An Enhanced Authentication System Using Face and Fingerprint Technologies

IOSR Journal of Computer Engineering (IOSR-JCE)
e-ISSN: 2278-0661,p-ISSN: 2278-8727, Volume 17, Issue 6, Ver. III (Nov – Dec. 2015), PP 74-84
www.iosrjournals.org
DOI: 10.9790/0661-17637484 www.iosrjournals.org 74 | Page
An Enhanced Authentication System Using Face and Fingerprint
Technologies
Ogbuokiri, B. O. *1
and Agu, M.N*2
*1
*2
.Department of Computer Science, University of Nigeria, Nsukka, Nigeria.
*1
blessing.ogbuokiri@unn.edu.ng*2
monica.agu@unn.edu.ng
Abstract: The primary aim of this paper is to develop an enhanced authentication system using a Cascaded-
Link Feed-Forward Neural Networks. In the end, the system overcomes some limitations of face recognition and
fingerprint verification systems by combining both. Experimental results demonstrate that the system performs
well. It meets the response time as well as the accuracy required.
Keywords: Multi-biometric, Face recognition, Fingerprint verification, Minutiae, pattern matching.
I. Introduction
Biometric authentication has been receiving extensive attention over the past decade with increasing
demands. Biometric is to identify individuals using physiological or behavioral characteristics, such as
fingerprint, face, iris, retina, palm-print, etc. Among all the biometric techniques, fingerprint recognition [1] is
the most popular method and is successfully used in many applications. Although fingerprint recognition
methods have been attributed with many flaws such as users fingerprint being dirty, wet, scratches and abrasion
which could lead to medical issues. Fingerprint technology employ feature-based image matching, where
minutiae (i.e., ridge ending and ridge bifurcation) are extracted from the registered fingerprint image and the
input fingerprint image, and the number of corresponding minutiae pairings between the two images is used to
recognize a valid fingerprint image [1].The feature-based matching provides an effective way of identification
for majority of people.
On the other hand, face recognition technologies is gradually taking the lead in the access control
system. Most public and private places prefer face for access control and security. This has attracted the
attention of vision researchers for several years. Face recognition may seem an easy task for humans, and yet
computerized face recognition system still cannot achieve a completely reliable performance [2]. The
difficulties arise due to large variation in facial appearance, head size, orientation and change in environment
conditions. Such difficulties make face recognition one of the fundamental problems in pattern analysis. In
recent years there has been a growing interest in machine recognition of faces due to potential commercial
applications such as film processing, law enforcement, person identification, access control systems, etc [2,3].
Owing to the foregoing, there is therefore increasing need for an enhanced authentication system that
can combine fingerprint and face. In this paper, attention is more on multi biometric to identify persons using
fingerprint and face. This paper is organized as follows: General architecture of the proposed system, an
Overview of fingerprint enrollment/detection. Face enrollment, detection and recognition, implementation of the
proposed architecture, testing and results. Finally, some concluding remarks.

