Face detection for online attendance system.

1. A study on
“IMAGE BASED ATTENDENCE SYSTEM”
A Report Submitted
in partial fulfilment for the award of degree of
Bachelor of Technology
In
Computer Science and Engineering
Under
The Assam Royal Global University
Submitted By-
Swarup Das (182025049)
Somodeep Seal (182025046)
Under the guidance of
Afsana Laskar
Lecturer
Department of Computer Science and Engineering
Royal School of Engineering & Technology
Guwahati-781035
September 2021-January-2022

3. DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
ROYAL SCHOOL OF ENGINEERING & TECHNOLOGY
Certificate to the Head of the Department
This is to certify that the project work entitled " IMAGE BASED ATTENDENCE SYSTEM "
is hereby approved as a Bonafede work of study as an engineering subject, carried out by the
students - Swarup Das (182025049) & Somodeep Seal (182025046) of 7th Semester, B.Tech,
Computer Science and Engineering Department under the guidance of Afsana Laskar, Lecturer,
Computer Science and Engineering Department, The Assam Royal Global University. The work
in the project is a genuine work carried out by the students as a prerequisite to the degree for which
it has been submitted.
Date:
Place: Guwahati __________________
Dr. Aniruddha Deka
Assistant Professor & H.O.D.,
Department of Computer Science
and Engineering
Royal School of Engineering &
Technology
I

4. DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
ROYAL SCHOOL OF ENGINEERING & TECHNOLOGY
CERTIFICATE OF APPROVAL
This is to certify that the project report entitled " IMAGE BASED ATTENDENCE SYSTEM "
submitted by Swarup Das (182025049) & Somodeep Seal (182025046) of 7th Semester, B.Tech,
Computer Science and Engineering Department, The Assam Royal Global University, Guwahati
in partial fulfilment for the award of the degree of B. Tech in Computer Science and Engineering
is a bonafede record of project work carried out by him under my supervision. The contents of this
report, in full or in parts, have not been submitted to any other Institution or University for the
award of any degree or diploma.
Date: Project Guide:
Place: Guwahati ___________________
Afsana Laskar
Lecturer,
Department of Computer Science
and Engineering
Royal School of Engineering &
Technology
II

6. DEPARTMENT OF COMPUTER SCIENCE AND ENGINEERING
ROYAL SCHOOL OF ENGINEERING & TECHNOLOGY
Declaration by the Candidate
We certify that this project entitled " IMAGE BASED ATTENDENCE SYSTEM ", a perquisite
towards partial fulfilment for the award of B.Tech degree in Computer Science and Engineering,
Royal School of Engineering & Technology, Guwahati contains no materials previously published
or written by another person, except where due reference has made in the text as in an accurate
record of our work carried under the guidance and supervision of Afsana Laskar ,Lecturer,
Department of Computer Science and Engineering.
Date-
Place: Guwahati
Swarup Das Somodeep Seal
(Roll No: 182025049) (Roll No: 182025046)
IV

7. ACKNOWLEDGEMENT
We would like to extend our gratitude and our sincere thanks to our honourable, esteemed guide,
Afsana Laskar, Lecturer, Computer Science and Engineering, Royal School of Engineering and
Technology. She is not only a helpful teacher with deep vision but also most importantly a kind
person. We sincerely thank for his exemplary guidance and encouragement. Her trust and support
inspired us in the most important moments of making right decisions and we are glad to work with
her.
We would like to thank all our other faculty members for their guidance and ideas that helped us
make this project successful. This project is by far the most significant accomplishment in our life
and it would be impossible without the people who supported us and believed in us.
We would like to thank all our friends for all the thoughtful and mind stimulating discussions that
we had, which made us think beyond the obvious. Last but not the least, we would like to thank
our parents, who taught us the value of hard work.
V

8. ABSTRACT
Face recognition, as one of the most successful applications of image analysis, has recently
gained significant attention, especially during the past several years. Our idea proposes an
automatic face recognition attendance system for students using computer vision and mini
cameras based on machine learning and narrowband Internet of things for cloud computing.
The system automatically detects and identifies faces and mark present/absent of students with
the help of face detection and it points out the technical challenges of building a face
recognition system. The system will automatically update the student’s presence in the class to
the student’s database and update all the data to the cloud. Face recognition processing,
including major components such as face detection, tracking, alignment, and feature extraction,
and it points out the technical challenges of building a face recognition system. We focus on
the importance of the most successful solutions available so far. It integrates face detection and
face recognition algorithms to build a fast and efficient application which will narrate the
description of the environment to the user.
VI

9. List of Figures
S.No Figure Name Page no.
Fig 1.1 Objective 5
Fig 3.1 Image Pre-processing 8
Fig 4.1 Block Diagram of Haar Cascade 11
Fig 4.2 Haar Cascade 12
Fig 4.3 Diagram of real time face recognition system 14
Fig 4.4 Haar Cascade (Working) 15
Fig 4.5 MTCNN (Working) 16
Fig 4.6 DNN (Working) 16
Fig 4.7 YoloV3 (Working) 17
Fig 4.8 Flow chart of face recognition 19
Fig 4.9 68 face landmarks 19
Fig 5.1 Conversion of Greyscale 21
Fig 5.2 Flow of Proposed Approach (Training part) 22
Fig 5.3 Flow of Proposed Approach (Recognition part) 23
Fig 6.1 Haar Feature 24
Fig 6.2 Algorithm of Working Process 26
Fig 7.1 Results 27
VII

