Fingerprint Recognition

Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)

Department of Computer Science & engineering
Indian Institute of technology, Kanpur

Project Report
Computer Vision and Image Processing (CS676)
Guided By
Prof. Simant Dube
Date: 12/11/2009
 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)

Abstract
Our Term Project is to study and implement a fingerprint recognition system based on Minutiae
based matching quite frequently used in various fingerprint algorithms and techniques. The
approach mainly involves extraction of minutiae points from the sample fingerprint images and
then performing fingerprint matching based on the number of minutiae pairings among two
fingerprints in question.

Our implementation mainly incorporates image enhancement, image segmentation, feature

(minutiae) extraction and minutiae matching. It finally generates a percent score which tells
whether two fingerprints match or not. The project is coded in MATLAB.

Computer Vision and Image Processing (CS676)Page 2

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)

Acknowledgement
We would like to express our sincere thanks and gratitude to Prof. Simant Dube for his
suggestions, help and support. Also we take this opportunity to thank Prof. Amitabha Mukerjee
for his valuable comments and feedback during our project presentations.

We would also like to appreciate our course TA’s Amit Kumar Gupta and Rahul Gupta for the
support.

Computer Vision and Image Processing (CS676)Page 3

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)

Table of Contents
1 Introduction 5
1.1 What is A Fingerprint 5
1.2 What is Fingerprint Recognition 5
1.3 Techniques for Fingerprint matching 6

2 Our Implementation 6
2.1 Design Description 6

3 Minutiae Extraction 7
3.1 Fingerprint Image Enhancement 7
3.2 Fingerprint Image Segmentation 9
3.3 Final Minutiae Extraction 10

4 Minutia matching 12
4.1 Minutiae Alignment 12
4.2 Minutiae Match 13

5 Experimentation Results 13
5.1 Performance Evaluation Indexes 13
5.2 Experiment Analysis 13

6 Conclusion 14

References 14

Computer Vision and Image Processing (CS676)Page 4

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)

1 Introduction
Fingerprint recognition or fingerprint authentication
refers to the automated method of verifying a match
between two human fingerprints. Fingerprints are one
of many forms of biometrics used to identify an
individual and verify their identity. Because of their
uniqueness and consistency over time, fingerprints
have been used for over a century, more recently
becoming automated (i.e. a biometric) due to (a) (b)
advancement in computing capabilities. Fingerprint Figure 1.2(a) two important minutia features
identification is popular because of the inherent ease (b) Other minutiae features
in acquisition, the numerous sources (ten fingers)
1.2 What is Fingerprint Recognition?
available for collection, and their established use and
Fingerprint recognition (sometimes referred to as
collections by law enforcement and immigration. dactyloscopy) is the process of comparing questioned
and known fingerprint against another fingerprint to
1.1 What is a Fingerprint? determine if the impressions are from the same finger
A fingerprint is the feature pattern of one finger (Figure or palm. It includes two sub-domains: one is fingerprint
1.1). It is an impression of the friction ridges and verification and the other is fingerprint identification
furrows on all parts of a finger. These ridges and (Figure 1.3). In addition, different from the manual
furrows present good similarities in each small local approach for fingerprint recognition by experts, the
window, like parallelism and average width. fingerprint recognition here is referred as AFRS
(Automatic Fingerprint Recognition System), which is
program-based.

Figure 1.1 Fingerprint image from a sensor

However, shown by intensive research on fingerprint
recognition, fingerprints are not distinguished by their
ridges and furrows, but by features called Minutia,
which are some abnormal points on the ridges (Figure Figure 1.3 Verification vs. Identification
1.2). Among the variety of minutia types reported in However, in all fingerprint recognition problems, either
literatures, two are mostly significant and in heavy verification(one to one matching) or identification(one
usage: to many matching), the underlining principles of well
 Ridge ending - the abrupt end of a ridge defined representation of a fingerprint and matching
 Ridge bifurcation - a single ridge that divides into remains the same.
two ridges

Computer Vision and Image Processing (CS676)Page 5

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)

