www.gusucode.com > 指纹识别源代码 > 指纹识别源代码/指纹识别的matlab源码/fprec.m
% Filterbank-Based Fingerprint Matching (A.K.Jain, S.Prabhakar, L.Hong and S.Pankanti, 2000)
%
% Abstract
% With identity fraud in our society reaching unprecedented proportions and with an increasing emphasis on
% the emerging automatic personal identification applications, biometrics-based verification, especially
% fingerprint-based identification, is receiving a lot of attention. There are two major shortcomings
% of the traditional approaches to fingerprint representation. For a considerable fraction of population,
% the representations based on explicit detection of complete ridge structures in the fingerprint are
% difficult to extract automatically. The widely used minutiae-based representation does not utilize a
% significant component of the rich discriminatory information available in the fingerprints. Local ridge
% structures cannot be completely characterized by minutiae. Further, minutiae-based matching has difficulty
% in quickly matching two fingerprint images containing different number of unregistered minutiae points.
% The proposed filter-based algorithm uses a bank of Gabor filters to capture both local and global details
% in a fingerprint as a compact fixed length FingerCode. The fingerprint matching is based on the Euclidean
% distance between the two corresponding FingerCodes and hence is extremely fast. We are able to achieve a
% verification accuracy which is only marginally inferior to the best results of minutiae-based algorithms
% published in the open literature. Our system performs better than a state-of-the-art minutiae-based system
% when the performance requirement of the application system does not demand a very low false acceptance rate.
% Finally, we show that the matching performance can be improved by combining the decisions of the matchers
% based on complementary (minutiae-based and filter-based) fingerprint information.
%
% Index Terms: Biometrics, FingerCode, fingerprints, flow pattern, Gabor filters, matching, texture, verification.
%
% Type "fprec" on Matlab command window to start image processing.
% This source code provides a new, improved GUI respect to the previous release.
% Simulation parameters can be changed in this file:
%
% n_bands: the number of concentric bands
% h_bands: the width of each band in pixels
% n_arcs: the number of arcs (each band has exactly n_arcs arcs)
% h_radius: the inner radius in pixel (the central band is not considered
% because it is a too small area)
% num_disk: the number of gabor filters
%
% n_sectors and h_lato rapresents respectively the total number of sectors
% (the length of the feature vector associated to each filter of filter-bank:
% there are num_disk gabor filters) and the height of the cropped image in pixels.
% N_secors and h_lato should not be changed.
%
%
% $Revision: 1.0 $ $Date: 2002.10.02 $
% $Revision: 2.0 $ $Date: 2003.11.29 $
% $Revision: 3.0 $ $Date: 2004.06.22 $ by Luigi Rosa
% email: luigi.rosa@tiscali.it
% mobile: +393403463208
% website: http://utenti.lycos.it/matlab
%
%
% Modified respect to the previous version:
%
% - Major bugs fixed
% - New GUI
% - 8 Gabor filters 0 22.5 45 67.5 90 112.5 135 157.5 degree
% - Convolution is performed in frequency domain
% - DataBase
% - Fingerprint matching
% - Error management
% - Complex filtering techniques
% - Improved core point determination
% - Robustness against noise
% - Modifiable simulation parameters
%
% Input fingerprint should be 256 x 256 image 8-bit grayscale @ 500 dpi.
% If these conditions are not verified some parameters in m-functions
% should be changed in a proper way (such as, for example, Gabor filter
% parameters in gabor2d_sub function). See the cited references for more
% details.
%
% M-files included:
%
% -fprec.m: this file. It initializes the entire image processing. The simulation
% parameters can be changed in this main file.
% -centralizing.m: a function which accept an input image and determines the coordinates
% of the core point. The core point is determinated by complex filtering.
% The region of interest is determinated fixing a minimum threshold value
% for the variance. Input image is divided into non-overlapping blocks and
% only blocks with a variance smaller than this threshold value are considered
% background. The logical matrix (associated to the region of interest) is first
% closed (Matlab function imclose), then eroded (Matlab function imerode) with two
% given structuring elements. The image is "mirrored" before convolution with complex
% filter, then it is re-cropped to its original sizes.
% -mirror.m: a function which is used to "mirror" input image in order to avoid undesired
% boundary effects (function used by centralizing.m).
% -recrop.m: a function used to resize the mirrored filtered image (function used by centralizing.m)
% -conv2fft.m: this function performs 2D FFT-based convolution. Type "help conv2fft" on Matlab command
% window for more details.