10. Table of Contents Page No
Abstract V
List of Figures VI
Problem Statement 01
Chapter 1 INTRODUCTION 02
1.1 Background 03
1.2 Problem Statement 04
1.3 Motivation 04
1.4 Aims and Objective 05
1.5 Project Description 06
1.6 System Requirements 06
Chapter 2 LITERATURE REVIEW 07
Chapter 3 IMAGE PRE-PROCESSING 08
3.1 Problem to Solve During Image pre-processing 08
3.2 Corrections 09
3.3 Enhancements 09
Chapter 4 Projectbackground 11
4.1 Face Detection 11
4.1.1 Haar Cascade 11
4.1.2 Feature of Haar Cascade 12
4.2 Approachto Solve the Problem 13
4.2.1 Algorithm Used 13
4.2.2 Extracting the Histograms 13
4.2.3 Performing Face Recognition 13
4.3 Flow Diagram 14
4.4 Some of the face detection models 15
4.5 Face Recognition 18
4.5.1 Stages in Face recognition 18
4.5.2 Face Landmark Estimation 19
4.5.3 Encoding of faces 19

11. Chapter 5 METHODOLOGY 20
5.1 Methodology 20
5.1.1 Input Images 20
5.1.2 Limitations of images 20
5.2 Face Detection 21
5.3 Pre-processing 21
5.3.1 Scaling of Image 21
5.3.2 Median Filtering 21
5.3.3 Conversion to Greyscale Image 21
5.3.4 Contrast Limited Adaptive Histogram Equalization 22
5.4 Block diagram 22
Chapter 6 IMPLEMENTATION 24
6.1 Dataset Creation 24
6.2 Face Detection 24
6.3 Face Recognition 25
6.4 Output/Attendance Updation 25
6.5 Working Algorithm 26
Chapter 7 RESULT, DISCUSSION AND DRAWBACKS 27
7.1 Result 27
7.2 Discussion 29
7.3 Drawbacks 30
Chapter 8 CONCLUSION & FUTURE SCOPE 31
8.1 Conclusion 31
8.2 Future Scope 31
REFERENCES 32

12. 1
Problem Statement
Student attendance system using face recognition
Challenge description with context
Our institute enrols approximately 500 students per year. Existing paper-based method is time
consuming and distracting to both students as well as faculties. It is also prone to human errors.
We propose face recognition based smart attendance system. Student attendance can be made more
robust. This also reduces the administrative work of faculties. The attendance data can be stored
on cloud for further processing.
Exact Problem
We propose face recognition with help of Computer (laptop) and mini camera (webcam) based
smart attendance system. By that way student attendance can be made more robust. This also the
reduces the administrative work of faculties. The attendance data can be stored on cloud for further
processing for cloud computing. That way we can sort out irregular students, less attendant student,
punctual students in real time basis.
Users
Any institute like Polytechnics, Degree colleges, Pharmacy Colleges, Medical Colleges, Any
industry, Traffic points, Stampede prone area. students, faculty, HOD, Principal will be stake
holders of the system.
Expected Outcomes
The project aims to streamline the communication system in the organization. One of the biggest
impacts will be ease of communication and less dependence on paper-based system. The project
will reduce the communication and it can lead to higher efficiency of work. Also, using the single
communication medium can increase the data security and data availability. Past data can also be
easily be searched and recorded in organised manner

13. 2
Chapter 1
INTRODUCTION
Human face plays an important role in our day-to-day life mostly for identification of a person.
Face recognition is a part of biometric identification that extracts the facial features of a face, and
then stores it as a unique face print to uniquely recognize a person. Biometric face recognition
technology has gained the attention of many researchers because of its wide application. Face
recognition technology is better than other biometric based recognition techniques like finger-
print, palm-print, iris because of its non-contact process. Recognition techniques using face
recognition can also recognize a person from a distance, without any contact or interaction with
person. The face recognition techniques are currently implemented in social media websites like
Facebook, at the airports, railway stations. The, at crime investigations. Face recognition technique
can also be used in crime reports, the captured photo can be stored in a database, and can be used
to identify a person. Facebook uses the facial recognition technique for automating the process of
tagging people. For face recognition we require large dataset and complex features to identify a
person in all conditions like change of illumination, age, pose, etc. Recent researches show there
is a betterment in facial recognition systems. In the last ten years there is huge development in
recognition techniques.
Attendance tends to be an important parameter in schools and universities. There seems to be a
direct correlation between attendance and academics. There are many factors that affect
academics, in this era of automation an automatic attendance system will tackle all the
shortcomings of a traditional attendance system.
The main objective of this project is to develop face recognition based automated student
attendance system. In order to achieve better performance, the test images and training images of
this proposed approach are limited to frontal and upright facial images that consist of a single face
only. The test images and training images have to be captured by using the same device to ensure
no quality difference. In addition, the students have to register in the database to be recognized.
The enrolment can be done on the spot through the user-friendly interface.

14. 3
1.1 Background
Face recognition is crucial in daily life in order to identify family, friends or someone we are
familiar with. We might not perceive that several steps have actually taken in order to identify
human faces. Human intelligence allows us to receive information and interpret the information in
the recognition process. We receive information through the image projected into our eyes, by
specifically retina in the form of light. Light is a form of electromagnetic waves which are radiated
from a source onto an object and projected to human vision. Robinson-Riegler, G., & Robinson-
Riegler, B. (2008) mentioned that after visual processing done by the human visual system, we
actually classify shape, size, contour and the texture of the object in order to analyse the
information. The analysed information will be compared to other representations of objects or face
that exist in our memory to recognize
The work on face recognition began in 1960. Woody Bledsoe, Helen Chan Wolf and Charles
Bisson had introduced a system which required the administrator to locate eyes, ears, nose and
mouth from images. The distance and ratios between the located features and the common
reference points are then calculated and compared. The studies are further enhanced by Goldstein,
Harmon, and Lesk in 1970 by using other features such as hair colour and lip thickness to automate
the recognition. In 1988, Kirby and Sirovich first suggested principle component analysis (PCA)
to solve face recognition problem. Many studies on face recognition were then conducted
continuously until today (Ashley DuVal, 2012).

