Fingerprint Matching Using Hough Transform

Abstract Biometric systems operate on behavioral and

physiological biometric data to identify a person. The physiological
characteristics such as face, fingerprint, and palm print and iris
remains unchanged throughout the life time of a person. The
biometric system operates as verification mode or identification mode
depending on the requirement of an application. Fingerprint is a
pattern of ridges, furrows and minutiae, which are extracted using
inked impression on a paper or sensors. A good quality fingerprint
contains 25 to 80 minutiae depending on sensor resolution and finger
placement on the sensor. In latent fingerprint matching case, it is very
challenging to estimate the orientation field based only on the image
due to the poor quality and small area of the latent. In this paper, we
propose a new fingerprint matching algorithm which is especially
designed for matching latent. The proposed algorithm modifies the
orientation based minutia to the texture descriptor by adding the ridge
frequency information and histogram equalization is also employed
in improving the intensity of the latent images. The experimental
results are performed on two different latent databases, NIST SD27
and WVU latent databases and the proposed algorithm is written in
mat lab.


Keywords- latent,minutiae,databases-NI STSD27,orientation field

Introduction

Fingerprints remain unique for every individual and thus it
plays a major role in identification of the culprits in forensics.
The latent fingerprint matching algorithms proposed includes
the image enhancement; feature extraction and matching of
the latents.The different characteristics of the images which
give a prominent difference between the databases are
identified and extracted. The accuracy of the fingerprint
matching algorithms depends on the image quality, image
enhancement methods and feature extraction algorithms. In
latent search, a latent print developed from a crime scene is
searched against the fingerprint database of known persons. It
is the matching between latent and rolled/plain fingerprints
that is of the utmost importance to apprehend suspects in
forensics.
1) Rolled, which is obtained by rolling the finger nail-
to-nail either on a paper (in this case ink is first
applied to the finger surface) or the platen of a
scanner.
2) Plain, this is obtained by placing the finger flat on a
paper or the platen of a scanner without rolling.
3) Latent, which are lifted from surfaces of objects that
are inadvertently touched or handled by a person
typically at crime scenes.









Minutias are the discontinuities of the ridges.
Endings, the points at which a ridge stops.
Bifurcations, the point at which one ridge divides into two.
Dots, very small ridges.
Ponds or lakes, empty spaces between two ridges.
Fingerprints are graphical patterns of ridges and valleys on the
surface of fingertips. One kind of widely-used features is
called minutiae, which is usually defined as the ridge ending
and the ridge bifurcation.


