Setisi 2015

Fingerprint Identification Based on Minutiae Point Using Probabilistic Neural NetworkEnny Indasyah, Septian Enggar S., Shi Jihn Horng, Ketut Edi P., Mauridhi Hery Purnomo1. Electrical Engineering Department - Institut Teknologi

Sepuluh Nopember2. Department of Computer Science and Information

Engineering - National Taiwan University of Science and Technology (NTUST)

1

outline

Introduction Contribution Algorithm Result Conclusion and Next Research

2

Introduction

Fingerprint identification problem Fingertip surface

conditionSpecial skin condition affects fingerprints identification perform

Feature points are hard to be extracted by image processing

3

Contribution

To produce better fingerprint identification result on special skin condition

4

AlgorithmImage acquisition

Image binarization

Object Thinning

Minutiae feature extraction

Identification based PNN

5

Image acquisition

Fingerprint image

Obtained from fingerprint sensor

6

Image binarizationRule:Pixel values is set to 255 if the pixel is less than the threshold level

Threshold level = 160

Before binarization

after binarization

7

Thinningto eliminate or reduce the width of ridges till the ridges are one pixel

to simplify the subsequent structural analysis of the image for the extraction of the fingerprint minutiae

Before thinning after thinning

8


By using Crossing Number (CN) concept

1. Whether a pixel, belongs to a ridge or not.

2. if the central pixel is 1 and has exactly 3 one-value neighbors, then the central pixel is a ridge bifurcation.

3. If the central pixel is 1 and has only 1 one-value neighbor, then the central pixel is a ridge ending (termination). 0 1 0

0 1 0

1 0 1

0 1 0

0 1 0

0 0 1

bifurcation termination

9


Removal spurious minutiae1. if the distance between a

termination and a bifurcation is smaller than D.

2. if the distance between two bifurcations is smaller than D.

3. if the distance between two terminations is smaller than D.

What is D?D is average distance between two neighboring ridges

Parameters of minutiae point a) bifurcation and b) ridge ending (termination)

Minutiae Extraction before removal process

Minutiae Extraction after Removal process

Green circle is bifurcation and red circle is termination

10

Identification based PNNInput layer:Minutiae feature

hidden layer:Classifying the minutiae

output layer:Result of identification

11

Result-Minutiae extraction

X Y Angle2 79 2.0095 2 113.002 128 2.00

145 2 177.002 192 4.0062 5 49.007 35 8.00

Parameter of termination

Parameter of bifurcationX Y Angle 1 Angle

2Angle

38 106 20.00 94.00 19.00

141 27 135.00 42.00 123.0050 117 52.00 185.00 57.0088 56 180.00 62.00 191.0063 45 64.00 84.00 66.00

186 68 32.00 75.00 29.0076 49 80.00 120.00 82.00

12

Result-PNN’s performance

0 10 20 30 40 50 60 70 80 90 1001

2

3

4

5

6

7

8

9

10The Result of Data Testing for PNN Classification

targetresult

100 testing data and 10 class fingerprint

the vertical axis is classification results from 10 different class. Horizontal axis is the total of testing data from 10 different class.

13

Conclusion & Next Research

Robust fingerprint preprocessing methods

Better results of identification

Next research

The experimental results show very good performance both on training and testing from database

Conclusion

14

Thank You

15