15. 4
1.2 Problem Statement
Traditional student attendance marking technique is often facing a lot of trouble. The face
recognition student attendance system emphasizes its simplicity by eliminating classical student
attendance marking technique such as calling student names or checking respective identification
cards. There are not only disturbing the teaching process but also causes distraction for students
during exam sessions. Apart from calling names, attendance sheet is passed around the classroom
during the lecture sessions. The lecture class especially the class with a large number of students
might find it difficult to have the attendance sheet being passed around the class. Thus, face
recognition student attendance system is proposed in order to replace the manual signing of the
presence of students which are burdensome and causes students get distracted in order to sign for
their attendance. Furthermore, the face recognition based automated student attendance system
able to overcome the problem of fraudulent approach and lecturers does not have to count the
number of students several times to ensure the presence of the students.
Hence, there is a need to develop a real time operating student attendance system which means the
identification process must be done within defined time constraints to prevent omission. The
extracted features from facial images which represent the identity of the students have to be
consistent towards a change in background, illumination, pose and expression. High accuracy and
fast computation time will be the evaluation points of the performance.
1.3 Motivation
The main motivation for this project was the slow and inefficient traditional manual attendance
system. So, why not make it automated fast and much efficiently. Also, such face detection
techniques are in use by the department of a criminal investigation where the usage of CCTV
footages and detecting the faces from the crime scene and comparing them with criminal database
to recognize them. It is also becoming as a feature of daily life in China, where authorities are
using it on the streets, in subway stations, and at airports.
We undertook this project to lessen the workload of the faculties and also minimize errors such as
a student’s missing their name during the roll call or his name being missed in the attendance sheet.
This will also help us tackle fraudulence by students who give proxies.

16. 5
1.4 Aims and Objectives
The objective of this project is to develop face recognition based automated student attendance
system. Expected achievements in order to fulfil the objectives are:
 To detect the face segment from the video frame.
 To extract the useful features from the face detected.
 To classify the features in order to recognize the face detected.
 To record the attendance of the identified student.
Fig.1.1 from [1]

17. 6
1.5 ProjectDescription
All the students of the class must register themselves by entering the required details and then their
images will be captured and stored in the dataset. During each session, faces will be detected from
live streaming video of classroom. The faces detected will be compared with images present in the
dataset. If match found, attendance will be marked for the respective student. At the end of each
session, list of absentees will be mailed to the respective faculty handling the session
1.6 System Requirements
The following components have been chosen for the development of this project:
i. Operating System: Windows/ Linux
ii. Software specifications:
o Python IDLE: IDLE is an integrated development environment for editing and
running python2.x or python 3 programs. Where we can see or check the output.
o Python: Python is a programming language. Which
has easy syntaxes to read that allows fewer lines of
code to the programmers. This language is also
suitable for other customized applications.

o PHP: PHP (Hypertext Pre-processor), It is backend language used for the
development of Web Application.
o HTML: HTML stands for Hypertext Mark-up Language.
HTML is used for creating web applications With
Cascading Style Sheets and JavaScript.

o AWS Cloud: In the Amazon Web Service Cloud “S3”
(simple storage service) is used to store the captured
images, those captured images are analysed and compared
using the dataset for cloud computing.

iii. Hardware specifications:
o Camera (Webcam)
o Laptop/ Computing unit

18. 7
Chapter 2
LITERATURE REVIEW
Arun Katara et al. (2017) mentioned disadvantages of RFID (Radio Frequency Identification) card
system, fingerprint system and iris recognition system. RFID card system is implemented due to
its simplicity. However, the user tends to help their friends to check in as long as they have their
friend’s ID card. The fingerprint system is indeed effective but not efficient because it takes time
for the verification process so the user has to line up and perform the verification one by one.
However, for face recognition, the human face is always exposed and contain less information
compared to iris. Iris recognition system which contains more detail might invade the privacy of
the user. Voice recognition is available, but it is less accurate compared to other methods. Hence,
face recognition system is suggested to be implemented in the student attendance system.
Authors in [3] proposed a model of an automated attendance system. The model focuses on how
face recognition incorporated with Radio Frequency Identification (RFID) detect the authorized
students and counts as they get in and get out form the classroom. The system keeps the authentic
record of every registered student. The system also keeps the data of every student registered for
a particular course in the attendance log and provides necessary information according to the need.
A plan and execution of a remote iris acknowledgment participation in the board framework. Iris’s
acknowledgment check is one of the most dependable individual recognizable proof techniques in
biometrics. This framework-based biometrics and remote method takes care of the issue of
misleading participation and the issue of laying the relating system.
Strong Face Recognition by means of Adaptive Sparse Representation. Inadequate Representation
(or coding) based Classification (SRC) has increased incredible accomplishment in face
acknowledgment as of late

19. 8
Chapter 3
Image Pre-Processing
Image pre-processing is the term for operations on images at the lowest level of abstraction.
These operations do not increase image information content but they decrease it if entropy is
an information measure. The aim of pre-processing is an improvement of the image data that
suppresses undesired distortions or enhances some image features relevant for further
processing and analysis task.
Think of image pre-processing as a sound system with a range of controls, such as raw sound with
no volume controls; volume control with a simple tone knob; volume control plus treble, bass, and
mid; or volume control plus a full graphics equalizer, effects processing, and great speakers in an
acoustically superior room
3.1 Problems to Solve During Image Pre-Processing
Image pre-processing by following the vision pipelines of four fundamental families of feature
description methods
 Local Binary Descriptors (LBP, ORB, FREAK, others)
 Spectra Descriptors (SIFT, SURF, others)
 Basis Space Descriptors (FFT, wavelets, others)
 Polygon Shape Descriptors (blob object area, perimeter, centroid)
Fig.3.1 from [2]