% -whichsector.m: a function used to determine (for each pixel of the cropped image) the corresponding
% sectors of the concentric bands (function used by sector_norm.m).
% -sector_norm.m: a function used to normalize input image and to calculate the features vector
% -cropping.m: this function is used to cropp the input fingerprint image after the core point is
% determinated.
% -gabor2d_sub.m: a function used to calculate the coefficients of the gabor 2D filters.
% -vedicentro.m: this simple routines uses the M-function centralizing.m and it is used to display
% the core point.
%
%
% A crucial step in fingerprint recognition is core point determination.
% If any error occurs while cropping image you can use the auxiliary m-file
% "vedicentro.m": it visualizes the input fingerprint and the core point
% calculated by the m-function "centralizing.m".
%
%
% Notes:
% The computational load can be significantly reduced by recursive filtering techniques.
% For a complete publication list of Lucas J. van Vliet please visit the following URL:
% http://www.ph.tn.tudelft.nl/~lucas/publications/papersLJvV.html
% Here you will find articles concernings a recursive implementation of the Gaussian filter,
% of the derivative Gaussian filter and of Gabor filter.
%
% If you want to optimize the proposed method an excellent article is the following one:
% Erian Bezhani, Dequn Sun, Jean-Luc Nagel, and Sergio Carrato, "Optimized filterbank
% fingerprint recognition", Proc. SPIE Intern. Symp. Electronic Imaging 2003, 20-24
% Jan. 2003, Santa Clara, California.
%
% This code was developed using:
% MATLAB Version 6.5 and Image Processing Toolbox Version 3.2 (R13)
% Operating System: Microsoft Windows 2000 Version 5.0 (Build 2195: Service Pack 4)
% Java VM Version: Java 1.3.1_01 with Sun Microsystems Inc. Java HotSpot(TM) Client VM
%
% Please contribute if you find this software useful.
% Report bugs to luigi.rosa@tiscali.it
%
%
% References:
%
% Cheng Long Adam Wang, researcher
% Fingerprint Recognition System
% http://home.kimo.com.tw/carouse9/FRS.htm
%
% A. K. Jain, S. Prabhakar, and S. Pankanti, "A Filterbank-based Representation for
% Classification and Matching of Fingerprints", International Joint Conference on
% Neural Networks (IJCNN), pp. 3284-3285, Washington DC, July 10-16, 1999.
% http://www.cse.msu.edu/~prabhaka/publications.html
%
% "Fingerprint Classification and Matching Using a Filterbank", Salil Prabhakar
% A DISSERTATION Submitted to Michigan State University in partial fulfillment
% of the requirements for the degree of DOCTOR OF PHILOSOPHY, Computer
% Science & Engineering, 2001
% http://biometrics.cse.msu.edu/SalilThesis.pdf
%
% Final Report 18-551 (Spring 1999) Fingerprint Recognition Group Number 19
% Markus Adhiwiyogo, Samuel Chong, Joseph Huang, Weechoon Teo
% http://www.ece.cmu.edu/~ee551/Old_projects/projects/s99_19/finalreport.html
%
% Kenneth Nilsson and Josef Bigun, "Localization of corresponding points in
% fingerprints by complex filtering", Pattern Recognition Letters, 24 (2003) 2135-2144
% School of Information Science, Computer and Electrical Engineering (IDE), Halmstad
% University, P.O. Box 823, SE-301 18, Halmstad, Sweden.
%
%
% ************************************************************************
% This code required a lot of time to be developed. Please send the author
% money, food, drinks or improvements to the code itself.
% This is my postal address:
% Luigi Rosa
% Via Centrale 35
% 67042 Civita di Bagno
% L'Aquila --- ITALY
%
% mobile +39 340 3463208
% email luigi.rosa@tiscali.it
% website http://utenti.lycos.it/matlab
% *************************************************************************
%
%
%
clear;
clc;
close all;
global immagine n_bands h_bands n_arcs h_radius h_lato n_sectors matrice num_disk
n_bands=4;
h_bands=20;
n_arcs=16;
h_radius=12;
h_lato=h_radius+(n_bands*h_bands*2)+16;
if mod(h_lato,2)==0
h_lato=h_lato-1;
end
n_sectors=n_bands*n_arcs;
matrice=zeros(h_lato);
for ii=1:(h_lato*h_lato)
matrice(ii)=whichsector(ii);
end
num_disk=8;
% 1--> add database
% 0--> recognition
%ok=0;
chos=0;
possibility=7;
messaggio='Insert the number of set: each set determins a class. This set should include a number of images for each person, with some variations in expression and in the lighting.';
while chos~=possibility,
chos=menu('Fingerprint Recognition System','Select image and add to database','Select image for fingerprint recognition','Info','Delete database',...
