International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169

Volume: 5 Issue: 2 273 276

_______________________________________________________________________________________________

Attendance System Using Face Recognition and Class Monitoring System

Prof. Arun Katara1, Mr. Sudesh V. Kolhe2, Mr. Amar P. Zilpe3, Mr. Nikhil D. Bhele4, Mr. Chetan J. Bele5
Department of Electronics
DMIETR, Sawangi (M), Wardha.
Email: arunkatara@gmail.com, sudeshkolhe@gmail.com, zilpe.amar00@gmail.com, nikhil.bhele03@gmail.com,
chetanbele1994@gmail.com.

Abstract In this paper, we propose a system that takes the attendance of students in the lecture. This system takes the attendance
automatically using face recognition. However, it is difficult to estimate the attendance exactly using each result of face recognition
independently because the face detection rate is not sufficiently high. In our paper, we propose a method for estimating the attendance exactly
using all the results of face recognition obtained by continuous observation. Continuous observation improves the performance for the
estimation of the attendance. We constructed the attendance system based on face recognition, and applied the system to classroom lecture. In
our system, we are using raspberry pi. we use OpenCv library which is installed in pi for face detection and recognition. The camera is
connected to raspberry pi and student database is stored in the pi. With the help of this system time will reduce and attendance will be marked.
In this paper first review similar works in the field of attendance system and recognition of face. Then, it showing our system structure and
plan. At the last, experiments are implemented to provide as manifest to support our plan. The result shows that uninterrupted observation
improved the performance for the approximation of the attendance.

Keywords- Face Recognition; Identification; Monitoring.

__________________________________________________*****_________________________________________________

thus it is time consuming and may distract the attention of the

I. INTRODUCTION student.
The Automated Attendance System is the progress that has In [2] the authors have proposed Automatic Control of
taken place in the field of automation replacing the attendance Students Attendance in Classroom Using RFID. In this system
marking activity traditionally. Automated Attendance Systems in which students carry a RFID tag type ID card and they need
are generally bio-metric , RFIDcard and MATLAB based. to place that on the card reader to record his attendance.
These systems are widely used in different organizations. RS232 is used to connect the system with computer and stored
Traditional method of attendance marking is very time the recorded attendance from the database. This system may
consuming and becomes difficult when the strength is high. give rise to the problem of fraudulent approach. An
Automation of Attendance System has boundary over unauthorized students may make use of RFID card and enter
conventional method as it saves time and also can be used for into the organization.
security purposes. This also helps to prevent fake attendance. In [3] the authors have proposed Wireless Fingerprint
Computer visual class library (OpenCV) which source Attendance Management System. This system uses iris
code is open, develops for application developers, including recognition system that does capturing the image of iris
plenty of functions which used to tackle the common
recognition, extraction, storing and matching. But the
problems in computer visual field, for example, motion
difficulty occurs to lay the transmission lines in the places
analysis and tracing, face recognition, three dimensional
where the quality of topography is poor. In [4] authors have
reconstruction and target recognition.
consider a system based on real time face recognition which is
We propose a method that take the attendance using
recognition of face based on continuous observation. In this reliable, secure and fast which needs improvement in different
paper, our purpose is to obtain the attendance, positions and lighting conditions.
images of students face, which are useful information in the III. PROPOSED MODEL
classroom lecture.
The system architecture is as shown in Figure 1. The
II. LITERATURE SURVEY proposed attendance system using face recognition system and
In [1] the authors have proposed a Automated Attendance class monitoring system. Camera is present inside the
Management System based On Face Recognition Algorithm. classroom where the students are seated. The camera
This system, which is based on face detection and recognition capturing the images of students from the video frame. A face
algorithm, automatically detects the student when he enters recognition system which automatically identifies and verifies
the class room and mark the attendance by recognition him. the identity of person from video frame from video source.
This technique is to be use in order to handle the treats like We used OpenCv library that can be formulated as given
spoofing. The problem with this approach is that it capture images are identify or verify all the students seated in front of
only one student image at a time when he enter the classroom camera from the stored database of faces.
273
IJRITCC | February 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 2 273 276
_______________________________________________________________________________________________
Face recognition have more advance than the other system D. Database Development
which are given in the table I. The steps in the propose As we chose bio-metric based system Enrollment of every
attendance system using face recognition system and class individual is required. This database development phase
monitoring system shown in the Fig. 1. consists of image capture of every individual and extracting
the bio-metric aspect, in this case it is face, and later it is
A. Image Capture enhanced using pre-processing techniques and stored in the
The Camera is placed inside the classroom at a distance database. In our project we have taken the images of
individuals in different angles, different expressions and also
from the students and taking the video frame of the whole
in different lighting conditions. A database of 50 individuals
class. From the video frame it takes anyone frame to verify
(MySQL-database) with 20 images of each has been collected
and identify the student to mark his attendance.
for this project. Faces stored in the database shown in fig
below.
TABLE I
Disadvantages of various Attendance Systems

