El4301832837

Mehala. G Int. Journal of Engineering Research and Applications www.ijera.com
ISSN : 2248-9622, Vol. 4, Issue 3( Version 1), March 2014, pp.832-837
www.ijera.com 832 | P a g e
Fingerprint Feature Extraction Algorithm
Mehala. G
Assistant Professor, Department of CSE, Hindusthan College of Engineering and Technology, Coimbatore-
641032.
ABSTRACT
The goal of this paper is to design an efficient Fingerprint Feature Extraction (FFE) algorithm to extract the
fingerprint features for Automatic Fingerprint Identification Systems (AFIS). FFE algorithm, consists of two
major subdivisions, Fingerprint image preprocessing, Fingerprint image postprocessing. A few of the challenges
presented in an earlier are, consequently addressed, in this paper. The proposed algorithm is able to enhance the
fingerprint image and also extracting true minutiae.
Keywords: Biometrics, Fingerprint Enhancement, Multistage Median Filter (MMF), Top & Miss Transformation,
Minutiae
I. INTRODUCTION
Biometrics is the science of identifying or
verifying the identification of a person based on
unique physiological and behavior characteristics.
Fingerprints were accepted formally as valid personal
identifier in the early twentieth century and have
since then become a de-facto authentication
technique in law enforcement agencies worldwide.
The FBI currently maintains more than 400 million
Fingerprint records on file. Fingerprints have several
advantages over other biometrics, such as following:
Highly universality, Highly distinctiveness, High
permanence, Easy collect ability, High performance,
Wide acceptability.
However, verification usually relies
exclusively on minutiae features. Most automatic
systems for fingerprint comparison are based on
minutiae matching. Minutiae are local features
marked by ridge discontinuities. There are about 18
distinct types of minutiae features that include ridge
endings, bifurcation, crossovers and islands. Minutiae
are minute details of the fingerprint and they are
shown in Fig.1.Automatic minutiae detection is an
extremely critical process, especially in low quality
fingerprints were noise and contrast deficiency can
originate pixel configurations similar to minutiae or
hide real minutiae. A ridge ending occurs when the
ridge flow abruptly terminates and a ridge bifurcation
is marked by a fork in the ridge flow.
Fig.1. Minutiae Types
Most matching algorithms do not even
differentiate between these two types since they can
easily get exchanged under different pressures during
acquisition.
In this paper, multiple sources of
information are consolidated to enhance the
performance of automatic fingerprint identification
systems. A critical step in automatic fingerprint
matching is to automatically and reliably extract
minutiae from the input fingerprint images. However,
the performance of a minutiae extraction algorithm
relies heavily on the quality of the input fingerprint
images. In order to ensure that the performance of an
automatic fingerprint identification/verification
system will be robust with respect to the quality of
the fingerprint images, it is essential to incorporate a
fingerprint enhancement algorithm in the minutiae
extraction module. To deal with this, the proposed
Fingerprint Feature Extraction algorithm will perform
RESEARCH ARTICLE OPEN ACCESS

Preprocessing method and Post processing method. A
series of image enhancement is applied to the
fingerprint image in Preprocessing method. With the
use of enhanced fingerprint image, extract the true
minutiae points in Post processing method
The Paper is organized as follows.
Preprocessing method is presented in Section 2, Post
processing method is presented in Section 3.
II. PREPROCESSING
In this section, a series of process is applied
to the fingerprint image.
Fig.2. Steps involved in Preprocessing
Local histogram equalization is used for
contrast expansion and Multistage Median
Filtering(MMF) for noise reduction. The binarization
process is applied by adaptive thresholding based on
the local intensity mean. Thinning is done using
morphological thinning operation, Hit and Miss
Transformation.
2.1 LOCAL HISTOGRAM EQUALIZATION:
Histogram equalization defines a mapping
of gray level p into gray levels q such that the
distribution of gray level q is uniform. The mapping
stretches the contrast (expands the range gray levels)
for gray levels near the histogram maximum. Since
contrast is expanded for most of the image pixels, the
transformation improves the detect ability of many
image features. The probability density function of a
pixel intensity level k , r is given by the equation,
n
n
rp k
kr )( (2-1)
Where: 0<= kr , k=0,1,2…255,. kn is the number of
pixels at intensity level kr and n is the total number
of pixels. The histogram is derived by plotting
)( kr rp against kr . A new intensity ks of level k is
defined as:
 


