Вы находитесь на странице: 1из 6

International Journal of Electrical, Electronics and Data Communication, ISSN: 2320-2084,

Volume- 1, Issue- 4

REAL TIME FACE RECOGNITION SYSTEM FOR TIME AND


ATTENDANCE APPLICATIONS
1

APARNA BEHARA, 2M.V. RAGHUNADH

Department of E and CE, NIT Warangal, INDIA 506004


Email: behara.aparna@gmail.com, raghunadhmv@gmail.com

Abstract In this paper, we have proposed an automated Face Recognition System for Time and Attendance application.
The model is developed with the help of real time OpenCV library The proposed system comprised of using the Viola Jones
algorithm for detecting the human faces and then the detected face is resized to the required size, this resized face is further
processed by using linear stretch contrast enhancement and finally it is recognized using a simple PCA / LDA. Once
recognition is done, automatically attendance will be updated in an Excel Sheet along with his name, date and time. An html
file is automatically updated by our system so that a remote authenticated user can access the attendance file .Our system is
integrated to an Automatic Attendance Management System, with the help of which some post attendance works like stipend
amount calculation, viewing attendance report for the required date, calculating the number of hours a person is present in
the class, searching for a required person in the classroom etc. Spoofing which is a major threat for our system can be
avoided using Eye Blink Detector algorithm. Our system can automatically update the Database for the newly enrolled
persons. The proposed system can execute the instructions given orally to it (eliminating the need of a mouse interface) and
can speak back to us.
Keywords Face, Attendance, Eye blink, GUI

I.

The research going on for Face Detection and


feature extraction are as discussed below:

INTRODUCTION

Person Recognition is one of the emerging


research fields in computer vision. The major
application for person recognition is for Classroom
Attendance system, where the attendance is
automatically updated onto an attendance sheet and
also the IN/OUT timings of the Students can be
stored as a soft data for later records. This system is
developed in order to avoid the manual drudgery for
lecturers in entering the data daily while taking
attendance and also to avoid proxy.
They are several biometrics used for Person
Recognition like Iris, Fingerprints, Face etc. Since
Iris and Fingerprints are very short-distance
biometrics, but our application requires a person to be
at a medium distance from the camera, which is fixed
at the centre of the classroom near the black board, so
that the view of the camera covers the entire
classroom. As a worst case, our system should be
able to recognise a person who is sitting at the last
bench, which might not be possible by using Iris or
Fingerprints as a biometric. Hence we go for a
medium range Biometric i.e., Face with the help of
which we can recognise a person and mark his
attendance.

A. Face Detection Methods:


The different techniques used for face detection
are as classified as shown below:
Knowledge Based Method
Feature Invariant Method
Template Matching Method
Appearance Based Method
B. Face Feature Extraction Methods:
The techniques used for Face Recognition can
be divided into two main categories:

Holistic Approach

Feature-based Approach
In the holistic approach, the whole face is taken as
input for recognition purpose. We can use PCA
(Principal Component Analysis), for reducing the
dimensionality of the data by projecting, it onto a
lower dimensional subspace [1]. We can go for LDA
(Linear Discriminate Analysis), where the
dimensionality reduction takes place such that the
within class variance is reduced and between class
variance is maximised [2].
In Feature based Approach, local features on face
such as eyes and nose are detected and based upon
which recognition is performed [3].

The rest of the paper is organized as follows: The


detailed literature survey is given in section II, the
proposed model is explained in section III, the
experimental results are shown in section IV, and
finally conclusion and future scope are discussed in
section V.
II.

III.

PROPOSED MODEL

The proposed model for Face Recognition system is


as shown in Fig.1.

LITERATURE SURVEY

Real Time Face Recognition System For Time and Attendance Applications
6

International Journal of Electrical, Electronics and Data Communication, ISSN: 2320-2084,

Fig. 1. Face Recognition Model

Real Time Face Recognition System For Time and Attendance Applications
7

Volume- 1, Issue- 4

International Journal of Electrical, Electronics and Data Communication, ISSN: 2320-2084,

A. Face Recognition System


The proposed model for face recognition system is
as shown in Fig. 1.
The main modules used are

A. Database Used
1) NIT Database: It is the database collected
from the 40 students of NIT Warangal. NIT face
database consists of frontal facial images taken under
different days, different lighting conditions, different
moods (smiling/crying), with and without spectacles
etc.

1) Pre-processing:
This stage consists of 3 main steps:
a) Face Detection: ViolaJones face detection
algorithm is used for Face detection.
b) Resizing: Once face is detected, it is resized
to a fixed resolution of 92 *112 pixel resolution
c) Contrast
Enhancement:
For
better
recognition accuracy, we have used Min-Max Linear
Contrast Stretch Enhancement Technique,
2)

Volume- 1, Issue- 4

B. Graphical User Interface (GUI):


Figure 9,10 shows the GUI that were made using
Microsoft Visual Studio 2010, as a Windows Forms
Application using Visual C++ as a programming
language and making use of OpenCV 2.4.2 (
computer vision library written in c++) and together
with some add on visual effects in the GUI .