Types of system Disadvantages

1) RFID-Based Fraudulent usage.
2) Fingerprint-Based Time consuming for students to
wait and give their attendance.
3) Iris-Based Invades the privacy of the user.

Fig. 2 Captured Database

E. Feature Extraction and Classification

The performance of a Face Recognition system depends on
the feature extraction and their classification to get the
accurate results. Feature extraction is achieved using holistic
Fig. 1 System Architecture
techniques. In holistic technique we can use of dimensionality
B. Face Detection reduction before the classification. We compared the results
of different holistic approaches used for feature extraction and
A proper and efficient face detection always enhances the classification in real time scenario.
performance of face recognition systems. Various algorithms Principal Component Analysis (PCA) was the first
are proposed for face detection such as Face geometry based algorithm that represents the faces economically. In PCA the
methods, Feature Invariant methods, Machine learning based face images are represented using eigen faces and their
methods. Out of all these methods we are proposed a corresponding projections along each eigen face. Instead of
framework which gives a high detection rate and is also fast.
using all the all the dimensions of an image only meaningful
Our system is efficient for real time application as it is fast
dimensions are considered to represent the image.
and robust. Hence we chose OpenCv Library for the detection
Mathematically an image using PCA is represented as
of faces. We system gives better results in different lighting
conditions. =WY +
where is the face vector, Y is vector of eigen faces, W is the
C. Pre-processing feature vector, and is the average face vector.
The detected face is extracted and deals with to pre- Now that you have detected the face we can used that face
processing. This pre-processing step involves with histogram image for recognition how however if you try to simply
equalization of the extracted face image and is resized to perform face recognition directly on normal photo images you
100x100. Histogram Equalization is the widely used will probably gate less than 10% accuracy. The problem with
Histogram Normalization technique. This improves the image representation we are given is its high dimensionality.
contrast of the image as it stretches the range of the intensities Two-dimensional pxq grayscale images span a m=pq
in an image by making it more clear. dimensional vector space, so an image with 100x100 pixels
lies in a 10,000 dimensional image space already. We can
only make a decision if theres any variance in data, so what
we are looking for are the components that account for the
information. The Principal Component Analysis (PCA) is
274
IJRITCC | February 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 2 273 276
_______________________________________________________________________________________________
used to turn a set of possibly correlated variables into a Performance is also evaluated in terms of recognition rate,
smaller set of uncorrelated variables. The idea is that a high- distance, false positive rate, time taken for training. False
dimensional data is repeatedly described by correlated Positive Rates are calculated by considering 50 real time
variables and therefore only a few meaningful dimensions image frames. It has been observed that LBP based algorithm
account for most of the information. The PCA method finds gives least false positive rate and good recognition rate as it
the directions with the greatest variance in the data, called correctly differentiates between the unknown and known
principal components. faces. LDA can make correct discrimination between the
images only if the discrimination is provided in the database
Algorithmic Description (for example images at different lighting conditions). Distance
also plays as a criteria in this system as the video frames are
Let be a random vector with taken when all students are seating in the classroom. So the
observations . face region captured at about 4 feet and 7 feet gives better
results for LBPH and other algorithms respectively. For a
Training data of 500 images training time is calculated. LBP
1. Compute the mean
based algorithm requires minimum time for training where as
SVM and Bayes classifiers take time for training.