k
j
k
j
jr
j
k rp
n
n
s
0
0
)( (2-2)
We apply the histogram equalization locally
by using a local window of 11x11 pixels. This results
in expanding the contrast locally, and changing the
intensity of each pixel according to its local
neighborhood. Figure 2.1 (right) present the
improvement in the image contrast obtained by
applying the local histogram equalization. Histogram
equalization is to expand the pixel value distribution
of an image so as to increase the perceptional
information.
Figure 2.1 original image and its Histogram
(left) and after equalization (right)
Thus Histogram equalization produces the
contrast enhanced fingerprint image for the
Multistage Median Filter (MMF) noise removal
process.
2.2 MULTISTAGE MEDIAN FILTER:
Performs filtering of a video stream using
the multistage implementation of the median filter.
The same procedure as in Moving Average Filter is
followed, with the difference that only one parameter
is given for the 3D window size, as it can only be
cubic.
Pixels in a low bit-rate encoded image are
categorized into two or more categories. In one
embodiment three pixel categories are used: 1) pixels
belonging to quasi-constant (QC) regions where the
pixel intensity values vary slowly 2) pixels
belonging to textured regions which are characterized
by many small edges and thin-line signals portions),
and 3) pixels belonging to dominant-edge (DE)
Local Histogram Equalization
Multistage Median Filter
Binarization
Hit & Miss Transformation

regions which are characterized by few sharp and
dominant edges. In one embodiment dominant-edge
pixels are categorized first, then quasi-constant and
textural pixels are distinguished from the remaining
pixels using the number of zero-crossings among the
pixels and a threshold. Conventional spatial filters
that are well suited for each pixel category type are
then used to enhance each region of the image. In one
embodiment, various combinations of spatial filters
are used to enhance the image.
The median filter has been frequently used
in image processing. Although it does a good job at
suppressing noise, features such as thin lines are
removed.
Let {x(.,.)} be a discrete two-dimensional
sequence, and consider the set of elements inside a
(2N+1)x(2N+1) square window W centered at the
(i,j)th
pixel.
Define the following four subsets of the
window W:
W0,1(i,j) ={x(i,j+k); -N≤k ≤N} (1)
W1,1(i,j) ={x(i+k,j+k);-N≤k ≤N} (2)
W1,0(i,j) ={x(i+k,j); -N≤k ≤N} (3)
W1,-1(i,j) ={x(i+k,j-k);-N≤k ≤N} (4)
Suppose that zs(i,j) (s=1,2,3,4) are the median values
of the elements in the four subsets, respectively, and
Yp(i,j)=min[z1(i,j), z2(i,j), z3(i,j), z4(i,j)] (5)
Yq(i,j)=max[z1(i,j), z2(i,j), z3(i,j), z4(i,j)] (6)
Then the output of the multistage median
filter (MMF) is defined by
Ym(i,j)=med[yp(i,j), yq(i,j), x(i,j)] (7)
One of the advantages of the MMF is its
ability to preserve details of images. For example, the
three pixel long bright line can be preserved with a
3x3 window. The fingerprint filtered by Multistage
Median Filter is shown in Fig.3.
Fig.3. Filtered Image
2.3 BINARIZATION:
The operation that converts a grayscale
image into binary image is known as binarization.
Binarization refers to the process of reducing the bit
depth of the grayscale image to just 1 bit per pixel.
The straight forward approach for binarization relies
on choosing a global threshold T. The binarized
image I is obtained as follows.



