MINUTIAE EXTRACTION ALGORITHM

Download Report

Transcript MINUTIAE EXTRACTION ALGORITHM

SECURED
AUTHENTICATION
USING
FINGER PRINT
BIOMETRY
-- IT Acumens.
COM
SYSTEM -DESIGN
USER
CHECK
AUTHENTICATION
AUTHENTICATED
REJECTED
STANDALONE
PC
MODIFIED SYSTEM
DESIGN
VC++
APPLICATION
FINGERPRINT
DATABASE
(BMP FILES)
MODULES
Our project has been modularized in
to three phases
1.Image Diagnostics Binarization
2. Minutiae Extraction
3. Enrollment & Authentication
BLOCK DIAGRAM
BMP
FILE
256
GRAY
SCALE
DIAGNOSTICS
AND
BINARISATION
BINARISED
MINUTIAE
EXTRACTION
MINUTIAE
TEMPLATE
DATA
ARRAY
TEMPLATE
DATABASE
AUTHENTICATION
& ENROLLMENT
DIAGNOSTICS AND
BINARISATION
This is the first module in
the project.
The input to this module
will be a
256 gray scale BMP file.
The output of this module
will be a binarised data of the
image.
BITMAP FILE TYPES
types:
Bitmap files are of two
DDB-Device Dependent
Bitmap
DIB-Device Independent
Bitmap
BITMAPS
DDB
DIB
DIB - CHARACTERS
Bitmaps that contain a
color table are deviceindependent
A color table describes
how pixel values
correspond to RGB
color values.
A DIB contains color
and dimension
information:
The color and
dimension information
is stored in a
BITMAPINFO structure.
BMP STRUCTURE
FILEHEADER
FILEINFORMATION
HEADER
RGBQUAD
DATA STORED
IN
INVERTED
FASHION
HANDLING DIB IN
VC++
The following structures are
maintained in any DIB file
 BITMAPHEADER
 BITMAPINFO
1) BITMAPINFOHEADER
2) RGBQUAD

DATA ARRAY
BITMAPFILEHEADER
The BITMAPFILEHEADER
structure contains the
following members,
1. bfType
2. bfSize
3. bfReserved1
4. bfReserved2
5. bfOffBits
BITMAPINFOHEADER
1. BISIZE
2. BIWIDTH
3. BIHEIGHT
4. BIBITCOUNT
5. biPlanes
6. biCompression
7. biSizeImage
8. biXPelsPerMeter
9. biYPelsPerMeter
10 .biClrUsed
11. biClrImportant
DATA EXTRACTION
START
A
OPEN FILE
C
F
MOVE TO DATA
B
F
A
PADDING
B
READ A PIXEL
T
CLOSE FILE
T
COMPUTE NO OF PADS
END OF
DATA
T
PAD BIT
C
STOP
F
SET PAD FLAG
A
STORE READ PIXEL
IN ARRAY
C
BINARISATION
START
C
READ PIXEL
T
SET PIXEL AS 0
VAL > 127
F
SET PIXEL AS 1
D
F
D
END OF PIXEL
T
STOP
C
MODULE-TWO
Extracting the minutiae
points from the fingerprint image
is the second module
The input will be the binarised
data pixels obtained from the
first module.
The output of this module will
be the image template containing
the minutiae details.
Minutiae Points
valley
bifurcation
ridge
bifurcation
Minutiae Extraction
• Input:
binarised image as a
256*256 array from the
previous module
•
Output:
set of minutiae points
with ( x , y ) and their
relative displacement
Algorithm
Step1:
Consider the inner 236*236 array
Step 2:
Scan the image from top to
bottom ,left to right order by
following only ridges
Step 3:
Find the 0-1 transition ,
calculate the width of the
ridge by noting the 1-0 transition
Step 4:
Move to the next row and follow
the same ridge
Note the width
11100000111111000000000111100000111
11000001111111000000001111010000111
10000011111100000100001111000001111
10000011111000001110001111000001111
00000011110000111110011111000001111
00001111110001111000011110000001111
00011111000001111000011110000111111
00011111000111110000011100000111111
00011111000111110000011100000111111
00111000001111100001111000001111110
00111000001111100001111000001111110
11111100001111000001110000001111000
• Step 5:
If the width >=width in previous
row there may be a top to bottom
bifurcation .Call the bifurcation
function to check if it is a minutiae
point
else
If the width =< width in previous
row there may be a bottom to top
ridge bifurcation.Call the bifurcation
function to check if it is a minutiae
point
• Step 6:
continue with the next row and
repeat this for all the ridges in the
given image or until 90 minutiae
points have been obtained.
Template Storage
• If a point is determined as a
minutiae point then the straight
line distance from the top and
left is determined and is copied
into the displacement array
• count is incremented
• Finally after getting a maximum
of 90 minutiae points the
template is stored.
Application
• New user entry:
This can be done by
only an administrator.The
template is extracted
using the above algorithm
and it is stored into the
template database stored
as a file
• Authentication:
Template is extracted
and checked if any 43
minutiae points are same
if so he is authenticated