journals/js/2021/
[2] https://recfaces.com/articles/facial-recognition-algorithms