An Enhanced Authentication System Using Face and Fingerprint Technologies
II. General architecture of the proposed system
Figure 1: Architecture of enhanced Authenticatin system
The enhanced authentication system is divided into two major parts namely; the fingerprint recognition
part and the face recognition part. When the application is run the user is requested to enroll his/her fingerprint
for a maximum of three times through a finger print scanner in other to extract a best feature. When a best
feature is extracted, it is then stored in the database for verification. In Addition, after a successful enrollment of
fingerprint, the user is then prompted to enroll his/her face using either a webcam or attached camera as many as
possible to capture a best quality needed. This is further stored in the database after extracting the best facial
feature. This also can be done vice versa. During verification, the fingerprint scanner, webcam or attached
camera is used to capture a fingerprint and a human face respectively. The captured image is extracted and
further compared with the template in the database for verification. We shall now go further to discuss in details
what happens in fingerprint enrollment / verification process and Face enrollment / recognition processes
respectively.
III. Fingerprint enrollment/Verification
A fingerprint is presented as a series of lines which is the high line for the friction skin. We call it the
ridge; as shown in Figure 2. The ridge of a fingerprint usually becomes the dark lines of a fingerprint image
which is acquired by the use of the sensor device called fingerprint scanner. Between the ridges are the low lines
which are the valleys. The valleys are showed as white lines in a fingerprint image. In Figure 3, some other
characteristics relative to fingerprint image processing is shown. When talking about fingerprints, we often see
some terms: local, global and minutiae. The local feature is a major representation of the ridges (the ridge
endings and the ridge bifurcations, as Figure 4a and 4b), the valleys and the pores on ridges. The global features
are a major representation relatives to a classification and information about locations of critical points (e.g.,
core and delta features, sometimes they are referred to as singularities points) on a fingerprint, see the
descriptions as Table 2. The minutiae features are the ridge endings and the ridge bifurcation, see as Figure 5 [4]
in [1].

Figure 2: The ridge and valley of a finger print
Figure 3: other characteristics of fingerprint image
Figure 4a: The position and orientation of the ridge ending and ridge bifurcation [1]

Figure 4b: Minutiae on a fingerprint [1]
1.2 Fingerprint enrollment Process
The design of the processing of the enrollment module according to the steps in Figure 5.
Figure 5: Fingerprint enrollment process
Step 1: Fingerprint image: Getting fingerprint image from the sensor device such finger print scanner.
Step2: Enhancement of fingerprint image: Fingerprint images that are acquired by ink or a scan sensor can
include noise signals. For accurate recognition, improvement of the clarity and continuity of ridge structures is
made. This is performed into 5 stages using wavelet transform; an algorithm for processing images, the stages
involved are normalization, wavelet decomposition, global texture filtering, local directional compensation and
wavelet reconstruction [6].
Step 3: Get needed features: There are some important features which are gotten from each fingerprint image.
These are the center point, the class type and the features of minutiae (the location, the orientation, etc.).

Knowing the class that each fingerprint belongs on the orientation under the center point.Extraction is now made
on the direction features of the lower portion of the octagon, according to Figure 6.
Figure 6: The portion of the octagon extracted for the direction features
Getting minutiae features is based on the thinned fingerprint image. Crossing number was used to
determine a minutiae point [5]. Crossing numberis defined as half of the sum of two adjacent pixels. According
to the result of crossing number, termination and bifurcation is recognized, as shown in Table 1.
Table 1: Cross number value to type of the minutiae
Step5: Data structure: After the needed features are gotten, the data structure for a fingerprint image is now
created as shown on Figure 7.
Figure 7: Data structure of fingerprint image
Step 6: Database: Saving the data structure of a fingerprint image into the database
3.2 Fingerprint verification process
The Fingerprint verification module uses the same representation which was used in enrollment
module. The difference with the enrollment module is the matching fingerprint step, see Figure 8

Figure 8: fingerprint verification module process
Due to the difficulty in matching a finger print, the input and template fingerprint features transform
into a common frame. The transformation of the input and the template fingerprint are rotation, translation and
scale. Two minutiae points are matched when they are within in the tolerance size [7, 9, 10].
For matching fingerprint, information in the fingerprint data structure is collected, involving two main
steps: The first step is to limit the number of templates fingerprint by the classification value of a fingerprint
image; that means a comparism of the input fingerprint with the sub template which has the same classification
value with the template.The second is the comparismofeach minutiae of the input fingerprint image with each
minutiae of the template fingerprint, see as Figure 9.
Figure 9: The processing of matching minutiae; (a)The minutiae set of the input fingerprint; (b)The
minutiae set of the template fingerprint; (c)The alignment based on the minutiae of the input fingerprint and