'Fingerprint image: visualization','Gabor Filter: visualization','Exit');
%--------------------------------------------------------------------------
%--------------------------------------------------------------------------
%--------------------------------------------------------------------------
% Calculate FingerCode and Add to Database
if chos==1
clc;
close all;
selezionato=0;
while selezionato==0
[namefile,pathname]=uigetfile({'*.bmp;*.tif;*.tiff;*.jpg;*.jpeg;*.gif','IMAGE Files (*.bmp,*.tif,*.tiff,*.jpg,*.jpeg,*.gif)'},'Chose GrayScale Image');
if namefile~=0
[img,map]=imread(strcat(pathname,namefile));
selezionato=1;
else
disp('Select a grayscale image');
end
if (any(namefile~=0) && (~isgray(img)))
disp('Select a grayscale image');
selezionato=0;
end
end
immagine=double(img);
if isa(img,'uint8')
graylevmax=2^8-1;
end
if isa(img,'uint16')
graylevmax=2^16-1;
end
if isa(img,'uint32')
graylevmax=2^32-1;
end
fingerprint = immagine;
N=h_lato;
[BinarizedPrint,XofCenter,YofCenter]=centralizing(fingerprint,0);
[CroppedPrint]=cropping(XofCenter,YofCenter,fingerprint);
[NormalizedPrint,vector]=sector_norm(CroppedPrint,0);
for (angle=0:1:num_disk-1)
gabor=gabor2d_sub(angle,num_disk);
ComponentPrint=conv2fft(NormalizedPrint,gabor,'same');
[disk,vector]=sector_norm(ComponentPrint,1);
finger_code1{angle+1}=vector(1:n_sectors);
end
img=imrotate(img,180/(num_disk*2));
fingerprint=double(img);
[BinarizedPrint,XofCenter,YofCenter]=centralizing(fingerprint,0);
[CroppedPrint]=cropping(XofCenter,YofCenter,fingerprint);
[NormalizedPrint,vector]=sector_norm(CroppedPrint,0);
for (angle=0:1:num_disk-1)
gabor=gabor2d_sub(angle,num_disk);
ComponentPrint=conv2fft(NormalizedPrint,gabor,'same');
[disk,vector]=sector_norm(ComponentPrint,1);
finger_code2{angle+1}=vector(1:n_sectors);
end
% FingerCode added to database
if (exist('fp_database.dat')==2)
load('fp_database.dat','-mat');
fp_number=fp_number+1;
data{fp_number,1}=finger_code1;
data{fp_number,2}=finger_code2;
save('fp_database.dat','data','fp_number','-append');
else
fp_number=1;
data{fp_number,1}=finger_code1;
data{fp_number,2}=finger_code2;
save('fp_database.dat','data','fp_number');
end
message=strcat('FingerCode was succesfully added to database. Fingerprint no. ',num2str(fp_number));
msgbox(message,'FingerCode DataBase','help');
end
%--------------------------------------------------------------------------
%--------------------------------------------------------------------------
%--------------------------------------------------------------------------
% Fingerprint recognition
if chos==2
clc;
close all;
selezionato=0;
while selezionato==0
[namefile,pathname]=uigetfile({'*.bmp;*.tif;*.tiff;*.jpg;*.jpeg;*.gif','IMAGE Files (*.bmp,*.tif,*.tiff,*.jpg,*.jpeg,*.gif)'},'Chose GrayScale Image');
if namefile~=0
[img,map]=imread(strcat(pathname,namefile));
selezionato=1;
else
disp('Select a grayscale image');
end
if (any(namefile~=0) && (~isgray(img)))
disp('Select a grayscale image');
selezionato=0;
end
end
immagine=double(img);
if isa(img,'uint8')
graylevmax=2^8-1;
end
if isa(img,'uint16')
graylevmax=2^16-1;
end
if isa(img,'uint32')
graylevmax=2^32-1;
end
fingerprint = immagine;
N=h_lato;
[BinarizedPrint,XofCenter,YofCenter]=centralizing(fingerprint,0);
[CroppedPrint]=cropping(XofCenter,YofCenter,fingerprint);
[NormalizedPrint,vector]=sector_norm(CroppedPrint,0);
% memoria per feature vector d'ingresso
vettore_in=zeros(num_disk*n_sectors,1);
for (angle=0:1:num_disk-1)
gabor=gabor2d_sub(angle,num_disk);
ComponentPrint=conv2fft(NormalizedPrint,gabor,'same');
[disk,vector]=sector_norm(ComponentPrint,1);
finger_code{angle+1}=vector(1:n_sectors);
vettore_in(angle*n_sectors+1:(angle+1)*n_sectors)=finger_code{angle+1};
end
% FingerCode of input fingerprint has just been calculated.