1.3 Fingerprint matching techniques

The large number of approaches to fingerprint
matching can be coarsely classified into three families.
• Correlation-based matching: Two fingerprint
images are superimposed and the correlation
between corresponding pixels is computed for
different alignments (e.g. various displacements
(a) (b)
and rotations).
Figure 2.1(a) Ridge Ending, (b) Ridge Bifurcation
• Minutiae-based matching: This is the most popular The outline of our approach can be broadly classified
and widely used technique, being the basis of the
into 2 stages - Minutiae Extraction and Minutiae
fingerprint comparison made by fingerprint matching. Figure 2.2 illustrates the flow diagram of the
examiners. Minutiae are extracted from the two
same.
fingerprints and stored as sets of points in the two-
dimensional plane. Minutiae-based matching
essentially consists of finding the alignment
between the template and the input minutiae sets
that results in the maximum number of minutiae
pairings
YES /NO
• Pattern-based (or image-based) matching: Pattern
based algorithms compare the basic fingerprint
patterns (arch, whorl, and loop) between a
previously stored template and a candidate Figure 2.2 System Flow Diagram
fingerprint. This requires that the images be The system takes in 2 input fingerprints to be matched
aligned in the same orientation. To do this, the and gives a percentage score of the extent of match
algorithm finds a central point in the fingerprint between the two. Based on the score and threshold
image and centers on that. In a pattern-based match value it can distinguish whether the two
algorithm, the template contains the type, size, fingerprints match or not. The input fingerprints are
and orientation of patterns within the aligned taken from the database provided by FVC2004
fingerprint image. The candidate fingerprint image (Fingerprint Verification Competition 2004).
is graphically compared with the template to
determine the degree to which they match. 2.1 Design Description
The above system is further classified into various
In Our project we have implemented a minutiae based
modules and sub-modules as given in Figure 2.3.
matching technique. This approach has been
Minutia extraction includes Image Enhancement,
intensively studied, also is the backbone of the current
Image Segmentation and Final Extraction processes
available fingerprint recognition products.
while Minutiae matching include Minutiae Alignment
and Match processes.
2 Our Implementation
We have concentrated our implementation on
Minutiae based method. In particular we are interested
only in two of the most important minutia features i.e.
Ridge Ending and Ridge bifurcation. (Figure 2.1)

Computer Vision and Image Processing (CS676)Page 6

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)
The first step in the minutiae extraction stage is
Fingerprint Image enhancement. This is mainly done to
improve the image quality and to make it clearer for
further operations. Often fingerprint images from
various sources lack sufficient contrast and clarity.
Hence image enhancement is necessary and a major
challenge in all fingerprint techniques to improve the
accuracy of matching. It increases the contrast
between ridges and furrows and connects the some of
Match the false broken points of ridges due to insufficient
amount of ink or poor quality of sensor input.
In our project we have implemented three techniques:
Histogram Equalization, Fast Fourier Transformation
and Image Binarization.
3.1.1 Histogram Equalization
Histogram equalization is a technique of improving the
global contrast of an image by adjusting the intensity
Figure 2.3 Detailed Design Description distribution on a histogram. This allows areas of lower
Under image enhancement step Histogram local contrast to gain a higher contrast without
Equalization, Fast Fourier Transformation increases the affecting the global contrast. Histogram equalization
quality of the input image and Image Binarization accomplishes this by effectively spreading out the most
converts the grey scale image to a binary image. frequent intensity values. The original histogram of a
Then image segmentation is performed which extracts fingerprint image has the bimodal type (Figure 3.1(a)),
a Region of Interest using Ridge Flow Estimation and the histogram after the histogram equalization
MATLAB’s morphological functions. occupies all the range from 0 to 255 and the
Thereafter the minutia points are extracted in the Final visualization effect is enhanced (Figure 3.1(b)).
Extraction step by Ridge Thinning, Minutia Marking and The result of the histogram equalization is shown in
Removal of False Minutiae processes. figure 3.2.
Using the above Minutia Extraction process we get the
Minutiae sets for the two fingerprints to be matched.
Minutiae Matching process iteratively chooses any two
minutiae as a reference minutia pair and then matches
their associated ridges first. If the ridges match well,
two fingerprint images are aligned and matching is
conducted for all remaining minutia to generate a
Match Score.
(a) (b)
Figure 3.1(a) Original histogram, (b) Histogram after
3 Minutiae Extraction equalization
As described earlier the Minutiae extraction process
includes image enhancement, image segmentation and
final Minutiae extraction.