LITERATURE SURVEY


Anil K. Jain, Fellow, IEEE, and Jianjiang Feng,
Member, IEEE [1] in 2011 proposed a system for matching
latent fingerprints found at crime scenes to rolled fingerprints
enrolled in law enforcement databases. The proposed system
utilized the feature extraction which encapsulated the features
like reference points (singularity), overall image
characteristics (ridge quality map, ridge flow map, and ridge
wavelength map), minutiae, and skeleton. Ridge flow map,
ridge wavelength map, and ridge quality map are obtained by
dividing the image into no overlapping blocks of size 16 * 16
and assigning single orientation, wavelength, and quality value
to each block. We define three quality levels for a block: level
(background), level 1 (clear ridge flow and unreliable
minutiae), and level 2 (clear minutiae). A skeleton is a one-
pixel-wide ridge, which is traced in the thinned image and
represented as a list of points. Secondary features (dots,
incipient ridges, and pores) are represented as a set of points.
To match various combinations of features, the baseline
matching algorithm was proposed which not only serves as a
matcher for minutiae-only templates, but also serves as a
framework to match and fuse various extended features.
The baseline matching algorithm takes only minutiae
as input and consists of the following steps:
Local minutiae matching: Similarity between each minutia of
latent fingerprint and each minutia of rolled fingerprint is
computed.
Global minutiae matching: Using each of the five most similar
minutia pairs found in initial step as an initial minutia pair, a
greedy matching algorithm is used to find a set of matching
minutia pairs.
Matching score computation: A matching score is computed
for each set of matching minutia pairs and the maximum score
is used as the matching score between the latent and rolled
prints.
Automated Fingerprint Identification Systems
(AFISs) have played an important role in many forensics and
civilian applications. There are two main types of searches in
forensics AFIS: tenprint search and latent search. In tenprint
search, the rolled or plain fingerprints of the 10 fingers of a
subject are searched against the fingerprint database of known
persons. In latent search, a latent print developed from a crime
scene is searched against the fingerprint database of known
persons. It is the matching between latents and rolled/plain
fingerprints that is of the utmost importance to apprehend
suspects in forensics. The manual latent identification process
can be divided into four steps, namely, analysis, comparison,
evaluation, and verification.
Alessandra A. Paulino, Student Member, IEEE,
Jianjiang Feng, Member, IEEE [2] proposed algorithm uses a
robust alignment algorithm (descriptor-based Hough
transform) to align fingerprints and measures similarity
between fingerprints by considering both minutiae and
orientation field information. To be consistent with the
common practice in latent matching (i.e., only minutiae are
marked by latent examiners), the orientation field is
reconstructed from minutiae. The proposed matching approach
uses minutiae and orientation field from both latent and rolled
prints. Minutiae are manually marked by latent examiners in
the latent, and automatically extracted using commercial
matchers in the rolled print. Based on minutiae, local minutiae
descriptors are built and used in the proposed descriptor-based
alignment and scoring algorithms. Orientation field is minutia
consists of five attributes, namely, x and y coordinates,
minutiae direction, type, and quality. The quality of minutia is
defined to have two levels: 0 (unreliable) and 1 (reliable). A
skeleton is a one-pixel-wide ridge, which is traced in the
thinned image and represented as a list of points. Secondary
features (dots, incipient ridges, and pores) are represented as a
set of points. The baseline matching algorithm is not only a
matcher for minutiae-only templates, but also serves as a
framework to match and fuse various extended features.
Eryun Liu, Anil K. Jain, Fellow, IEEE, and Jie Tian,
Fellow, IEEE [3] proposed a coarse to fine matching strategy
based on minutiae clustering and minutiae match propagation
is designed specifically for palm print matching. To deal with
the large number of minutiae, a local feature-based minutiae
clustering algorithm is designed to cluster minutiae into
several groups such that minutiae belonging to the same group
have similar local characteristics. The coarse matching is then
performed within each cluster to establish initial minutiae
correspondences between two palm prints. Starting with each
initial correspondence, a minutia match propagation algorithm
searches for mated minutiae in the full palm print. The
proposed palm print matching algorithm has been evaluated
on a latent-to-full palm print database consisting of 446 latent
and 12,489 background full prints. The minutiae clustering
algorithm is proposed to significantly reduce latent-to-full
palm print matching complexity by avoiding the similarity
computation between minutiae with different local
characteristics, which are deemed not to be matched.
Jianjiang Feng and Jie Zhou IEEE in 2011 [4]
proposed different minutia descriptors have been proposed in
the literature, which can be coarsely classified into three
categories: image-based, texture-based, and minutia-based.
These descriptors have not been evaluated systematically. A
typical minutia matching algorithm consists of two stages:
local matching and global matching. In the local matching
stage, each minutia in the first fingerprint (query) is compared
to each minutia in the second fingerprint (template) to find
candidate minutia pairs. In the global matching stage, minutiae
sets of the two fingerprints are aligned and a set of minutia
pairs is determined as the final matching minutiae pairs. Local
matching is very important for the final matching result.
Ideally, local matching should output all correct pairs and as
few as possible false pairs with low computational complexity.
Jianjiang Feng, Member, IEEE, Yuan Shi, and Jie
Zhou, Senior Member, IEEE [5] propose improved versions of
the basic algorithm for two special but frequent cases1) the
mated template fingerprint of one component fingerprint is
known and 2) the two component fingerprints are from the
same finger. We propose a robust orientation field estimation
algorithm (called the basic algorithm) for latent overlapped
fingerprints whose core is the constrained relaxation labeling
algorithm. The separating algorithm consists of the following
steps:
1) Region segmentation: The overlapped fingerprint image is
divided into background region, overlapped region, no
overlapped regions of two component fingerprints.
2) Initial orientation field estimation: One dominant
orientation is estimated in the no overlapped region while two
dominant orientations are estimated in the overlapped region.
3) Orientation field separation: The initial orientation field is
separated into two component orientation fields.
Joshua Abraham, Paul Kwan and Junbin Gao [6]
presented the concept of hybrid shape and orientation
descriptor in their paper Fingerprint Matching using A
Hybrid Shape and Orientation Descriptor. Spurious and
missing minutiae can both introduce errors in minutiae
correspondence. The hybrid descriptor can effectively filter
out spurious or unnatural minutiae pairings while
simultaneously using the additional ridge orientation cues in
improving match score calculation. Fingerprint matching
algorithms largely follow 3 different classes: correlation-
based, minutiae-based, and non-minutiae feature based
matching. Correlation-based matching involves superimposing
2 fingerprint images together and calculating pixel-wise
correlation for different displacement and rotations. Minutia-
based matching uses extracted minutiae from both fingerprints
in order to help perform alignment and retrieve minutiae
pairings between both fingerprint minutiae sets. Non-minutiae
feature based matching use non-minutiae features, such as
ridge shape, orientation and frequency images in order to
perform alignment and matching.
Jinwei Gu, Student Member, IEEE, Jie Zhou, Senior Member,
IEEE, and Chunyu Yang [7] proposed a method of combining
the decisions of the matchers based on the global structure
(orientation field) and the local cue (minutiae). This paper
proposes a novel representation for fingerprints which
includes both minutiae and model-based orientation field.
Here the matching is based only on the orientation field. If the
two fingerprints of a given matching are from a same finger, it
is a genuine matching, otherwise an imposter matching. The
performance of a biometric system operating in verification
mode can be specified in terms of false acceptance rate (FAR)
and false rejection rate (FRR). Given a threshold, FAR is
defined as the percentage of that imposter matchings whose
scores are greater than the threshold, and FRR is defined as
the percentage of those genuine matchings whose scores are
less than the threshold. The receiver operating curves (ROC)
plotting FAR versus FRR under different thresholds is often
used to evaluate the systems performance.