% Checking with DataBase
if (exist('fp_database.dat')==2)
load('fp_database.dat','-mat');
%---- alloco memoria -----------------------------------
%...
vettore_a=zeros(num_disk*n_sectors,1);
vettore_b=zeros(num_disk*n_sectors,1);
best_matching=zeros(fp_number,1);
valori_rotazione=zeros(n_arcs,1);
% start checking ---------------------------------------
for scanning=1:fp_number
fcode1=data{scanning,1};
fcode2=data{scanning,2};
for rotazione=0:(n_arcs-1)
p1=fcode1;
p2=fcode2;
% ruoto i valori dentro disco
for conta_disco=1:num_disk
disco1=p1{conta_disco};
disco2=p2{conta_disco};
for old_pos=1:n_arcs
new_pos=mod(old_pos+rotazione,n_arcs);
if new_pos==0
new_pos=n_arcs;
end
for conta_bande=0:1:(n_bands-1)
disco1r(new_pos+conta_bande*n_arcs)=disco1(old_pos+conta_bande*n_arcs);
disco2r(new_pos+conta_bande*n_arcs)=disco2(old_pos+conta_bande*n_arcs);
end
end
p1{conta_disco}=disco1r;
p2{conta_disco}=disco2r;
end
% ruoto i dischi circolarmente
for old_disk=1:num_disk
new_disk=mod(old_disk+rotazione,num_disk);
if new_disk==0
new_disk=num_disk;
end
pos=old_disk-1;
vettore_a(pos*n_sectors+1:(pos+1)*n_sectors)=p1{new_disk};
vettore_b(pos*n_sectors+1:(pos+1)*n_sectors)=p2{new_disk};
end
d1=norm(vettore_a-vettore_in);
d2=norm(vettore_b-vettore_in);
if d1<d2
val_minimo=d1;
else
val_minimo=d2;
end
valori_rotazione(rotazione+1)=val_minimo;
end
[minimo,posizione_minimo]=min(valori_rotazione);
best_matching(scanning)=minimo;
end
[distanza_minima,posizione_minimo]=min(best_matching);
beep;
message=strcat('The nearest fingerprint present in DataBase which matchs input fingerprint is : ',num2str(posizione_minimo),...
' with a distance of : ',num2str(distanza_minima));
msgbox(message,'DataBase Info','help');
else
message='DataBase is empty. No check is possible.';
msgbox(message,'FingerCode DataBase Error','warn');
end
end % fine caso 2
if chos==3
clc;
close all;
helpwin fprec;
end % fine caso 3
if chos==4
clc;
close all;
if (exist('fp_database.dat')==2)
button = questdlg('Do you really want to remove the Database?');
if strcmp(button,'Yes')
delete('fp_database.dat');
msgbox('Database was succesfully removed from the current directory.','Database removed','help');
end
else
warndlg('Database is empty.',' Warning ')
end
end % fine caso 4
if chos==5
clc;
close all;
selezionato=0;
while selezionato==0
[namefile,pathname]=uigetfile({'*.bmp;*.tif;*.tiff;*.jpg;*.jpeg;*.gif','IMAGE Files (*.bmp,*.tif,*.tiff,*.jpg,*.jpeg,*.gif)'},'Chose GrayScale Image');
if namefile~=0
[img,map]=imread(strcat(pathname,namefile));
selezionato=1;
else
disp('Select a grayscale image');
end
if (any(namefile~=0) && (~isgray(img)))
disp('Select a grayscale image');
selezionato=0;
end
end
figure('Name','Selected image');
imshow(img);
end % fine caso 5
if chos==6
clc;
close all;
figure('Name','Gabor Filter');
mesh(gabor2d_sub(0,num_disk));
end % fine caso 6
end % fine while