F. Post-processing
In the proposed system, after recognizing the faces of the
students, the information of students are updated into an
2. Compute the the Covariance Matrix S
MySQL database. The data of present student are then send to
server computer and stored so that authorised person can see
it.

IV. GRAPHICAL USER INTERFACE (GUI)

3. Compute the eigenvalues and eigenvectors of The GUI is developed using WampSever, it is a window
web development environment. It allow you to create web
application with Apache2, PHP and a MySQL database.
Alongeside, PhpMyAdmin allows you to manage easily your
databases.
The system provides the following functions.
4. Order the eigenvectors descending by their Name
eigenvalue. The principal components are the Date
eigenvectors corresponding to the largest Time
eigenvalues. Attendance: Present/Absent

The principal components of the observed vector are then Excel Sheet is generated when Recognition is completed.
given by: Figure 3 shows the extraction of face region and updating to
the database after pre-processing. Figure 3 shows the
recognition process. Post-processing step in-cludes updating
the excel sheet with students names who are present as shown
in Figure 4.
where .

The reconstruction from the PCA basis is given by:

where .

So Face Recognition involves in two stages, feature

extraction and classification. The above mentioned feature
extractors combined with classifiers are compared in various
real world scenarios such as lighting conditions, Unintentional Fig.3 Recognizing the Faces
facial feature changes (occluded faces), Expressions. System
275
IJRITCC | February 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________
International Journal on Recent and Innovation Trends in Computing and Communication ISSN: 2321-8169
Volume: 5 Issue: 2 273 276
_______________________________________________________________________________________________

Roll No. Name Attendance Class/Time

10 Sudesh Present 8.00AM

11 Nikhil Absent 8.00AM

13 Chetan Present 8.00AM

Fig.4 Attendance Excel Sheet.

V. CONCLUSION AND FUTURE WORK

Attendance system using face recognition and class
monitoring system based on face recognition techniques thus
proved to be time saving and secure. In real scenarios LBPH
outperforms algorithms with better recognition rate and low
false positive rate.
In future work is to improve recognition rate and Using
Raspberry pi the current project can be modified by infrared
camera interfacing , it can be used in smart surveillance
monitoring security system.

ACKNOWLEDGMENT

The authors would like to thank Prof. Arun Katara for his
admirable guidance and assistance during this project.

REFERENCES
[1] Shireesha Chintalapati, M.V. Raghunadh, Automated
Attendance Management System Based On Face
Recognition Algorithms,2013 IEEE International
Conference on Computational Intelligence and Computing
Research.
[2] Ruian Liu Mimi Zhang Shengtao Ma, Design of Face
Detection and Tracking System, 2010 IEEE 3rd
International Congress on Image and Signal Processing.
[3] M.C. Santana, M.H. Tejera, J.C. Gmez, Encara:real-time
detection of frontal faces, IEEE ICIP.
[4] Adam Schmidt, Andrzej Kasinski, The Performance of the
Haar Cascade Classifiers Applied to the Face and Eyes
Detection, Computer Recognition Systems 2.
[5] B. K. Mohamed and C. Raghu, Fingerprint attendance
system foe classroom need, in India conference, 2015
Annual IEEE, IEEE.
[6] Amit Kumar, P.K. Varma, Srinivas Perala, P. R. Chadha,
Automatic Attendance System By Visual Programming
Language LabVIEW, in IEEE International Conference On
Power Electronics, Intellegent Control and Energy System
2016.
[7] Francisco Silva, Victor Filipe, Antonio Pereira, Atomaatic
Control of Student Attendance In Classroom Using RFID,
2008 IEEE International Conference.
[8] Maddu Kamaraju, Penta Anil Kumar, Wireless Fingerprint
Attendance Management System, 2015 IEEE

276
IJRITCC | February 2017, Available @ http://www.ijritcc.org
_______________________________________________________________________________________