 

otherwise
MeanLocalnnIif
nn
old
new
0
),(1
),(
21
21
(2-3)
This is a very well studied problem in the
area of image processing and computer vision. There
exists optimal approach to determine this threshold T
based on the statistics of the gray scale distribution
within the image. However, due to the non-stationary
nature of the fingerprint, such algorithms are not
suitable or are sub-optimal at best. The brightness of
the image may vary across different regions of the
fingerprint image. Therefore an adaptive binarization
is often used for this purpose. In such techniques, the
threshold T is determined locally considering the
properties of the local neighborhood.
The local adaptive thresholding is applied to
the directionally filtered image, which produces the
final enhanced binary image. This involves
calculating the average of the gray-level values
within an image window at each pixel, and if the
average is greater than the threshold, then the pixel
value is set to a binary value of one; otherwise, it is
set to zero. The gray-level image is converted to a
binary image, as there are only two levels of interest,
the foreground ridges and the background valleys.
The fingerprint images possess ridge flow
patterns with slowly changes in directions. They may
have various gray-level values due to non-uniformity
of the ink intensity, non-uniform contrast with the
sensors by users or changes in illumination and
contrast during image acquisition process. Obviously,
Global thresholding method fails to create good
quality binary images for further feature extractions.
The resultant image of the given fingerprint image is
shown in Fig.4.
Fig.4. Binarized Fingerprint Image
We carried out the binarization process
using an Adaptive threshold that separates the
foreground from the background with non uniform
illumination. In this approach, each pixel is
successively examined and assigned white and black
value. This is of particular importance, since the
binarization step is very critical to the process of
minutiae detection. Errors in this stage propagate into
subsequent phases of feature extraction. Therefore,
the binarization process has to be robust and resilient
to the quality of input image.
2.4 THINNING
The final image enhancement step typically
performed prior to minutiae extraction is thinning.
Thinning is a morphological operation that

successively erodes away the foreground pixels until
they are one pixel wide. A standard thinning
algorithm, Hit-or-Miss Transformation is employed,
which performs the thinning operation using two sub
iterations. Each sub iteration begins by examining the
neighbourhood of each pixel in the binary image, and
based on a particular set of pixel-deletion criteria, it
checks whether the pixel can be deleted or not. These
sub iteration continues until no more pixel can be
deleted.
The thinning of a Binarized image is
achieved by using the structuring elements{B} which
is shown in Fig.5.
{B} = {B1
,B2
,….…B8
}
The process of thinning is done by the
following equation,
(Binarized Image ¤ B1
) ¤ B2
…..¤ B8
(2.4)
B1
B2
……………………………….. … B8
Fig.5. Structure Elements
The application of the thinning algorithm to
a fingerprint image preserves the connectivity of the
ride structures while forming a skeleton version of
the binary image. This skeleton image is then used in
the subsequent extraction of minutiae. Automatically
detected minutiae may be slightly perturbed from its
original location because of the noise introduced
during the binarizing and thinning processes. Hence
we should apply the thinning process without
affecting the minutiae locations. The resultant of
thinning process is shown in Fig.5.1.
(a) Original image (b) Binary image (c) Thinned
image
Fig.5.1. Fingerprint Image after Applying
Binarization and Thinning
III. POSTPROCESSING:
The post processing stage of the proposed
algorithm involves the detection of true minutiae, the
ridge ending and bifurcation.
3.1 DETECTION OF TRUE MINUTIAE
The candidate minutiae are detected in a
single scan of the thinned ridge map image using 3 x
3 windows. Each ridge pixel (value 1) is classified
based on the number of 0 to 1 transitions meet when
making full clockwise trip along its 8 neighborhood
pixels. If there is only one such transition then the
pixel is classified as a candidate ridge ending. If the
number of transition is 3 then the pixel is classified as
a candidate ridge bifurcation. In all other cases the
pixel is not included into the candidate minutiae set.
The number of candidate minutiae detected
as above is much larger than the number of genuine
minutiae in the fingerprint image. Most of the
candidate minutiae have no meaning corresponding
to the fingerprint image, they being associated with
different false minutiae structures. The post
processing algorithm present in the following tests
the validity of each candidate minutiae by analyzing
the thinned ridge map image in W x W neighborhood
of the minutiae.
3.2 ALGORITHM
For each candidate minutiae (ridge ending
and ridge bifurcation):
1. Create and initialize with 0 am image L of size
W x W. each pixel of L corresponds to a pixel of
thinned image which is located in a W x W
neighborhood centered in the candidate minutiae.
Fig.6. Example of Minutia validations.
2. Label with -1 the center pixel of L (Fig. 6(a),
7(a)). This is the pixel corresponding to the
candidate minutiae point in the thinned ridge
map image.
3. if the candidate minutiae is ridge ending then:
a) Label with 1 all the pixel in L with correspond
to the pixel connected with the candidate ridge
ending in the thinned ridge map
image(Fig.6(b)).
b) Count the number of 0 to 1 transition ( 01 )
met when making full clock wise trip along the
border of the L image (Fig.6(c)).
c) If 01 =1, then validate the candidate minutiae
as true ridge ending.