3.1 Fingerprint Image Enhancement

Computer Vision and Image Processing (CS676)Page 7

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)

(a) (b) (a) (b)

Figure 3.2(a) Original Image, (b) Enhanced Image after Figure 3.3(a) Enhanced Image after FFT, (b) Image before FFT
histogram equalization The enhanced image after FFT has the improvements
as some falsely broken points on ridges get connected
3.1.2 Fast Fourier Transformation
and some spurious connections between ridges get
In this method we divide the image into small
removed.
processing blocks (32 x 32 pixels) and perform the
3.1.3 Image Binarization
Fourier transform according to equation:
Image Binarization is a process which transforms the 8-
bit Gray image to a 1-bit image with 0-value for ridges
and 1-value for furrows. After the operation, ridges in
(1)
the fingerprint are highlighted with black color while
for u = 0, 1, 2, ..., 31 and v = 0, 1, 2, ..., 31. furrows are white.
In order to enhance a specific block by its dominant A locally adaptive binarization method is performed to
frequencies, we multiply the FFT of the block by its binarize the fingerprint image. In this method image is
magnitude a set of times. Where the magnitude of the divided into blocks of 16 x 16 pixels. A pixel value is
original FFT = abs (F (u, v)) = |F (u, v)|. then set to 1 if its value is larger than the mean
So we get the enhanced block according to the intensity value of the current block to which the pixel
equation: belongs (Figure 3.4).

  (2)
-1
where F (F (u, v)) is given by:

(3)
For x = 0, 1, 2 …31 and y = 0, 1, 2 ...31.
The k in formula (2) is an experimentally determined
constant, which we choose k=0.45 to calculate. A high
value of k improves the appearance of the ridges by
filling up small holes in ridges, but too high value of k (a) (b)
can result in false joining of ridges which might lead to Figure 3.4(a) Binarized Image after FFT, (b) Image before
a termination become a bifurcation. binarization
Figure 3.3 presents the image after FFT enhancement.

3.2 Fingerprint Image Segmentation

Computer Vision and Image Processing (CS676)Page 8

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)
After image enhancement the next step is fingerprint The direction map is shown in the following diagram
image segmentation. In general, only a Region of (Figure 3.5).
Interest (ROI) is useful to be recognized for each
fingerprint image. The image area without effective
ridges and furrows is first discarded since it only holds
background information. Then the bound of the
remaining effective area is sketched out since the
minutiae in the bound region are confusing with those
spurious minutiae that are generated when the ridges
are out of the sensor.
To extract the region of interest, two steps are
followed: Block direction estimation and ROI extraction (a) (b)
by Morphological methods. Figure 3.4(a) Binarized Image, (b) Direction map of image
3.2.1 Block direction estimation 3.2.2 ROI Extraction by Morphological
Here the fingerprint image is divided into blocks of size operations
16 x 16 pixels (W x W) after which the block direction ROI extraction is done using two Morphological
of each block is calculated according to the algorithm: operations called OPEN and CLOSE. The OPEN
operation can expand images and remove peaks
I. Calculate the gradient values along x-direction (g x)
introduced by background noise (Figure 3.6). The
and y-direction (gy) for each pixel of the block. Two
‘CLOSE’ operation can shrink images and eliminate
Sobel filters are used to fulfill the task.
small cavities (Figure 3.7).
II. For each block, use following formula to get the
Least Square approximation of the block direction.
2 ( g x∗g y )
tan2ß =
( g x 2−g y 2 )
for all the pixels in each block.

The formula is easy to understand by regarding

gradient values along x-direction and y-direction as
cosine value and sine value. So the tangent value of the Figure 3.5 Original image area
block direction is estimated nearly the same as the way
illustrated by the following formula.
2 sin cos
tan2 =
cos 2−sin 2
After finished with the estimation of each block
direction, those blocks without significant information
on ridges and furrows are discarded based on the
following formulas:
2 ( g x∗g y ) + ( g x 2−g y 2 )
E= Figure 3.6 After CLOSE Figure 3.7 After OPEN
W∗W ∗( g x 2+ g y 2 )

For each block, if its certainty level E is below a

threshold, then the block is regarded as a background
block.

Computer Vision and Image Processing (CS676)Page 9

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)
3.3.2 Minutiae Marking
Minutiae marking is now done using templates for
each 3 x 3 pixel window as follows.
If the central pixel is 1 and has exactly 3 one-value
neighbors, then the central pixel is a ridge branch
(Figure 3.10).

Figure 3.8 Final ROI

Figure 3.8 show the final ROI of the fingerprint which is
the bound area after subtraction of the closed area
from the opened area. Then the leftmost, rightmost, Figure 3.10
uppermost and bottommost blocks out of the bound If the central pixel is 1 and has only 1 one-value
area are discarded. neighbor, then the central pixel is a ridge ending
(Figure 3.11).
3.3 Final Minutiae Extraction
Now that we have enhanced the image and segmented
the required area, the job of minutiae extraction closes
down to four operations: Ridge Thinning, Minutiae
Marking, False Minutiae Removal and Minutiae
Representation.
Figure 3.11
3.3.1 Ridge Thinning
There is one case where a general branch may be triple
In this process we eliminate the redundant pixels of
counted (Figure 3.12). Suppose both the uppermost
ridges till the ridges are just one pixel wide. This is
pixel with value 1 and the rightmost pixel with value 1
done using the MATLAB’s built in morphological
have another neighbor outside the 3x3 window due to
thinning function.
some left over spikes, so the two pixels will be marked
bwmorph(binaryImage,’thin’,Inf)
as branches too, but actually only one branch is located
The thinned image is then filtered, again using
in the small region. Thus this is taken care of.
MATLAB’s three morphological functions to remove
some H breaks, isolated points and spikes (Figure 3.9).