20. 9
lists common image pre-processing operations, with examples from each of the four descriptor
families, illustrating both differences and commonality among these image pre-processing steps,
which can be applied prior to feature description.
3.2 Corrections
During image pre-processing, there may be artifacts in the images that should be corrected prior
to feature measurement and analysis. Here are various candidates for correction.
 Sensor corrections. Discussed in Chapter 1, these include dead pixel correction, geometric
lens distortion, and vignetting.
 Lighting corrections. Lighting can introduce deep shadows that obscure local texture and
structure; also, uneven lighting across the scene might skew results. Candidate correction
methods include rank filtering, histogram equalization, and LUT remap.
 Noise. This comes in many forms, and may need special image pre-processing. There are
many methods to choose from, some of which are surveyed in this chapter.
 Geometric corrections. If the entire scene is rotated or taken from the wrong perspective,
it may be valuable to correct the geometry prior to feature description. Some features are
more robust to geometric variation than others
 Color corrections. It can be helpful to redistribute color saturation or correct for
illumination artifacts in the intensity channel. Typically color hue is one of the more
difficult attributes to correct, and it may not be possible to correct using simple gamma
curves and the sRGB color space. We cover more accurate colorimetry methods later in
this chapter.
3.3 Enhancements
Enhancements are used to optimize for specific feature measurement methods, rather than fix
problems. Familiar image processing enhancements include sharpening and color balancing. Here
are some general examples of image enhancement with their potential benefits to feature
description.
 Scale-space pyramids. When a pyramid is constructed using an octave scale and pixel
decimation to sub-sample images to create the pyramid, sub-sampling artifacts and jagged
pixel transitions are introduced. Part of the scale-space pyramid building process involves

21. 10
applying a Gaussian blur filter to the sub-sampled images, which removes the jagged
artifacts.
 Illumination. In general, illumination can always be enhanced. Global illumination can be
enhanced using simple LUT remapping and pixel point operations and histogram
equalizations, and pixel remapping. Local illumination can be enhanced using gradient
filters, local histogram equalization, and rank filters.
 Blur and focus enhancements. Many well-known filtering methods for sharpening and
blurring may be employed at the pre-processing stage. For example, to compensate for
pixel aliasing artifacts introduced by rotation that may manifest as blurred pixels which
obscure fine detail, sharpen filters can be used to enhance the edge features prior to gradient
computations. Or, conversely, the rotation artifacts may be too strong and can be removed
by blurring.

22. 11
Chapter 4
PROJECTBACKGROUND
4.1 Face Detection
Face detection is a computer technology being used in a variety of applications that identifies
human faces in digital images. Face detection also refers to the psychological process by which
humans locate and attend to faces in a visual scene.
4.1.1 HaarCascade
So, what is Haar Cascade? It is an Object Detection Algorithm used to identify faces in an image
or a real time video. The algorithm uses edge or line detection features proposed by Viola and
Jones in their research paper “Rapid Object Detection using a Boosted Cascade of Simple
Features” published in 2001. The algorithm is given a lot of positive images consisting of faces,
and a lot of negative images not consisting of any face to train on them.
Fig.4.1 Block diagram [1]

23. 12
4.1.2 Feature ofHaar Cascade
Fig 4.2 The rectangle on the left is a sample representation of an image with pixel values 0.0 to
1.0. The rectangle at the center is a haar kernel which has all the light pixels on the left and all the
dark pixels on the right. The haar calculation is done by finding out the difference of the average
of the pixel values at the darker region and the average of the pixel values at the lighter region. If
the difference is close to 1, then there is an edge detected by the haar feature
A sample calculation of Haar value from a rectangular image section has been shown here. The
darker areas in the haar feature are pixels with values 1, and the lighter areas are pixels with values
0. Each of these is responsible for finding out one particular feature in the image. Such as an edge,
a line or any structure in the image where there is a sudden change of intensities. For ex. in the
image above, the haar feature can detect a vertical edge with darker pixels at its right and lighter
pixels at its left.
The objective here is to find out the sum of all the image pixels lying in the darker area of the
haar feature and the sum of all the image pixels lying in the lighter area of the haar feature. And
then find out their difference. Now if the image has an edge separating dark pixels on the right and
light pixels on the left, then the haar value will be closer to 1. That means, we say that there is an
edge detected if the haar value is closer to 1. In the example above, there is no edge as the haar
value is far from 1.
Fig.4.2 Haar Cascade [1]

24. 13
4.2 APPROACHTO SOLVE THE PROBLEM
The system consists of a camera that records the video of the class captures the images of the
person then the image is sent to the image enhancement module. After enhancement, the image
comes in the Face Detection and Recognition modules and then the attendance is marked. At the
time of enrolment, samples of face images of individual persons are stored in the Face database.
● Face Detection:
● Face Recognition:
● The face recognition systems can operate basically in two modes
1. Verification or authentication of a facial image
2. Identification of facial recognition
4.2.1 Algorithm Used:
Local Binary Pattern (LBP) is a simple yet very efficient texture operator which labels the pixels
of an image by thresholding the neighbourhood of each pixel and considers the result as a binary
number.
4.2.2 Extracting the Histograms:
Now, using the image generated in the last step, we can use the Grid X and Grid Y parameters to
divide the image into multiple grids, as can be seen in the following image: Based on the image
above, we can extract the histogram of each region as follows: As we have an image in grayscale,
each histogram (from each grid) will contain only 256 positions (0~255) representing the
occurrences of each pixel intensity.
4.2.3 Performing face recognition:
In this step, the algorithm is already trained. Each histogram created is used to represent each
image from the training dataset. So, to find the image that matches the input image we just need
to compare two histograms and return the image with the closest histogram. We can assume
that the algorithm has successfully recognized if the confidence is lower than the threshold
defined.

