International Journal of Scientific Engineering and Technology

Volume No.5 Issue No.5, pp: 225-227

ISSN:2277-1581
01 May 2016

Fingerprint Authenication Using Graph Theory

Harsh P. Bajaj1,Arun Kodnani2 ,Harshit Singhal3
1

CSE,VIT University,Vellore,India
[email protected],[email protected],[email protected]
II. RELATED WORK

Abstract : In this paper we have tried to implement the idea is

to classify fingerprints through identifying the core of the
fingerprint. The fingerprint is taken as an input via an image.
The image is then transformed for necessary preprocessing.
Then we make a graph with ridges ending and bifurcations all
around the Centre, acting as vertices. Then we compare the
graph constructed using the connection of vertices. The idea is
to calculate distance between neighboring vertices for each
vertex. Our aim is to increase the efficiency of fingerprint
identification in different orientation.
Keywords-Ridges,Minutae,eccentricity,curvature of graph
I. INTRODUCTION
Fingerprint identification is the process of identifying a
person based on the patterns of his fingerprint denoted by the
patterns of the ridges on the surface of his finger. The ridges in
the image of a fingerprint are denoted by the black lines and the
spaces between them known as valleys and are denoted by white
spaces. Also one or more ridges in a fingerprint pattern are
connected[1].The unique factors of a fingerprint are: crossover,
core, bifurcation, ridge ending, island, delta and pore.
The two fundamental principles of fingerprints identification
are:
Immutability: The fingerprint pattern remains consistent for a
particular person throughout his life. They only change by
accident, skin disease or change of size.
Uniqueness: Fingerprint pattern develops in a purely random
manner during foetal development in the womb. There is
sufficient variability in the arrangement that no two patterns
match each other even the patterns of two different fingers of the
same person are unique.
The current research only matches fingerprints using neighbor
of a particular ridge to match the patterns or by reducing the
ridges to a smaller pixel value. These algorithms are not able to
overcome the problem of identification when there is some noise
in the image of the fingerprint because of dirt, scar etc[2]/.
This paper proposes a solution to overcome the problems of
orientation and noise of a fingerprint image during the process of
matching. The algorithm considers the image as a whole and
compares using minutiae.

IJSET@2016

One of the papers by HongleiWei[1] has proposed a fingerprint

matching algorithm which can withstand the elastic distortion
produced while taking a fingerprint. For fingerprint matching
they have used the ridge line which connects the minutia point
with the associated ridge point. Basically, algorithm is divided
into two stages based on the ridge lines.
In first stage, estimation of transformation is done by
matching the input with local quadrilateral sets. Here, the
quadrilateral is formed by connecting two ridges minutia point
and ridge points respectively. Input graph and quadrilateral sets
are matched by comparing their respective parameters.
In second stage, a coarse pair set is achieved by matching two
aligned ridge line sets in relatively slack matching qualifications,
and then graph matching algorithm is applied again to match
those coarse pair set again to remove the false matched ridge
lines which might be present due to slack matching
qualifications. The coarse pair set is represented as graph of input
image and quadrilateral template graph respectively. Each node
in the graph corresponds to a ridge line and each edge
corresponds to a quadrilaterals. A similarity graph can be
achieved by matching the two graphs, in which an edge between
two nodes corresponds to the matching score of corresponding
quadrilaterals composed with the corresponding ridge lines. Final
matching score is then calculated on the basis of both the
matching score of ridge line sets and the matching score of
orientation fields.
One another paper[2] has proposed core based algorithm for
Fingerprint verification. Most of the non-core based matching
algorithm is time consuming. Therefore, they are not suitable for
online application. So to increase the efficiency of algorithm,
firstly, they used core detection algorithm to get the core position,
and then defined some local structure of the core area. Using this
local Structure, we find some correspondent points of the two
fingerprint image. Secondly, we use the correspondent points in
the first stage to march the global feature of the fingerprint.
Experimental results showed that the performance of the
proposed algorithm is good but not the best.
Another paper[3] proposed a fingerprint minutia matching
technique, in which matching is done by using both the local and
global structures of fingerprint minutiae. The local structure of a
minutia describes a rotation and translation invariant feature of
the minutia in its neighborhood. It is used to find the

doi : 10.17950/ijset/v5s5/501

Page 225

International Journal of Scientific Engineering and Technology

Volume No.5 Issue No.5, pp: 225-227
correspondence of two minutiae sets and increase the
reliability of the global matching. While the global structure of
minutiae reliably determines the uniqueness of fingerprint.
Therefore, the local and global structures of minutiae together
provide a solid basis for reliable and robust minutiae matching.
The scheme is suitable for an on line processing due to its high
processing speed.
III. PROPOSED SOLUTION
Given an image of fingerprint take the higher and the
lower pixel values from it and then recognize the friction ridges
pattern. We are using binary grey values for the image. Here we
mark each and every ridge by a set of nodes for every pixel value.
The ridges having nodes will be connected by ensuring the
connectedness of the graph and we traverse in bottom up fashion.
When the graph begins to be disconnected then we start making
the new graph by taking nodes from its one end point to other [6].
Now, each ridge has become the connected path of the fingerprint
graph.