template fingerprint; (d)the matching result where the template minutiae and their correspondences are
connected [7, 1].
IV. Face detection and recognition system
The Face recognition system is made up of the following main steps: Data Preprocessing, Feature
Extraction and Classification. Data Preprocessing involves steps like face detection, noise reduction, image
resizing, scaling and so on. Feature Extraction involves extraction of data from the images that are relevant to
face recognition and removal of any redundant data leading to reduction in size of the data set [8]. Classification
will involve determining the class of the input test data based on the features extracted from the training data set.
See figure 10.
Figure 10: Architecture of the face detection and recognition system. [2]
4.1 Face detection
Face detection can be viewed as two-class recognition problem in which an image region is classified
as being a “Face” or “nonFace”. Consequently, face detection is one of the few attempts to recognize from
images a class of objects for which there is a great deal of within-class variability. Face detection also provide
interesting challenges to the underlying pattern classification and learning techniques. The class of face and no
face image are decidedly characterized by multimodal distribution function and effective decision boundaries
are likely to be nonlinear in the image space [11, 12]. Artificial Neural Network (ANN) Multi-Layer Perception
(MLP) was employed for the face detection see figure 11.
Figure 11: Face detection using neural network [2].
The MLP neural network [14] has feed-forward architecture within input layer, a hidden layer, and an
output layer. The input layer of this network has N units for an N dimensional input vector. The input units are
fully connected to the I hidden layer units, which are in turn, connected to the J output layers units, where J is
the number of output classes. A Multi-Layers Perceptron (MLP) is a particular kind of artificial neural network
[13]. Assuming that there is access to a training dataset of l pairs (xi, yi) where xi is a vector containing the
pattern, while yiis the class of the corresponding pattern. In this case a 2-class task, yicanbe coded 1 and -1. In
the proposed system, the dimension of the retina is 15x15 pixels representing human faces and non-face, the
input vector is constituted by 225 neurons, the hidden layer has 15 neurons [12, 13].

1.3 Pre processing
The appearance of the face varies due to relative camera-face pose, between full frontal images and
side-profile images; in-situ occlusions such as facial hair (e.g.beard, moustache), eye-glasses and make-up;
facial expressions can significantly influence the appearance of a face image; overlapping occlusions where
faces are partially occluded by other faces present in the picture or by objects such as hats, or fans; conditions of
image acquisition where the quality of the picture, camera characteristics and in particular the illumination
conditions can strongly influence the appearance of a face. For better system performance, in the recognition
stage, a set of pre-processing techniques was applied: the first step in pre-processing is to bring all images into
the same color space and to normalize the size of face regions. This normalization process is critical to
improving the final face recognition rate and some experimental results will be presented in the later [15,16,17].
Figure 12: Face detection showing preprocessing stage using neural network [12].
4.3 Feature extraction
The feature extraction technique used is implemented by scanning the image with a fixed-size window
from left-to-right and top-to-bottom. A window of dimensions h × w pixels begins scanning each extracted face
region from the left top corner sub-dividing the image into a set number of h × w sized blocks. On each of these
blocks a transformation is applied to extract the characterizing features which represent the observation vector
for that particular region. Then the scanning window moves towards right with a step-size of n pixels allowing
an overlap of o pixels, where o = w − n. Again features are extracted from the new block. The process continues
until the scanning window reaches the right margin of the image. When the scanning window reaches the right
margin for the first row of scanned blocks, it moves back to the left margin and down with m pixels allowing an
overlap of v pixels vertically. The horizontal scanning process is resumed and a second row of blocks results,
and from each of these blocks an observation vector is extracted [18,19,20,21,22]. The scanning process and
extraction of blocks is shown on Figure 12.
Figure 13: Block extraction from a face image.