25. 14
4.3 FLOW DIAGRAM
Fig.4.3 Diagram of real time face recognition system [6]

26. 15
4.4 Some face detectionmodels
Face detection is one of the most fundamental aspects of computer vision. It is the base of many
further studies like identifying specific people to marking key points on the face.
Some pre-trained models like
Haar Cascades
It is super fast to work with and like the simple CNN, it extracts a lot of features from images. The
best features are then selected via Adaboost. This reduces the original 160000+ features to 6000
features. But applying all these features in a sliding window will still take a lot of time. So they
introduced a Cascade of Classifiers, where the features are grouped. If a window fails at the first
stage, these remaining features in that cascade are not processed. If it passes then the next feature
is tested and the same procedure is repeated. If a window can pass all the features then it is
classified as a face region.
Dlib Frontal Face Detector
Dlib is a C++ toolkit containing machine learning algorithms used to solve real-world problems.
Although it is written in C++ it has python bindings to run it in python. It also has the great facial
landmark keypoint detector which I used in one of my earlier articles to make a real-time gaze
tracking system.
The frontal face detector provided by dlib works using features extracted by Histogram of Oriented
Gradients (HOG) which are then passed through an SVM. In the HOG feature descriptor, the
distribution of the directions of gradients is used as features. Moreover, Dlib provides a more
advanced CNN based face detector
Fig.4.4 Haar Cascade [7]

27. 16
MTCNN
It was introduced by Kaipeng Zhang, et al. in 2016 in their paper, “Joint Face Detection and
Alignment Using Multi-task Cascaded Convolutional Networks.” It not only detects the face but
also detects five key points as well. It uses a cascade structure with three stages of CNN. First,
they use a fully convolutional network to obtain candidate windows and their bounding box
regression vectors, and the highly overlapped candidates are overlapped using on-maximum
suppression (NMS). Next, these candidates are passed to another CNN which rejects a large
number of false positives and performs calibration of bounding boxes. In the final stage, the facial
landmark detection is performed.
DNN Face Detectorin OpenCV
It is a Caffe model which is based on the Single Shot-Multibox Detector (SSD) and uses ResNet-
10 architecture as its backbone. It was introduced post OpenCV 3.3 in its deep neural network
module. There is also a quantized Tensorflow version that can be used but we will use the Caffe
Model
Fig.4.5 MTCNN [8]
Fig.4.6 DNN [9]

28. 17
YOLO
YOLO algorithm employs convolutional neural networks (CNN) to detect objects in real-time.
The algorithm requires only a single forward propagation through a neural network to detect
objects. The CNN is used to predict various class probabilities and bounding boxes simultaneous ly.
YOLO divides every image into a grid of S x S and every grid predicts N bounding boxes and
confidence. The confidence reflects the precision of the bounding box and whether the bounding
box in point of fact contains an object in spite of the defined class. YOLO even forecasts the
classification score for every box for each class.
Fig.4.7 Working of YOLO model[13]

29. 18
4.5 Face Recognition
Face recognition in simple terms is the process of identifying an individual from their face. The
actions involved when a person tries to identify an individual is to capture an image of the
individual’s face with the eyes, analyse the image of the face, and compare patterns of the
individual’s face with the faces that are already known.
The series of problems in identifying a face are:
● To find all the faces in an image
● To focus on each face and be able to detect a face even if it is turned sideways or in a tilted
orientation
● To be able to pick out unique features from each face that can be used to tell apart one face
from another
● To compare the unique features of each face to all the people we already know to
determine the person’s name.
Similar process is used by computers when it comes to face recognition.
4.5.1 Stagesin Face Recognition
1. Face Detection: This stage detects and locates human faces in a frame. Both Machine
learning or Deep learning algorithms can be used for face detection. The facial features
extracted by the algorithms are then used in a classifier to determine if it is a face.
2. Posing and Projecting Faces: This stage deals with being able to understand that even if
a face is turned in a weird direction or in bad lighting, it is still the same person. In order
to accomplish this, 68 specific points (called face landmarks) that exist on every face are
found out and the image is sheared so that the eyes and mouth are centered as best as
possible.
3. Encoding: This step deals with extraction of a few basic measurements from each face.
From these measurements, a pre trained CNN is used to generate encodings.
4. Comparing the face: Finally, the encoding of a face is compared to all the encodings of
an already known face to determine the person’s name.