I. PROPOSED METHOD



Fig: Overview of the proposed method


This is the block representation of the proposed method. The
latent fingerprints uplifted from the crime scene are enhanced
by using the technique of histogram equalization. Minutiae are
manually marked by latent examiners in the latent, and
automatically extracted using commercial matchers in the
rolled print. Based on minutiae, local minutiae descriptors are
built and used in the proposed descriptor-based alignment and
scoring algorithms. Orientation field is reconstructed from
minutiae location and direction for the latent. The eight
different typical features of the fingerprints are extracted by
using the minutiae extraction algorithms. As the features are
extracted for the matching of the rolled prints available with
the agencies, the matching of these features with the original
prints involves the utilization of different adaptive algorithms.
Orientation field can be used in several ways to improve
fingerprint matching performance, such as by matching
orientation fields directly and fusing scores with other
matching scores, or by enhancing the images to extract more
reliable features. Fingerprint alignment or registration consists
of estimating the parameters (rotation and translation) that
align two fingerprints. There are a number of features that may
be used to estimate alignment parameters between two
fingerprints, including singular points, orientation field,
ridges, and minutiae. There are also a number of methods to
align two fingerprints: Generalized Hough Transform, local
descriptors, energy minimization. An alignment method for
minutiae matching that estimates rotation, scale, and
translation parameters using a Hough Transform. Given two
sets of points (minutiae), a matching score is computed for
each transformation in the discretised set of all allowed
transformations.

DESIGN PHASE
Minutiae extraction: The two main methods of minutiae
feature extraction either require the gray-scale image to be
converted to a binary image, or work directly on a raw or
enhanced gray-scale image. In the binary image based method,
the binarization of the gray-scale image is the initial step. This
requires each gray-scale pixel intensity value to be
transformed to a binary intensity of black (0) or white (1). The
simplest approach is to apply a global threshold where each pixel is mapped according to
I(x, y) = 1 if I(x, y) t,
0 Otherwise








tilization of different adaptive algorithms. Orientation field can be used in several ways to improve fingerprint
matching performance, such as by matching orientation fields directly and fusing scores with other matching scores,
or by enhancing the images to extract more reliable features. Fingerprint alignment or registration consists of
estimating the parameters (rotation and translation) that align two fingerprints. There are a number of features that
may be used to estimate alignment parameters between two fingerprints, including singular points, orientation
field, ridges, and minutiae. There are also a number of methods to align two fingerprints: Generalized Hough
Transform, local descriptors, energy minimization. An alignment method for minutiae matching that estimates
rotation, scale, and translation parameters using a Hough Transform. Given two sets of points (minutiae), a matching
score is computed for each transformation in the discretised set of all allowed transformations.

MATCHING
Fingerprints from the latent databases are not catagorised to the different classes and hence to find the specified
class to which it belongs we have adopted a method of training the fingerprints from the known database and thus
give the test vector. Fig 5 represents the known fingerprints belonging to three different classes. Now the image
from each class is trained and the corresponding matching of the latent is done by using the kNN classifier.

Hough Transform
The Hough transform is a feature extraction technique used in image analysis, computer vision, and digital image
processing. The purpose of the technique is to find imperfect instances of objects within a certain class of shapes by a
voting procedure. This voting procedure is carried out in a parameter space, from which object candidates are obtained
as local maxima in a so- called accumulator space that is explicitly constructed by the
algorithm for computing the Hough transform. The classical Hough transform was concerned with the
identification of lines in the image, but later the Hough transform has been extended to identifying positions of
arbitrary shapes, most commonly circles or ellipses.

LATENT DATABASES
NIST Special Database 27 (NIST SD27): NIST Special Database 27 is the only publicly available database comprising
latent fingerprints from operational scenarios (latents collected at crime scenes). It consists of 258 latent fingerprint
images and 258 corresponding (mated) rolled prints. Both latents and rolled prints are available at 500 ppi. The
quality of the latent in NIST SD27 varies, reflecting the operational (casework) quality. NIST SD27 contains latent
prints of three different qualities, termed good, bad, and ugly, which were classified by latent examiners.
Although this classification of latent prints as good, bad, and ugly is subjective, it has been shown that such a
classification is correlated with the matching performance. Another indicator of fingerprint quality that affects
the matching performance is the number of minutiae in the latent print.


Summary

A fingerprint matching algorithm is designed for matching
latents to rolled or /plain fingerprints which is based on texture
based Hough transform alignment. The performance of the
proposed matcher is compared for two different latent
databases. The texture based descriptor is used to improve the
matching accuracy especially when the overlap between the
latents and rolled prints is more.