bwmorph(binaryImage, ’hbreak’, k)
bwmorph(binaryImage, ’clean', k)
bwmorph(binaryImage, ’spur', k)
Figure 3.12

3.3.3 False Minutiae Removal

At this stage false ridge breaks due to insufficient
amount of ink & ridge cross connections due to over
inking are not totally eliminated. Also some of the
earlier methods introduce some spurious minutia
points in the image. So to keep the recognition system
consistent these false minutiae need to be removed.
(a) (b)
Here we first calculate the inter ridge distance D
Figure 3.9(a) Image before, (b) Image after thinning which is the average distance between two

Computer Vision and Image Processing (CS676)Page 10

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)
neighboring ridges. For this scan each row to calculate So each minutia is completely characterized by the
the inter ridge distance using the formula: following parameters 1) x-coordinate, 2) y-coordinate,

Inter ridge distance =

∑ all pixels with value 1 3) orientation and 4) ridge associated with it (Figure
row length 3.14)
Finally an averaged value over all rows gives D.
All we label all thinned ridges in the fingerprint image
with a unique ID for further operation using a MATLAB
morphological operation BWLABEL.
Now the following 7 types of false minutia points are
removed using these steps (Figure 3.13).
Figure 3.14
Actually a bifurcation can be broken down to three
terminations each having their own x-y coordinates
(pixel adjacent to the bifurcating pixel), orientation and
an associated ridge.
The orientation of each termination (tx, ty) is estimated
by following method. Track a ridge segment whose
starting point is the termination and length is D. Sum
Figure 3.13 up all x-coordinates of points in the ridge segment.
 If d(bifurcation, termination) < D & the 2 minutia Divide above summation with D to get sx. Then get sy
are in the same ridge then remove both of them using the same way.
(case m1) −1 sy−ty
Get the direction from: tan
 If d(bifurcation, bifurcation) < D & the 2 minutia are sx−tx
in the same ridge them remove both of them (case Results after the minutia extraction stage (Figure 3.15-
m2, m3) 3.17)
 If d(termination, termination) ≈ D & the their
directions are coincident with a small angle
variation & no any other termination is located
between the two terminations then remove both of
them (case m4, m5, m6)
 If d(termination, termination) < D & the 2 minutia
are in the same ridge then remove both of them
(case m7)
where d(X, Y) is the distance between 2 minutia
points.

Figure 3.15 Thinned image

3.3.4 Minutiae Representation
Finally after extracting valid minutia points from the
fingerprint they need to be stored in some form of
representation common for both ridge ending and
bifurcation.

Computer Vision and Image Processing (CS676)Page 11

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)

Now we choose one minutia from each set to find the

ridge correlation factor between them. The ridge
associated with each minutia is represented as a series
of x-coordinates (x1, x2…xn) of the points on the ridge. A
point is sampled per ridge length L starting from the
minutia point, where the L is the average inter-ridge
length. And n is set to 10 unless the total ridge length is
less than 10*L.
So the similarity of correlating the two ridges is derived
from:
Figure 3.16 Minutiae after marking
m

S=

√ ∑ xiX i
i=0
m

∑ x i2 X i2
i=0

where (xi..xn) and (Xi..Xn) are the set of x-coordinates

for each of the 2 minutia chosen. And m is minimal one
of the n and N value. If the similarity score is larger
than 0.8, then go to step 2, otherwise continue to
match the next pair of ridges.
2. The approach is to transform each set according to
its own reference minutia and then do match in a
unified x-y coordinate.
Let M ( x , y ,θ ) be reference minutia found from step
1(say from I1). For each fingerprint, translate and rotate
Figure 3.17 Real Minutiae after false removal all other minutiae ( xi , yi ,θi ) with respect to the M
according to the following formula:
xi new cos θ −sin θ 0 xi−x
4 Minutiae Matching
After successfully extracting the set of minutia points
of 2 fingerprint images to be tested, we perform
( )[
yi new = sin θ cos θ 0
θi new 0 0 1 ][ ]
yi− y
θi−θ
The new coordinate system is originated at reference
Minutiae Matching to check whether they belong to
minutia M and the new x-axis is coincident with the
the same person or not.
direction of minutia M. No scaling effect is taken into
We use an iterative ridge alignment algorithm to first
account by assuming two fingerprints from the same
align one set of minutiae w.r.t other set and then carry-
finger have nearly the same size.
out an elastic match algorithm to count the number of
So we get transformed sets of minutiae I1’ & I2’
matched minutia pairs.