Figure 14: Feature extraction from a face image using neural network [12].
4.4.1 Face Recognition
4.4.2 Training
Multi-Layer Perceptron (MLP) with a back propagation learning algorithms was chosen for the
proposed system because of its simplicity and its capability in supervised pattern matching. It has been
successfully applied to many pattern classification problems [11]. The problem has been considered to be
suitable with the supervised rule since the pairs of input-output are available. For training the network, the
classical back propagation algorithm was used. An example is picked from the training set, the output is
computed. The error is calculated as the difference between the actual and the desired output. It is minimized by
back-propagating it and by adjusting the weights. Although back-propagation can be applied to network with
any number of layers, it has been shown that one layer of hidden units suffices to approximate any function
[11,13,14]. Therefore, in most application, a MLP Neural Networks (NN) with a single layer of hidden units is
used with a sigmoid activation function for unit f (a)=
1
1+ 𝑒−𝑎 . This function has the interesting property of
having an easy to compute derivative f’(a) = f(a)[1 – f(a)]. The MLP training is amount to: repeatedly presented
with sample inputs and desired targets. Then the output and targets are compared and the error measured. At
last, adjusts weights until correct output for every input.
4.4.3 Evaluation
In the evaluation process, a model of a newly acquired test subject iscompared against all existing
models in the database and the most closely correspondingmodel is determined. If these are sufficiently close, a
recognition event is triggered [2].
V. Implementation
The system was implemented using vb.net on a visual studio 2010 and MySQL as relational database.
At first the neural network architecture was created and then trained with training set of faces and non-faces. In
neural network architecture some functions are used for training purpose (training function trainscg ()),
initialization purpose (Layer Initialization Functions initnw()), and for performance purpose (Performance
Functions msereg ()) etc. In the face detection algorithms Fast Fourier transform are used with training function
trainscg() [13,23,]

VI. Testing and results
6.1 Testing
Different faces and fingerprints were used to assess the performance of the system. For each person,
one fingerprint with acceptable quality was selected as the template. All but one face images were used to train
the face recognition subsystem. The remaining face image was paired with the fingerprint to form a test sample.
This process was repeated several times. An example of the identification is shown on figure 13a and 13b
Figure 15a Face Image acquisition and identification
Figure 15b fingerprint Image acquisition
1.4 Results
The identification accuracy of the integrated system as well as identification accuracies of face
recognition and fingerprint identification is compiled in table 1. The response time of the integrated system for a
typical identification is 3 seconds see table 2. In comparison, identification using only fingerprints on a database
of 5 persons takes 9 seconds. From this, the system can achieve desirable identification accuracy with an
acceptable response time. This was developed and run on a machine with following configuration; Intel core i5
2.5 GHZ processor with window 7 32bit operating system, 2 GB RAM.
Table 1: Identification accuracy of face recognition and fingerprint identification
No of Faces No of fingerprints No of faces detected
at each compare
No of fingerprints
detected at each
compare
Accuracy (%)
5 5 1 1 100

