Poster Magang

You might also like

Download as ppt, pdf, or txt
Download as ppt, pdf, or txt
You are on page 1of 1

Fingerprint Minutiae Extraction Algorithm in

Real Time Fingerprint Recognition System


using FPGA Device
Sunny Arief SUDIRO1,2), Michel PAINDAVOINE1), Tb. Maulana KUSUMA2)
1)
LE2I-UMR CNRS 5158 University of Burgundy, Dijon, France
2)
Gunadarma University, Jakarta, Indonesia
{sunny-arief.sudiro,paindav}@u-bourgogne.fr  {sunny,mkusuma}@staff.gunadarma.ac.id

INTRODUCTION
There are many ways to identify or verify a person, manually or automatically. An increasing number of applications and
system that use verification process to access their services, they need a fast and reliable system in identifying a
person. Biometric recognition system needs a process to extract biometric feature. In fingerprint recognition system,
some research in fingerprint feature extraction process concern on pixel representation value '1' that lead to ROI
problem, and using trigonometry calculation that involving the floating point number when obtaining the minutiae
parameters, this method is not simple and difficult to implement in hardware environment. Other method is using global
feature that also involve in floating point calculation, also bigger size in template. Using artificial neural network
or fuzzy rule in detecting minutiae points, event it is possible but this method is really not simple and difficult to
implement in hardware environment.
In this research, main perspectives of developing and implementing fingerprint extraction algorithm as a part of
fingerprint recognition system is focused. First, developing a simple algorithm to extract fingerprint feature and test
this algorithm on PC. The second thing is implementing this algorithm into FPGA devices. The major research topics on
which the proposed approach is developing and modifying fingerprint extraction feature algorithm using crossing number Figure. 1. Fingerprint image
structure and their minutiae
method on pixel representation value '0', so there is no need a process concerning ROI segmentation. And specially in points detected.
obtaining their parameters using Angle Calculation Block avoiding floating points calculation. As this method is local
feature that usually involve with 60-100 minutiae points, makes the template is small in size.

SOFTWARE BASED APROACH & RESULT


Minutiae extraction is based on Crossing Number Algorithm, this algorithm is working
on pixel representation ‘0’ as a concern, but the decision of minutiae point can be
selected for each pixel value. Ideally, a good skeleton is strictly one pixel to
obtain good result of minutiae points detection. Minutiae points extraction
algorithm can be applied with crossing number (CN) at a point P and Pi is the binary
pixel value in the neighborhood of P, expressed as:

8
CN  0,5 | Pi  Pi 1 | with P9 = P1 (1)
i 1

Usually fingerprint sensor provides the grayscale image of fingerprint. Extraction


process needs binary image, so binarization process must be applied. Then thinning
process is performed to obtain the representation of one pixel width of fingerprint
structure, especially for representation of pixel value ’0’ for the skeleton image.
Figure 2 present this algorithm and angle calculation block method for obtaining
minutiae point parameter.

Figure. 2. Flowchart of proposed algorithm and Angle Calculation method


for obtaining minutiae point parameters, (a) Bifurcation point and (b)
End Point.

The first result of fingerprint extraction feature algorithm evaluation using


human visual analysis has 12,37 % minutiae points not detected. Using bozorht3
algorithm as matching algorithm for 80 fingerprint images we get EER 5.89 % at
threshold value 180.

The performance evaluation of proposed algorithm by providing FAR, FRR and EER
also Receiver Operation Characteristic will be based on mindtct extraction and
bozorth3 matching algorithm from NFIS2 provided by NIST (see fig. 3), this
Figure. 3. ROC curve for minutiae points extraction algorithm for all performance evaluation is performed using three dataset consisting around 1880
three dataset.
of fingerprint images.

HARDWARE BASED APROACH & RESULT


Finally in relation to time computation constraint, we intend to implement full algorithm using
hardware system as FPGA device based on hardware schematic we provided. This hardware-based
approach consists of several component/module, such as: thinning component, sensor module,
absolute value module, crossing number module and others (see figure 4 and 5). For sensor module
will generate signal controller needed by fingerprint sensor to capture fingerprint image. For
sensor module a MBF 200 from Fujitsu is used and constructing this device on prototype board to
be connected to FPGA board Zest SC1.
Fingerprint sensor read module has
been developed and simulated. The
result, initialization process takes
33.154 uS. And for each pixel of
Figure. 5. Hardware schematic for (a) sensor module,
fingerprint image, the processing (b) absolute value module and (c) crossing number module
time is 0.64 uS and for all pixels,
it requires 50.99 mS with 12 MHz
clock signal. As we can see in
figure 6, this module has provided a
good signal in sequence to drive
fingerprint sensor.

Figure. 6. Sensor signal controller read by Logic


Figure. 4. Accessing 8 bit neighbor in serial data and thinning component/module in FPGA device.
Analyzer.

CONCLUSSION
We propose an adaptable fingerprint minutiae extraction algorithm into hardware implementation with the complexity of O(n) and 14.05 % of EER,
better than reference algorithm, which is 20.39 % and 18 seconds computational time better than similar method, which is 60-90 seconds just for
preprocessing step. The first step of algorithm implementation in hardware environment (embedded) using FPGA Device by developing IP Core without
using any soft processor is presented.

FJC ‘2009 UB Dijon

You might also like