Then we determine the curvature of the ridges using

combinatorial curvature[7].The connected path and circuits where
f ~v indicates that vertex v is on the boundary of face f, and |f|
isthe number of vertices on the boundary of face f[7] Find the
combinatorial curvature of the graph from the last graph
constructed. The least curvature value for the set of graphs is
nearest to center. So, in order to find the center of the fingerprint
we take the lowest curvature ridge and below it on the one grey
value patch are marked as center. Then after finding the center
point we make a wheel graph around it with the help of
surrounding ridges. Then for matching purposes we use graceful
labeling to label each and every ridges node uniquely based on
the direction and distance from the center point[5]. A graceful
labeling is one in which labeling is done with distinct integers
from the set {0, 1, 2, ..., m} (where m is the number of edges)
such that when each edge vertex u to vertex is given the value | f
(u)-f (v) | the edges are uniquely labeled from set of numbers {1,
2, m}[7].Then the similar labeling is done in the graph after
angle of 20 and by this can get distance measure and direction of
ridges of fingerprint graph. Use the planarity of graphs to detect
the crossover. The graph are said to be planar if edges in the
plane do not intersect each other. The graph in non-planar if
edges intersect but they can be made non-intersecting with the
given edges and vertices. So with this if we detect intersecting
edges i.e. their non-planar we can say that there is a cross-over
for the two edges we store their labels. The node in the path
having highest eccentricity is known as ridge ending. As
eccentricity is labeled relative to distance from a particular vertex
to the others is maximum. Then we get the end points as their
eccentricity is maximum.

ISSN:2277-1581
01 May 2016

These end points of the ridges at each 20 and note their

graph label and edge labels. The lake is other important
component of minutiae of fingerprint. Lake can be determined by
using concept of multiple edges. The two adjacent edges having
deg(3) are considered. Then check their edge connectivity is two
then there is lake formed by these to vertices. The edge
connectivity gives the minimal number of edges required to make
graph disconnected. Degree of a node can be done by finding the
incident vertices on a vertex. The incidence can be calculated by
the number of non-parallel edges which ends up on the same
vertex. The number of incident vertices on a graph gives the
degree of that graph.
Fig. 1

For detection the stored edge labels of each parameter is taken.

Then the labels are compared with labels of test images[8]. The
noises on the image as discussed can be removed by comparing
matching each parameters edge labels up to 80% only so that
even off there is noise the test image nearly matches the stored
image based on basic parameters. The parameters are weighted
based on their uniqueness from person to person. There are three
major types of minutiae features: the ridge ending, the
bifurcation, and the dot (also called short ridge). The ridge
ending is, as indicated by the name, the spot where a ridge ends.
A bifurcation is the spot where a ridge splits into two
ridges. Spots are those fingerprint ridges that are significantly
shorter than other ridges. Also, the ridges in the fingerprint have a
specific grey value of 0 which will help distinguish from noises
like scars, ink and dust.

where, d(v,vi) is the distance of from farthest vertex vi[4].

IJSET@2016

doi : 10.17950/ijset/v5s5/501

Page 226

International Journal of Scientific Engineering and Technology

Volume No.5 Issue No.5, pp: 225-227
Fig. 2

ISSN:2277-1581
01 May 2016

fingerprint is taken at different orientations. Now after finding

the core of a fingerprint the wheel graph will be formed around
the core and labeling of each vertex will be done. This step will
help us to uniquely identify fingerprint of the respective person.
Since the presence of noises in the image is possible so 80%
match could be considered. Therefore it can be concluded that
the proposed algorithm is efficient way of fingerprint matching.
V. REFERENCES

IV. CONCLUSION
This study described a new fingerprint matching technique based
on identifying the core of a fingerprint. To find out the core of the
fingerprint we are finding the curvature of the ridges. Since the
finger print matching technique is based on core of a fingerprint
so it can efficiently recognize the fingerprint even if the

IJSET@2016

i. http://www.biometrics.gov/documents/fingerprintrec.pdf
ii. Isenor, D. K. &Zaky, S. G.Fingerprint
identification
using
graph
matching
Pattern Recogn., Elsevier Science Inc., 1986, 19, 113-122
iii. Honglei Wei, ZongyingOu and Jianxin Zhang. Fingerprint
Identification based on Ridge Lines and Graph Matching.
iv. Weiwei Zhang, Yangsheng Wang. Core-Based Structure
Matching Algorithm of Fingerprint Verification.
v. Xudong Jiang and Wei-Yun Yau. Fingerprint Minutiae
Matching Based on the Local And Global Structures.
vi. Werner lz and Walter G. Kropatsch. Graph Representation
of Fingerprint Topology. In DanijelSkocaj, editor, Computer Vision CVWW'04, pages pp. 5158, Ljubljana, Slovenija, February 2004.
Slovenien Pattern Recognition Society.
vii. West, Douglas B. Introduction to Graph Theory; 2ndedition.
Prentice Hall: Upper Saddle River, NJ. 2001. pgs 89-94
viii. Xudong Jiang and Wei-Yun Yau. Fingerprint Minutiae
Matching Based on the Local And Global Structures.
ix. M. Gromov, Hyperbolic groups, in Essays in Group Theory, S.
M. Gersten (Ed.), M.S.R.I. Publications, Springer Verlag, New York,
1987, pp. 75-263.
x. Raymond, J. W. & Willett, P. Maximum
common
subgraph isomorphism algorithms for the matching of chemical
structures Journal of Computer-Aided Molecular Design, 2002, 16,
2002.

doi : 10.17950/ijset/v5s5/501

Page 227