Table 2: Response time of the system
Face location (Seconds) Face Retrieval (Seconds) Fingerprint Verification
(Seconds)
Total (Seconds)
0.5 0.5 2.0 3.0
VII. Conclusion
This system has successfully integrated and implemented face and fingerprint technologies for personal
identification. The system overcomes some limitations of face recognition and fingerprint verification systems.
Experimental results demonstrate that the system performs well. It meets the response time as well as the
accuracy required.
References
[1] Tran, T.N. and Nguyen, T.H. Building a fingerprint recognition module and making it interact with the enrollment fingerprint
application.MSc Thesis, HO CHI MINH CITY, 2012.
[2] Peter, M. C. and Claudia, I. Automatic Face Recognition Systemfor Hidden Markov Model Techniques, College of Engineering
&Informatics,National University of Ireland Galway, Ireland, 2010.
[3] Adini, Y., Moses, Y. and Ullman, S. Face recognition: the problem of compensating for changes in illumination direction. IEEE
Trans. on Pattern Analysis and Machine Intelligence, 19(7):721–732, 1997.
[4] Lee, H.C. and Giessen, R.E, Adcances in fingerprint technology, Elsevier, New York, 1991.
[5] Henry, E. R., Classification and uses of fingerprints, London: Routledge ,2000.
[6] Ching-Tang, H., Eugene, Li., and You-Chuang, W., An effective algorithm for fingerprint image enhancement base on wavelet
transform, Pattern Recognition vol. 36 pp. 303-312, 2003.
[7] Zho, R. W., Quek, C. and NG, G. S., Novel single pass thinning algorithm, Pattern Recognition Letter, Vol. 16, pp. 1267-1275.
2005.
[8] Vijaya, K. N., Gnaneswara, R, N. and A.L.N.R, RTL: Reduced texture spectrum with lag value based image retrieval for medical
images, International Journal of Engineering Science and Technology Vol. 2, No. 4, December, 2009
[9] Ravi, J. Enhanced fingerprint verification system, International Journal of Engineering Science and Technology Vol.1(2),2009, 35-
42
[10] Anil, J., Lin, H., Sharath, P., and Ruud, B., An identity authentication system using fingerprints, Elsevier, New York, 2010.
[11] Paul, V. and Michael, J. Rapid Object Detection using a Boosted Cascade of Simple Features, Computer Vision and Pattern
Recognition, USA. 2001
[12] Henry, A. R., Shumeet, B., and Takeo, K., Neural Network-Based Face Detection. IEEE Trans. On Pattern Analysis and Machine
Intelligence, vol.20, No. 1, Page(s). 39-51, 2008.
[13] Smach, F., Atri, M., Mitéran, J. and Abid, M. Design of a Neural Networks Classifierfor Face Detection.World academy of
Science, engineering and technology, 2005.
[14] Tarun, K., Kushal, V.S. and Shekhar, M. Artificial Neural Network in Face Detection.International Journal of Computer
Applications (0975 – 8887) Volume 14– No.3, January 2011
[15] Anil, J., Lin, H., and Yatin, K. F2ID: A personal identification system using faces and fingerprints, Department of Computer
science, Michigan state University. 2001
[16] Ming-Husan, Y., David, J., and Narendra, A. DetectingFaces in Images: A Survey. IEEE transaction on pattern analysis and
machine intelligence, vol.24 no.1, January 2002.
[17] Zhang, Z., Zhu, L., Li S.Z., Zhang H. J., Real-time multi-view face detection.Proceeding of the Fifth IEEE International Conference
on automatic Face and Gesture Recognition, Page(s): 142-147, 20-21 May 2002.
[18] Feraund, R., Bernier, O.J., Viallet, J., Collobert, M . A fast and accurateface detector based on neural network. IEEE Transactions
on Pattern Analysis and Machine Intelligence, Volume: 23 Issue: 1,Pages(s):42-53, Jan.2001.
[19] Smith, W.A.P. and Hancock, E.R. Single image estimation of facial albedo maps. Lecture Notes in Computer Science, 3704:517–
526, 2005.
[20] Steinberg, E., Corcoran, P., Prilutsky, Y., Bigioi, P., Ciuc, M., Ciurel, S., and Vertran, C.,Classification system for digital images
using workflow, face detection, normalization, and face recognition, US Patent 7,555,148, June 2009.
[21] Viola, P. and Jones, M. , Robust real-time object detection, presented at the 2nd International workshop on Statistical and
Computational Theories of Vision, Vancouver, Canada, July 13th, 2001.
[22] Zhao, W., Chellappa, R., Rosenfeld, A., and Phillip, P.J. Face recognition: A literature survey. ACM Computing Surveys, pages
399–458, 2003

An Enhanced Authentication System Using Face and Fingerprint Technologies

More Related Content

What's hot (19)

Similar to An Enhanced Authentication System Using Face and Fingerprint Technologies (20)

More from iosrjce (20)

Recently uploaded (20)

An Enhanced Authentication System Using Face and Fingerprint Technologies