4.1 Minutiae Alignment

4.2 Minutiae Match
Let I1 & I2 be the two minutiae sets given by,

Computer Vision and Image Processing (CS676)Page 12

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)
An elastic string ( x , y ,θ ) match algorithm is used to find 5.2 Experiment Analysis
number of matched minutia pairs among I1’ & I2’. A fingerprint database from the FVC2002 (Fingerprint
According to the elastic string match algorithm Verification Competition 2002) is used to test the
minutia mi in I1’ and a minutia mj in I2’ are program’s performance. A series of correct and
incorrect match score is recorded.
considered "matching," if the spatial distance (sd)
Following is the distribution curve obtained after
between them is smaller than a given tolerance r 0
experiments (Figure 4.1).
and the direction difference (dd) between them is
smaller than an angular tolerance Ѳ0.

sd = √ (xi−xj)2 +( yi− yj)2 ≤ r0

dd = min (|θi−θj|, 360−|θi−θj|) ≤ Ѳ0
❑

Let mm(.) be an indicator function that returns 1 in

the case where the minutiae mi and mj match
according to above equations.

1,∧s d ( mi , m j ) ≤r 0∧dd (m i , m j)≤θ 0

mm(mi,mj)=
{ 0 ,∧otherwise
Figure 5.1 Distribution of Correct Scores and Incorrect Scores
(Red: Incorrect Scores, Green: Correct Scores)
Now the total number of matched minutiae pair given
by, In our experiments distribution curve gives an average
num (matched minutiae) = ∑ mm(m i ,m j ) correct match score of about 30 and average incorrect
match score of 25 on the database chosen.
and final match score is given by, The FAR and FRR curve as claimed by the algorithm is
num(matched minutiae) shown under (Figure 5.2)
Match Score =
max ⁡(num of minutiae ¿ I 1 , I 2 )

5 Experimental Results
5.1 Performance Evaluation Index
Two indexes are well accepted to determine the
performance of a fingerprint recognition system:
 False Rejection Rate (FRR): For an image database,
each sample is matched against the remaining
Figure 5.1 FRR and FAR curve (Red: FAR, Blue: FRR)
samples of the same finger to compute the False
Rejection Rate In our experiments FAR and FRR values were 30-35%
 False Acceptance Rate (FAR): Also the first sample approximately. Thus at a threshold match score of
of each finger in the database is matched against about 28 the verification rate of the algorithm is about
the first sample of the remaining fingers to 65-70%.
compute the False Acceptance Rate. The relatively low percentage of verification rate is due
to poor quality of images in the database and the

Computer Vision and Image Processing (CS676)Page 13

 Fingerprint Recognition
Rohit Singh (Y6400), Utkarsh Shah (Y6510), Vinay Gupta (Y6534)
inefficient matching algorithm which lead to incorrect  Fingerprint Classification and Matching by Anil Jain
matches. (Department of Computer Science & Engineering,
Michigan State University) & Sharath Pankanti
(Exploratory Computer Vision Group IBM T. J.
6 Conclusion Watson Research Centre) 2000
The above implementation was an effort to understand  Fingerprint database - FVC2002 (Fingerprint
how Fingerprint Recognition is used as a form of Verification Competition 2002)
biometric to recognize identities of human beings. It  Wikipedia link -
includes all the stages from minutiae extraction from http://en.wikipedia.org/wiki/Fingerprint_recognition
fingerprints to minutiae matching which generates a
match score. Various standard techniques are used in
the intermediate stages of processing.
The relatively low percentage of verification rate as
compared to other forms of biometrics indicates that
the algorithm used is not very robust and is vulnerable
to effects like scaling and elastic deformations. Various
new techniques and algorithm have been found out
which give better results.
Also a major challenge in Fingerprint recognition lies in
the pre processing of the bad quality of fingerprint
images which also add to the low verification rate.

References
 Handbook of Fingerprint Recognition by Davide
Maltoni, Dario Maio, Anil K. Jain & Salil Prabhakar
 Fingerprint Recognition, Paper by WUZHILI
(Department of Computer Science & Engineering,
Hong Kong Baptist University) 2002

Computer Vision and Image Processing (CS676)Page 14