The document discusses a fingerprint matching method that uses ridge count and associative tables to directly match fingerprint skeleton images. It details preprocessing steps, similarity measures for ridge curves and patterns, and evaluates the performance of the algorithm, highlighting its efficiency in handling elastic distortions compared to other methods. The algorithm shows an average matching time of 0.025 to 0.33 seconds per match, making it a robust solution for fingerprint verification.

1. PRESENTED BY: AAKANKSHA JAIN
MTECH SCHOLAR
MDS UNIVERISTY

2. • Introduction
• Fingerprint Image
• Matching the ridges
• Evaluation of results
• Advantages
• References
CONTENT

3. What is Fingerprint Matching using Ridges?
Fingerprint matching using ridge count is a method with
which two fingerprint skeleton images are matched
directly using association tables.
Association Table: Association table describes relation
between ridges in neighborhood ,to handle whole pattern
of ridge.
• For matching purpose there are 2 similarity measures
(1) Matching ridge curves
(2) Matching ridge patterns
Introduction

4. For obtaining fingerprint skeleton image , we have to go through
several preprocessing steps like:
• Segmentation
• Filtering
• Thinning
• Binarization
Fingerprint Image

5. The neighborhood relationships among ridges are invariant during one’s life
time and are robust to elastic distortions of fingerprint images.These steady
relationships make the base of the ridge-based fingerprint matching
method.
As shown in figure, ridges R1 and R3 are neighbor ridges of ridge R2. A ridge
curve may have more than one neighbor on its each side in the skeleton
image
Matching Ridges

6. A. Similarity measure of two ridge curves :
Suppose Pm and Pn are respectively the starting point and the ending point of
ridge f , and Pm and Pn could be ridge end, ridge bifurcation or ridge broken
points. The curvature γ of curve f is defined as:
γ describes a curve’s winding degree, and it’s an invariant to image rotation and
translation
B. Associate table :
The associate table is constructed by sampling ridge R with interval d from its
starting point to end point, and obtain one sampling point-set £ and its associate
point-sets “up and “down. All the points in “up and “down are labeled by their
corresponding ridges (NULL for empty).
The labels and the sampling point set G make up of ridge R’s associate table.
Continue..

7. Continue..
we choose the sampling interval of 7 pixels in order to depict the neighborhood
relationships of the ridges.The associate tables of all ridges contain all information and
features the image has.
C. Similarity measure of two ridge patterns :
The similarity measure of two fingerprints is defined as:
score = N=(C *distortion)
Where N is the total length of all matched ridges, more ridges matched would achieve
higher score; C is a scaling constant

8. The distortion describes the distortion between the ridge structures formed
by matched ridge pairs, wrong matched ridge pair always leads to higher
distortion value and lower score.
Continue..

9. Figure: Ridge based fingerprint matching results
Evaluation of results..
• Fingerprint Matching algorithm using ridge count ,when applied on the database
of FVC2002, givesThe average matching time is 0.025s to 0.33s per match.
• Algorithm handles the elastic distortion problem well and it also helps to
eliminate the matching uncertainty (such as caused by not having enough
minutiae)since it fully utilizes the ridge information.

10.  Algorithm matches fingerprint skeleton images directly .
 Concept of Associate tables describe the neighborhood relations
among ridge curves.
 Similarity measures Matching ridge curves and Matching ridge
patterns properly handle the elastic distortion.
 Better performance is achieved by this method compared to minutiae-
based matching method.
Advantage

11. 1. A.K.Jain, L.Hong, and R.M.Bolle. On-line Fingerprint Verification. IEEE
Trans. on Pattern Analysis and Machine Intelligence. 19 (4): 302-313, April
1997.
2. Xiaohui Xie, Fei Su, Anni Cai and Jing’ao Sun, ”A Robust Fingerprint
Matching Algorithm Based on the Support Model”. Proc. International
Conference on Biometric Authentication (ICBA), Hong Kong, China, July
15-17, 2004 .
3. D.Maio, D.Maltoni, R.Cappelli, J.L.Wayman, A.K.Jain. FVC2002: Second
Fingerprint Verification Competition. Pattern Recognition, 2002,
Proceedings.
16th International Conference on., 11-15 Aug. 2002 Page(s): 811-814 vol.3
4. Chen PH and Chen XG, A New Approach to Healing the Broken Lines in
the Thinned Fingerprint Image. Journal of China Institute of
Communications.
25(6):115-119, June 2004
References