Fig.7. Changes in the L images after different steps of
the algorithm.
4. if the candidate minutiae is a ridge bifurcation
then:
d) Make a full clockwise trip along its 8
neighborhood pixels of the candidate ridge
bifurcation, and label in L with 1,2 and r
respectively the three connected components
met during this trip(Fig. 7(b)).
e) For l=1,2,3 (Fig. 7 (c),(d),(e), label with l all
pixels in L which:
i. Have the label 0;
ii. Are connected with an l labeled pixels;
iii. Correspond to 1 valued pixels in thinned
ridge map;
f) Count the number of 0 to 1, 0 to 2, 0 to 3
transitions met while making a full clockwise
trip along the border of the L image. The
above three numbers are denoted by 01 , 02
and 03 respectively as shown in (Fig.7(f)).
g) If 01 =1 ^ 02 =1^ 03 =1, then validate the
candidate minutiae as a ridge bifurcation.
The dimension W of the neighborhood
analyzed around each candidate minutiae is chosen
larger then two times the average distance between
two neighborhood ridges.
IV. FINAL RESULS PREPROCESSING:
A.Histogram Equalized B. Multistage Median
Image Filtered Image
C. Binarized Image D.Thinned Image
Fig.8. Output of Preprocessing
POSTPROCESSING:
A.True Ridge End B. True Ridge
Bifurcation
Fig.9. Output of Post processing
V. CONCLUSION
The primary goal of Fingerprint Feature
Extraction(FFE) is the enhancement of the fingerprint
images using Multistage Median Filter and also
extract true minutiae from the thinned fingerprint
images. Firstly, this paper focused a series of
techniques for fingerprint image enhancement to
facilitate the extraction of minutiae. Experiments
were then conducted using a combination of both
good test images and poor fingerprint images in order
to provide a well balance evolution on a more
qualitative and accurate measure of the performance.
Whereas poor images rely on qualitative measures of
inspection, but can provide a more realistic
evaluation as they provide a natural representation of
fingerprint imperfection such as noise and corrupted
elements.
The final results have shown that combined
with an accurate estimation of the minutiae, the
multistage median filter is able to effectively enhance
the clarity of ridge structures while reducing noise. In
contrast, for low quality images that exhibit high
intensities of noise, the filter is highly effective in

enhancing the image.
Overall, the results have shown that the
implemented enhancement algorithm is a useful step
to employ prior to minutiae extraction. Thus the
proposed Fingerprint Feature Extraction (FFE)
algorithm implements a set of reliable techniques for
fingerprint image enhancement and minutiae
extraction.
REFERENCES
[1] I. Emiroglu, M.B. Akhan, University of
Hertforshire, UK, “Pre-Processing of
Fingerprint Images” in European
Conference on Security and Detection, 28-
30 April 1997, Conference Publication No.
437.
[2] Marius Tico, Pauli Kuosmanen, Tamper
University of Technology, Finland, “An
Algorithm for fingerprint Image
Postprocessing”, 0-7803-6514-3/00.
[3] Virginia Espinosa, Polytechnic University of
Catalonia, “ Fingerprints Thinning
Algorithm”, IEEE AES Systems Magazine,
September 2003.
[4] Virginia Espinosa-Dur6, Polytechnic
University of Catalonia, “Minutiae
Detection Algorithm for Fingerprint
Recognition”, CICYT TIC2000-1669-C04-
02.
[5] D. Simon-Zorita, J. Ortega-Garcia, S. Cruz-
Llanas and J. Gonzalez-Rodriguez,”
Minutiae Extraction scheme for fingerprint
recognition systems” 0-7803-6725-1.
[6] J C Amengual, A Juan, J C Perez, F Prat, S
Saez and J M Vilar, “ Real time minutiae
extraction in fingerprint images.

El4301832837

More Related Content

What's hot (20)

Viewers also liked (20)

Similar to El4301832837 (20)

Recently uploaded (20)

El4301832837