Feature Extraction and Recognition


As shown in the GUI as shown in Fig 9, certain
provisions are provided for selecting few options like

For feature extraction and recognition we have


used PCA/LDA Algorithm. Since we want real time
processing of data, we have chosen the simple
PCA/LDA
algorithms
where
computational
complexity is less when compared to the other face
recognition algorithms.

Database Selection: ECE / CSE Database


Algorithm Selection: PCA /LDA Algorithm
Image Type Selection: Offline/Real-Time
Camera Selection :Blackboard /Frontal Door
Specialization :ACS / VLSI/ EIE

Identified person announcement button which has the


Microsoft text to speech converter integrated in it to
announce the name of the recognized person as
Mr.xxx. is present. Update Attendance button will
automatically update the attendance along with time
and date of his arrival in an Excel sheet as shown in
Fig 3.Update html button will allow access for a
remote authenticated user to retrieve the attendance
sheet as shown in Fig 4 and Fig.5. Figure 6 shows the
Login form for Principal /Head of the Department.
Once authenticated GUI as shown in Fig.10 gets
opened, where we have certain options like searching
a particular person in a classroom with the help of his
image, an attendance sheet as shown in Fig. 7.Charts
depicting the number of hours an Employee works
and attendance percentage for the chosen month as
shown in Fig. 8 and Fig.
11.And also we have options for applying leave and
stipend calculation.

Fig. 2. Eye Blink Detector

3)
Eye Blink Detector: Spoofing attack (copy
attack) is a major threat for our proposed system. If
instead of a live person, if we keep a photograph, it
marks the attendance of the person onto the Excel
sheet, which shouldnt be the case for
Classroom/Employee
attendance
Applications.
Figure. 2 show the eye blink detector algorithm.
IV.

EXPERIMENTAL RESULTS
Fig. 3. Excel sheet with Attendance marked
Real Time Face Recognition System For Time and Attendance Applications
8

International Journal of Electrical, Electronics and Data Communication, ISSN: 2320-2084,

Volume- 1, Issue- 4

Fig. 4. Html Login Page

Fig. 7. GUI:Attendance Report

Fig. 5. Html page Attendance Report

Fig. 8. GUI:Graph plotting number of hours worked on each


day

Fig. 6. GUI Attendance Management System :Login Page

Real Time Face Recognition System For Time and Attendance Applications
9

International Journal of Electrical, Electronics and Data Communication, ISSN: 2320-2084,

Fig. 9. GUI developed : Face Recognition System

Fig. 10. GUI developed : Attendance Management System


Real Time Face Recognition System For Time and Attendance Applications
10

Volume- 1, Issue- 4

International Journal of Electrical, Electronics and Data Communication, ISSN: 2320-2084,

Volume- 1, Issue- 4

Table II
LDA Algorithm
Database Size

Offline Images

Real time Video

( Recognition
Rate)
0.62

(Recognition
Rate)
0.60

0.8

0.72

0.82

0.75

0.85

0.82

0.9

0.87

0.89

0.85

0.8

0.79

0.82

0.76

CONCLUSION AND FUTURE SCOPE


Due to the mounting demands for an Automated
Attendance System, Person Recognition, systems
have gained a lot of importance these days. Hence
depending upon the size of the Training Database,
recognition rate varies for both of these
algorithms.And also we have observed that the
recognition rate achieved for real time is mouch
lesser when compared to offline images.
Performance of the proposed system reduces by 3-5%
of recognition accuracy, when tested with or without
spectacles. Typical variations where our system is
likely to fail are intentional masking of the face
/tonsuring of the head /beard/obscuring the face with
a scarf. Other type of Biometrics like Gait can be
used for better recognition in longer distance
recognition.

Fig. 11. GUI:Graph plotting attendance percentage for each


student

C. Analysis
Table I shows the comparison of the recognition rates
between offline images and for Webcam videos using
PCA for varying Database size.
Table I
PCA Algorithm
Database Size

Offline Images

Real time Video

( Recognition
Rate)
0.6

(Recognition
Rate)
0.56

0.7

0.64

0.79

0.70

0.8

0.75

0.85

0.79

0.89

0.86

0.9

0.88

0.93

0.91

REFERENCES
[1] M. Turk, A. Pentland, Eigenfaces for Recognition, Journal of
Cognitive Neurosicence, Vol. 3, No. 1, 1991, pp. 71-86
[2] J. Lu, K.N. Plataniotis, A.N. Venetsanopoulos, Face
Recognition Using LDA-Based Algorithms , IEEE Trans. on
Neural Networks, Vol. 14, No. 1, January 2003, pp. 195-200
[3] Ramesha, K and Raja Feature extraction based face
recognition, gender and age classification ,(IJCSE)
International Journal on Computer Science and Engineering,
Vol. 02, No.01S, 2010,pp 14-23.

Similarly Table II shows the comparison using LDA


Algorithm.

Real Time Face Recognition System For Time and Attendance Applications
11

Вам также может понравиться