30. 19
4.5.2 FaceLandmark Estimation
The basic idea in Face landmark Estimation is to come up with 68 specific points (called
landmarks) that exist on every face — the top of the chin, the outside edge of each eye, the inner
edge of each eyebrow, etc. Then a machine learning algorithm finds these 68 specific points on
any face. Then the image is sheared so that the eyes and mouth are centered as best as possible.
The aim is that no matter how the face is turned, the facial features should be centered.
4.5.3 Encoding of faces
Deep Convolutional Neural Networks are trained to generate 128 measurements(embedding) for
each face. The embeddings are basically numbers generated by the neural network. The embedding
of a face is compared to all the embedding of an already known face to determine the person’s
name.
Fig.4.8 Flowchart of stages of face recognition. [14]
Fig.4.9 68 face landmarks [7]

31. 20
Chapter 5
METHODOLOGY
5.1 Methodology
The approach performs face recognition-based student attendance system. The methodology flow
begins with the capture of image by using simple and handy interface, followed by pre-processing
of the captured facial images, then feature extraction from the facial images, subjective selection
and lastly classification of the facial images to be recognized. Both LBP and PCA feature
extraction methods are studied in detail and computed in this proposed approach in order to make
comparisons. LBP is enhanced in this approach to reduce the illumination effect. An algorithm to
combine enhanced LBP and PCA is also designed for subjective selection in order to increase the
accuracy
All the students of the class must register themselves by entering the required details and then their
images will be captured and stored in the dataset. During each session, faceswill be detected from
live streaming video of classroom. Thefaces detected will be compared with images present in
thedataset. If match found, attendance will be marked for the respective student. At the end of each
session, list of absentees will be mailed to the respective faculty handling the session
5.1.1 Input Images
Although our own database should be used to design real time face recognition student attendance
system, the databases that are provided by the previous researchers are also used to design the
system more effectively, efficiently and for evaluation purposes.
For our own database, the images of students are captured by using laptop built in camera and
mobile phone camera. Each student provided four images, two for training set and two for testing
set. The images captured by using laptop built in camera are categorized as low quality images,
whereas mobile phone camera captured images are categorized as high quality images. The high-
quality images consist of seventeen students while low quality images consists of twenty-six
students.
5.1.2 Limitations of the Images
The input image for the proposed approach has to be frontal, upright and only a single face.
Although the system is designed to be able to recognize the student with glasses and without
glasses, student should provide both facial images with and without glasses to be trained to increase
the accuracy to be recognized without glasses. The training image and testing image should be
captured by using the same device to avoid quality difference. The students have to register in
order to be recognized. The enrolment can be done on the spot through the user-friendly interface.

32. 21
5.2 Face Detection
Viola-Jones object detection framework will be used to detect the face from the video camera
recording frame. The working principle of Viola-Jones algorithm is mentioned. The limitation of
the Viola-Jones framework is that the facial image has to be a frontal upright image, the face of
the individual must point towards the camera in a video frame.
5.3 Pre-Processing
Testing set and training set images are captured using a camera. There are unwanted noise and
uneven lighting exists in the images. Therefore, several pre-processing steps are necessary before
proceeding to feature extraction.
Pre-processing steps that would be carried out include scaling of image, median filtering,
conversion of colour images to grayscale images and adaptive histogram equalization.
5.3.1 Scaling ofImage
Scaling of images is one of the frequent tasks in image processing. The size of the images has to
be carefully manipulated to prevent loss of spatial information. (Gonzalez, R. C., & Woods, 2008),
In order to perform face recognition, the size of the image has to be equalized. This has become
crucial, especially in the feature extraction process, the test images and training images have to be
in the same size and dimension to ensure the precise outcome. Thus, in this proposed approach test
images and train images are standardize at size 250 × 250 pixels.
5.3.2 MedianFiltering
Median filtering is a robust noise reduction method. It is widely used in various applications due
to its capability to remove unwanted noise as well as retaining useful detail in images. Since the
colour images captured by using a camera are RGB images, median filtering is done on three
different channels of the image. If the input image is a grayscale image, then the median filtering
can be performed directly without separating the channels
5.3.3 Conversionto Grayscale Image
Camera captures color images, however the proposed contrast improvement method CLAHE can
only be performed on grayscale images. After improving the contrast, the illumination effect of
the images able to be reduced. LBP extracts the grayscale features from the contrast improved
images as 8 bit texture descriptor (Ojala, T. et al., 2002).Therefore, color images have to be
converted to grayscale images before proceeding to the later steps. By converting color images to
grayscale images, the complexity of the computation can be reduced resulting in higher speed of
computation (Kanan and Cottrell, 2012). Figure shows the conversion of images to grayscale
image.
Fig.5.1 conversion of greyscale [3]

33. 22
5.3.4 ContrastLimited Adaptive Histogram Equalization
Histogram equalization or histogram stretching is a technique of image contrast
Enhancement. The contrast improvement is usually performed on the grayscale images. Image
contrast is improved by stretching the range of its pixel intensity values to span over the desired
range of values, between 0 and 255 in grayscale. The reason that Contrast Limited Adaptive
Histogram Equalization (CLAHE) is used instead of histogram equalization is because histogram
equalization depends on the global statistics. Hence, it causes over enhancement of some parts of
image while other parts are not enhanced properly. This distorts the features of the image. It is a
serious issue because the features of the image have to be extracted for the face recognition
5.4 Block Diagram
Fig.5.2 Flow of the Proposed Approach (Training Part) [8]

34. 23
Fig.5.3 Flow of the Proposed Approach (Recognition Part) [8]

35. 24
Chapter 6
IMPLEMENTATION
We propose a low-cost solution in recording student attendance by employing face detection
technique. Our solution consists of four stages: Data set creation, face detection, Face recognition
and output. We named our proposed solution Image Based Attendance System. The system is
designed to improve the time efficiency and to reduce the staff-workload, which would ultimately
improve the reliability of the attendance record.
6.1 DatasetCreation
Images of students are captured using a web cam. Multiple images of single student will be
acquired with varied gestures and angles. These images undergo pre-processing. The images are
cropped to obtain the Region of Interest (ROI) which will be further used in recognition process.
Next step is to resize the cropped images to particular pixel position. Then these images will be
converted from RGB to gray scale images. And then these images will be saved as the names of
respective student in a folder.
6.2 Face Detection
Face detection here is performed using Haar-Cascade Classifier with OpenCV. Haar Cascade
algorithm needs to be trained to detect human faces before it can be used for face detection. This
is called feature extraction. The haar cascade training data used is an xml file haarcascade_
frontalface default. The haar features shown in Fig. will be used for feature extraction.
Fig.6.1 Haar Feature [1]

