IJCSI International Journal of Computer Science Issues, Vol.

7, Issue 3, No 9, May 2010 45

ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Image Compression Algorithms for Fingerprint System

Preeti Pathak
CSE Department, Faculty of Engineering, JBKP, Faridabad, Haryana,121001, India

Abstract
minutiae based automatic identification technique first
Fingerprint-which have been used for about 100 years are the locates the minutiae point and matches their relative
oldest biometric signs of identity. Humans have used fingerprints placement in a given finger and the stored template,
for personal identification for centuries and the validity of shown in figure 1.
fingerprint identification has been well established. In fact,
fingerprint technology is so common in Human Identification that
it has almost become the synonym of biometrics. Fingerprints are
believed to be unique across individuals and across fingers of
same individual. Even identical twins having similar DNA, are
believed to have different fingerprints. The analysis of
fingerprints for matching purposes generally requires the
comparison of several features of the print pattern. These include
patterns, which are aggregate characteristics of ridges, and minutia
points, which are unique features found within the patterns. is also
necessary to know the structure and properties of human skin in
order to successfully employ some of the imaging technologies. A
major approach for fingerprint recognition today is to extract
Fig. 1: Fingerprint Recognition System
minutiae from fingerprint images and to perform fingerprint
matching based on the number of corresponding minutiae pairings.
One of the most difficult problems in fingerprint recognition has The fingerprint image may be obtained from a thumb pad
been that the recognition performance is significantly influenced fingerprint scanner device scanning at 500 dpi [2] . A
by fingertip surface condition, which may vary depending on good quality fingerprint contains between 60 and 80
environmental or personal causes. Addressing this problem this minutiae, but different fingerprint have different number
paper propose some extra features that can be used to strengthen of minutiae. A fingerprint image essentially consists of a
the present approaches followed in developing Fingerprint
set of minutiae on the plane. Minutiae are the
recognition system. To increase security and accuracy we can use
Infrared technique and technique to assign a score value to each of terminations and bifurcations of ridge lines in a
extracted minutiae. fingerprint image. In order to extract these minutiae, we
have to undergo the operation linearization followed by
the process of thining. Thus, the set minutiae those are
Key Terms Biometric, Minutiae, Binarization, Thinning,
Median Filter.
well defined and more prominent then the rest are given
higher relevance and importance in the process of
minutiae matching.
1. Introduction

Biometric authentication has been receiving extensive

attention over the past decade with increasing demands
in automated personal identification . 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 Fig.2 : Terminations and bifurcations of ridge lines in a fingerprint
and is successfully used in many applications. Typical image
fingerprint recognition methods 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 2. System Architecture
the number of corresponding minutiae pairings between
the two images is used to recognize a valid fingerprint
The fingerprint recognition system is shown in figure3. It
image [1].The feature-based matching provides an
consists of four components
effective way of identification for majority of people.The
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 3, No 9, May 2010 46
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

1. User Interface (with Infrared Sensor) the regions that can not be reliable recovered. The
enhanced fingerprint image is fed to the minutiae
2. System Database extractor again. The task of authentication module is
to authenticate the identity of the person who intends
3. Enrollment Module to access the system. The person to be authenticated
indicates his / her identity and places his / her finger
on the fingerprint scanner ; a digital image of his her
4. Authentication Module fingerprint is captured ; minutiae pattern is extracted
from the captured fingerprint image and fed to a
Infrared matching algorithms which matches it against the
Sensor persons minutiae templates stored in the system
database to establish the identity To increase
security and accuracy we can use Infrared technique
and technique to assign a score value to each of
extracted minutiae.

Fig 3 : System Architecture

The user interface provides mechanisms for a user to

indicate his/her identity and input his/her
fingerprints into the system. Here along with finger
print scanner the Infrared Sensor will be used. The
system database consists of a collection of records,
each of which corresponds to an authorized person
that has access to the system. Each record contains
the following fields which are used for
authentication purpose:

User name of the person , Fig. 4: Block diagrams of Enrollment ,Verification, Identification Task
Minutiae templates of the persons fingerprint,
and other information (e.g., specific user 3. Development of proposed system
privileges). Alongwith thermal parameter.
Proposed system consist of following sub modules -
The task of enrollment module is to enroll persons
and their fingerprints into the system database.
3.1 Normalization
When the fingerprint images and user name of a
person to be enrolled are fed to the enrollment
module, a minutiae extraction algorithm is first By normalizing an image, the colors of the image are
applied to the fingerprint images and the minutiae spread evenly throughout the gray scale. A normalized
patterns are extracted. A quality checking algorithm image is much easier to compare with other images, and
is used to ensure that the records in the system the quality of the image is easier determined.
database only consist of fingerprints of good quality, Normalization is a pixel-wise operation that does not
in which a significant number (default value is 25) change the clarity of the ridge and valley structures .
of genuine minutiae may be detected. If a fingerprint Normalization reduces the variations in gray level values
image is of poor quality, it is enhanced to improve along ridges and valleys, which facilitates the subsequent
the clarity of ridge/vally structures and mask out all processing steps.
Let I(x; y) denote the grayscale value at pixel (x; y), M and
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 3, No 9, May 2010 47
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

V, the estimated mean and variance of grayscale values in

this 64 X 64 window, respectively, and N(x; y), the Median filtered
normalized grayscale value at pixel (x; y). For all the
Pixels in the window, the normalized image is defined as: * * *
* 4 *
V 0 (T ( x, y) M 2 * * *
M0 , ifI ( x, y) M
N(x, y) = V
(1) Center value (previously 97) is replaced by the median of
all nine values. Note that for the first (top) example, the
V 0 (T ( x, y) M2
M0 , ifI ( x, y) M median filter would also return a value of 5, since the
V ordered values are 1, 2, 3, 4, 5, 6, 7, 8, 9. For the second
(bottom) example, though, the mean filter returns the
value 16 since the sum of the nine values in the window
is 144 and 144 / 9 = 16. This illustrates one of the
In Eq. (1) , M0 and V0 are the desired mean and variance celebrated features of the median filter: its ability to
Values , respectively. Normalization is a pixel-wise remove 'impulse' noise (outlying values, either high or
Operation and does not change the clarity of the ridge and low). The median filter is also widely claimed to be
Valley structures . For our experiments, we set the values of 'edge-preserving' since it theoretically preserves step
both M0 and V0 to 100. The values of M0 and V0 should edges without blurring. However, in the presence of
be the same across all the training and test sets [7].
noise it does blur edges in images slightly.

3.2 Median Filter Median filtering is a simple and very effective noise
removal filtering process. Its performance is particularly
good for removing shot noise. Shot noise consists of
In image processing it is usually necessary to perform
strong spike like isolated values.
high degree of noise reduction in an image before
performing higher-level processing steps, such as edge
detection. The median filter is a non-linear digital Shown below are the original image and the same image
filtering technique, often used to remove noise from after it has been corrupted by shot noise at 10%. This
images or other signals. The idea is to examine a sample means that 10% of its pixels were replaced by full white
of the input and decide if it is representative of the signal. pixels. Also shown are the median filtering results using
This is performed using a window consisting of an odd 3x3 and 5x5 windows; three (3) iterations of 3x3 median
number of samples. The values in the window are sorted filter applied to the noisy image; and finally for
into numerical order; the median value, the sample in the comparison, the result when applying a 5x5 mean filter
center of the window, is selected as the output. The to the noisy image.
oldest sample is discarded, a new sample acquired, and
the calculation repeats.

Median filtering is a common step in image processing.

It is particularly useful to reduce speckle noise and salt
and pepper noise. Its edge-preserving nature makes it
useful in cases where edge blurring is undesirable.

The median filter is also a spatial filter, but it replaces

the center value in the window with the median of all the
pixel values in the window. The kernel is usually square
but can be any shape. An example of median filtering of
a single 3x3 window of values is shown below. Fig 5: a) Original image; b) Added Shot Noisy at 10\%

Unfiltered values 3.3 Binarization (Regional Average Threshold)

6 2 0
3 97 4 Making an image binary, transforms the gray scale
19 3 10 image into
a binary image (black and white). Either a global or
In order: 0, 2, 3, 3, 4, 6, 10, 15, 97 localized threshold value
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 3, No 9, May 2010 48
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