36. 25
Here we are using detectMultiScale module from OpenCV. This is required to create a rectangle
around the faces in an image. It has got three parameters to consider- scaleFactor, minNeighbors,
minSize. scaleFactor is used to indicate how much an image must be reduced in each image scale.
minNeighbors specifies how many neighbors each candidate rectangle must have. Higher values
usually detects less faces but detects high quality in image. minSize specifies the minimum object
size. By default, it is (30,30) [8]. The parameters used in this system is scaleFactor and
minNeighbors with the values 1.3 and 5 respectively.
6.3 Face Recognition
Face recognition process can be divided into three stepsprepare training data, train face recognizer,
prediction. Here training data will be the images present in the dataset. They will be assigned with
a integer label of the student it belongs to. These images are then used for face recognition. Face
recognizer used in this system is Local Binary Pattern Histogram. Initially, the list of local binary
patterns (LBP) of entire face is obtained. These LBPs are converted into decimal number and then
histograms of all those decimal values are made. At the end, one histogram will be formed for each
images in the training data. Later, during recognition process histogram of the face to be
recognized is calculated and then compared with the already computed histograms and returns the
best matched label associated with the student it
belongs to [9].
6.4 Output/Attendance Updation
After face recognition process, the recognized faces will be marked as present in the excel sheet
and the rest will be marked as absent and the list of absentees will be mailed to the respective
faculties. Faculties will be updated with monthly attendance sheet at the end of every month.

37. 26
6.5 Algorithm of working process
Fig.6.2 Algorithm [5]

38. 27
Chapter 7
RESULTS AND DISCUSSION
7.1 Results
The users can interact with the system using a GUI. Here users will be mainly provided with three
different options such as, Enter ID, Enter name, Notification, Take Images, Train Images, Track
Images, Attendance, and Quit. The students are supposed to enter all the required details in the
student registration form. After clicking on register button, the web cam starts automatically and
window as shown in Fig. pops up and starts detecting the faces in the frame. Then it automatically
starts clicking photos until 100 samples are collected or CRTL+Q is pressed. These images then
will be pre-processed and stored in training images folder. Click on the train image button makes
to process into the image. From next time onward no registration is required, simple press the
button Track Image it automatically recognize the student and make an entry on attendance sheet.
Lastly using quit button, it closes the current window.

39. 28
Fig.7.1 Results

40. 29
7.2 Discussion
This proposed approach provides a method to perform face recognition for student attendance
system, which is based on the texture-based features of facial images. Face recognition is the
identification of an individual by comparing his/her real-time captured image with stored images
in database of that person. Thus, training set has to be chosen based on the latest appearance of an
individual other than taking important factor for instance illumination into consideration.
Viola-Jones object detection framework is applied in this approach to detect and localize the face
given a facial image or provided a video frame. From the detected face, an algorithm that can
extract the important features to perform face recognition is designed.
Some pre-processing steps are performed on the input facial image before the features are
extracted. Median filtering is used because it is able to preserve the edges of the image while
removing the image noises. The facial image will be scaled to a suitable size for standardizing
purpose and converted to grayscale image if it is not a grayscale image because CLAHE and LBP
operator work on a grayscale image.
Contrast Limited Adaptive Histogram Equalization (CLAHE) is proposed in pre-processing in
order to improve the image contrast and reduce the illumination effect. Most of the previous
researchers have implemented histogram equalization in their approach

41. 30
7.3 Drawbacks
The drawbacks associated with the system proposed in this project are summarized as below:
 Low lighting conditions: Images acquired in an environment with low lighting condition
leads to poor performance of the system.
 Image quality: The quality of the reference image plays an important role in the
identification process. If the resolution of the said image is not high enough, it can cause
cameras to be tricked into believing that the person being scanned is not the same as in the
photo. An easy solution is to ensure that both the reference images and scanning are
performed by similar cameras.
 Storage: Depending on the quality of the input data, a system would need an appropriate
amount of storage. This could be troublesome if the data collected is of high quality and
requires large amounts of storage space especially for events with a large expected
attendance.
 Angles: Many non-premium facial recognition systems cannot account for faces that are
captured at angles other than straight into the capturing camera. The disadvantage of this
is that it makes the attendance marking process slower and less efficient.

42. 31
Chapter 8
CONCLUSION & FUTURE SCOPE
8.1 Conclusion
Recording class attendance is a common practice in many educational institutions; particularly for
those who enforce compulsory attendance policy. Roll-call and sign-in-sheet are widely accepted
as the conventional practices for recording student attendance in a classroom. However, there are
number of issues introduced in these conventional practices, such as time inefficiency, labour
intensiveness, human error, and distraction.
In this paper, we proposed Image Based Attendance System as a low-cost solution in recording
student attendance by utilising face detection technique. students face are automatically located,
and students then registered their attendance by simply identifying their face on the records.
We have used face recognition concept to mark the attendance of student and make the system
better. The system performs satisfactory in different poses and variations. In future this system
need be improved because these system sometimes fails to recognize students from some distance,
also we have some processing limitation, working with a system of high processing may result
even better performance of this system.
8.2 Future Scope
Practically all academic institutions require attendance record of students and maintaining
attendance physically can be hectic as well as time consuming task. Hence maintaining attendance
automatically with the help of face recognition will be exceptionally useful and less prone to
mistakes or errors as compared to manual procedure. This will also reduce the manipulation of
attendance record done by students and reduces time consumption too.
We can use our system at public places with CCTV camera. As well as railway station, bus stations,
crowded area, to reduce crimes. We can store each person data in database with the help of Aadhar
card. Therefore, identification of person gets easy. Our system is helpful for police.