is used. Gray level images cannot be operated on for the 2. The algorithm finds out (x,y) location of the
determination of key features as the range of values of first block pixel which is not processed yet in
pixel intensities very widely. So it is very difficult to the original binary image.
discriminate conspicuously the valleys from the ridges, 3. A black pixel is inserted into thinned image at
as well as ridges bifurcations from ridge endings. Hence (x,y) location (gray pixels) and the black pixel
some half toning technique needs to be applied on the is removed from the original binary image at the
finger print image. A popular technique is that of location.
Threshold of Binarization [3] [4] [5]. Then algorithm looks at the ridge continuity , if
In many image-processing applications it is desired to there is ridge continuity , it follows the ridge.
convert the the gray images into white and black (bi-
level) image. Thesholding involves looking at each pixel 3.5 Noise Removal
and deciding whether it can be converted into white (0)
or black (255), the decision is made by comparing the
numerical pixel value against a fixed number called a After applying thining algorithm we are lift with an
threshold level (T). If any pixel of image(x ,y) is less image that still has got some noise , i.e. some ridges
than the threshold level (T), the pixel is set to zero are not of one pixel width or some other noise , this
(background point), otherwise it is set to 255 (object module is used to remove this noise
point). In this we check if in the neighborhood of a pixel
there are more than five pixels. Less than five pixels
Algorithm: imply that this pixel is a ridge end or bifurcation or
some part of a ridge. More than five pixels imply
1. Divide the image into 4*4 regions. that this pixel is associated with some noise.
2. Calculate the average of gray level in the first
4*4 region. 3.6 Depuration
3. Threshold the leftmost region of 4*2 by using
average gray level calculated in stage2. The image obtained from previous module still not
4. Move the 4*4 operation window by 2 pixels to suitable for minutie detection . This module removes
the right . If right edge of the image is reached , some more defects .Depuration of the ridge map
then move the window 4 pixels up and return to involves removal of the spurious elements,
the left edge . identified as undesirable spikes, and to join the
5. Repeat stage 2 to stage 4 until the entireimage is broken lines using a smoothening procedure . This
processed by RAT(Regional average depuration process is carried out by simple rules
thresholding). like .

To remove small isolated lines.

3.4 Thinning To merge all the lines who have end points with
similar direction and the distance between them
Hear thinning algorithm is based on ridge following.
is small.
Such an algorithm has a better probability of enhancing
the required properties of image . Ridges exist in all
Algorithm :
finger prints and they form the features either by ending
(ridge endings) or by forking bifurcations . In our work
1. First identify the pixel which is a ridge end.
we are using black pixels as the ridges in fingerprints.
2. Then find out another black pixel in the 7*7
The current work uses thining algorithm which deals
matrix neighborhood of this pixel.
with just black pixels, which are between two white
3. If this pixel exists then we find out the slope of
pixels [2] [3] [5].
two respective ridges at their respective ridge
ends in 7*7 matrixes.
Algorithm:
4. If the slopes obtained are comparable then join
ridge ends by using DDA algorithm.
1. The image is read from bottom left to the right
side line by line and the algorithm always tries
to find day any of black pixels in the original
3.7 Minutiae Extraction
image . Because it is obvious that any of black
pixels may be constituent of ridge.
Minutiae extraction was carried out using the crossing
number approach. Crossing number of pixel p is
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 3, No 9, May 2010 49
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

defined as half the sum of the differences between pairs The number of minutiae in a given area is also limited
of adjacent pixels defining the 8-neighborhood of p. therefore the minutiae density must also be kept in check.
Mathematically in eq .(2). In order to filter out these false minutiae a 3 level-
filtering process is applied:
cn( p ) 1
2 | val ( p
i 1..8
i mod 8 ) val ( pi 1) | (2)
Level 1: Removes the false ridge endings created as a
result of the application of minutiae extraction algorithm
Where p0 to p7 are the pixels belonging to an ordered at the ends of the thinned image.
sequence of pixels defining the 8-neighborhood of p and
val (p) is the pixel value. Level 2: Removes the first five types of minutiae
mentioned above using the rule based morphological
minutiae filtering approach given by .

Level 3: This stage limits the maximum number of

minutiae present in the thinned image to a pre-specified
threshold.

A minutiae m is described by the triplet m={x, y, },

where x, y indicate the minutiae location coordinates and
denotes the minutiae orientation, which is the
Fig.6: cn (p)=2,cn (p)=3 and cn (p)=1 representing a orientation evaluated for the minutiae location from the
nonminutiae region, a bifurcation and a ridge ending orientation image obtained during the enhancement
process. The minutiae type is not being used during the
Crossing numbers 1 and 3 correspond to ridge endings matching process since minutiae type can be inverted
and ridge bifurcations respectively. An intermediate due to enhancement and binarization steps [6].
ridge point has a crossing number of 2. The minutiae
obtained from this algorithm must be filtered to preserve
only the true minutiae. The different types of false
minutiae introduced during minutiae extraction include
spike, bridge, hole, break, Spur, Ladder, and
Misclassified Border areas. (See figure 7)