43. 32
References
1. Chin, H. (2018). Face recognition based automated student attendance system (Doctoral
dissertation, UTAR).
2. Krig, Scott. Computer vision metrics: Survey, taxonomy, and analysis. Springer Nature, 2014.
3. Cuimei, L., Zhiliang, Q., Nan, J., & Jianhua, W. (2017, October). Human face detection algorithm
via Haar cascade classifier combined with three additional classifiers. In 2017 13th IEEE
International Conference on Electronic Measurement & Instruments (ICEMI) (pp. 483-487). IEEE.
4. Sidney Fussell. (2018). NEWS Facebook's New Face Recognition Features: What We Do (and Don't) Know.
[online] Available at: https://gizmodo.com/facebooks-new-face-recognition-features-what-we-do-an-
1823359911 [Accessed 25 Mar. 2018].
5. Tabora, V. (2011). Face Detection Using OpenCV With Haar Cascade Classifiers. línea]. Available:
https://becominghuman. ai/face-detection-using-opencv-with-haarcascade-classifiers-
941dbb25177.[Último acceso: 12 08 2019].
6. Kar, N., Debbarma, M. K., Saha, A., & Pal, D. R. (2012). Study of implementing automated
attendance system using face recognition technique. International Journal of computer and
communication engineering, 1(2), 100.
7. Kim, M., Lee, D., & Kim, K. Y. (2015). System architecture for real-time face detection on analog
video camera. International Journal of Distributed Sensor Networks, 11(5), 251386.
8. Kaziakhmedov, Edgar, Klim Kireev, Grigorii Melnikov, Mikhail Pautov, and Aleksandr Petiushko.
"Real-world attack on MTCNN face detection system." In 2019 International Multi-Conference on
Engineering, Computer and Information Sciences (SIBIRCON), pp. 0422-0427. IEEE, 2019.
9. Chintalapati, S., & Raghunadh, M. V. (2013, December). Automated attendance management
system based on face recognition algorithms. In 2013 IEEE International Conference on
Computational Intelligence and Computing Research (pp. 1-5). IEEE.
10.Raja, R. "Face recognition using OpenCV and Python: A beginner’s guide. Retrieved May 26, 2020."
(2020).
11.D'Souza, J. W., Jothi, S., & Chandrasekar, A. (2019, March). Automated attendance marking and
management system by facial recognition using histogram. In 2019 5th International Conference on
Advanced Computing & Communication Systems (ICACCS) (pp. 66-69). IEEE.
12.Balcoh, N. K., Yousaf, M. H., Ahmad, W., & Baig, M. I. (2012). Algorithm for efficient attendance
management: Face recognition based approach. International Journal of Computer Science Issues
(IJCSI), 9(4), 146.
13.Tian, Y., Yang, G., Wang, Z., Wang, H., Li, E., & Liang, Z. (2019). Apple detection during different
growth stages in orchards using the improved YOLO-V3 model. Computers and electronics in
agriculture, 157, 417-426.
14.Indra, Evta, Muhammad Yasir, Andrian Andrian, Delima Sitanggang, Oloan Sihombing, Saut
Parsoran Tamba, and Elviana Sagala. "Design and Implementation of Student Attendance System
Based on Face Recognition by Haar-Like Features Methods." In 2020 3rd International Conference
on Mechanical, Electronics, Computer, and Industrial Technology (MECnIT), pp. 336-342. IEEE,
2020.
15.Chin, H., Cheah, K. H., Nisar, H., & Yeap, K. H. (2019, September). Enhanced Face Recognition
Method Based On Local Binary Pattern and Principal Component Analysis For Efficient Class
Attendance System. In 2019 IEEE International Conference on Signal and Image Processing
Applications (ICSIPA) (pp. 23-28). IEEE.
16.Zhang, S., Wen, L., Shi, H., Lei, Z., Lyu, S., & Li, S. Z. (2019). Single-shot scale-aware network for
real-time face detection. International Journal of Computer Vision, 127(6), 537-559.
17.Fleischer, Rachel S. "BIAS IN, BIAS OUT: WHY LEGISLATION PLACING REQUIREMENTS ON
THE PROCUREMENT OF COMMERCIALIZED FACIAL RECOGNITION TECHNOLOGY MUST
BE PASSED TO PROTECT PEOPLE OF COLOR." Public Contract Law Journal 50, no. 1 (2020):
63-89.
18.Mahesh, PC Senthil, K. Sasikala, and M. Rudra Kumar. "Face Recognition Based Automated
Student Attendance System using Deep Learning." International Journal 9, no. 3 (2020).

44. 33
19.Al Sheikh, Rahmah, Raghad Al-Assami, Mariam Al-Bahar, Muntaha Al Suhaibani, Mutasem
Alsmadi, Muneerah Alshabanah, Daniah Alrajhi et al. "Developing and implementing a barcode
based student attendance system." International Research Journal of Engineering and Technology
(IRJET) Volume 6 (2019).
20.Hapani, Smit, Nandana Prabhu, Nikhil Parakhiya, and Mayur Paghdal. "Automated Attendance
System Using Image Processing." In 2018 Fourth International Conference on Computing
Communication Control and Automation (ICCUBEA), pp. 1-5. IEEE, 2018.