Fig. 8: Filtered and Unfiltered Minutiae Sets

In this process we use text files as databases. We

consider three text files for storing the bifurcation
information , the ridge information , and the thermal
parameters respectively. Every time a new fingerprint
image is operated upon , a set of two text files are
generated along with another file for keeping information
about the person whose fingerprint is under checking. In
Fig. 7: Types of false minutiae case of verification , the text files generated are matched
AB with the existing ones and the numbers of matches are
CD displayed.
EF
A. Spike, B. Bridge, C. Hole, D.Break, E. Spur F. Ladder
3.8 Minutiae Matching
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 3, No 9, May 2010 50
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

Let T and I be the representation of the template and more reliable . This system needs some extra cost
input fingerprint, respectively. Let the minutiae sets of and more processing time, but one can not
the two fingerprints be given by: compromise with the security.

T {m1, m 2,.........., mn}

I {m'1, m'2,.........., m' n} References
(3)
mi {xi, yi,i}, i 1.....m [1] D.Maltoni, D.Maio, A.K.Jain, and Prabhakar, Hand book

mj {x' j , y ' j , ' j}, j 1.....n

of fingerprint Recognition . Springer,2003

[2] V.Espinosa-Duro., Minutiae detection algorithm for

fingerprint recognition, IEEE Aerospace Electron . Syst. Mag.
A minutia mj in I and a minutia mi in T are 17(3),7-10,2002
considered to be matched if their spatial and [3] D. Manolescu, Feature extaction a pattern for
orientation differences are within specified information retrieval, In Proceedings of the 5th Pattern
Languages of Programming, Monticello, Illinois, USA, August
thresholds ro and o. Minutia matching was 1998.
carried out by using the approach given in . In
[4] A.K.Jain, Fundamentals of Digital Image Processing,
this approach the minutiae sets are first PHI, ISBN-13:978-013336150
registered using a derivative of the Hough
[5] Dario Maio, Davide Maltoni, Direct Graay-Scale Minutiae
transform followed by fingerprint matching Detection In Fingerprints, IEEE Transactions on Pattern
using spatial and orientation-based distance Analysis and Machine Intelligence v.ol. 9 n.o.l, pp., 27-40,
January 1997
computation. The matching algorithm returns a
percentage match score, which is then used to [6] F.A.Afsar, M.Arif and M.Hussain, Fingerprint
Identification and Verification System using Minutiae Matching,
take the match-no match decision based on the National conference on Emerging Technologies 2004
security criterion[6].
[7] Bhupesh Gour, T. K. Bandopadhyaya, Sudhir Sharma,
Fingerprint Feature Extraction Using Midpoint ridge Contour
method and Neural Network, IJCSNS International Journal
4. Result and discussion of Computer Science and Network Security, VOL.8
No.7, July 2008
To study the performance of the proposed work, we
consider the database containing 350 images. There
are 10 different impressions per finger. Here we
About the Author
assume that images should be of good quality with
at least 500 dpi and noise level, if any, of the image The Author, Preeti Pathak, working as Lecturer in Department,
Faculty of Engineering, JBKP, Faridabad, India. She has received
should be removed by the preprocessing. The her M.Tech (Comp. Sc. & Tech) from Maharishi Dayanand
performance of the system has shown the efficiency University, Rohtak along with two other Master degrees, M.Sc. in
about 97% in 5.8 seconds .It has been seen that Physics from Agra University, Agra and M.C.A. (Master in
when the core point is correctly located, the Computer Applications) from IGNOU, New Delhi, INDIA. She
translation invariant property of features is satisfied has also served as Lecturer in other Eng. Colleges like BSA and
and the rotation handled in the matching stage is Arravali Eng. College Faridabad. She has been involved in many
very fast. As result of which the matching process Training, Seminars and conferences. She has also presented her
becomes very fast. Technical Papers in Two International Conferences and two
National Conferences. She has also completed her Training on
Selected Topics in Software Engineering sponsored by AICTE
and HRD of India held in March 18-24, 2010 at IIT, Kharagpur.
She has begged many such Certificates and achievements in her
5. Conclusion educational career. Her research interest includes Software
Engineering, Fuzzy System, Image Processing and Computer
The paper presents different steps involved in the Networks & Internet Technology.
development of a fingerprint based person
identification and verification system. And the
proposed fingerprint recognition system uses both
frequency and orientation information available in
fingerprint. Due to this , the existing system become