1. Digital Image Processing - (monsoon 2003) FINAL PROJECT REPORT
FINGERPRINT MATCHING
Project Members
Sanyam Sharma
Sunil Mohan Ranta
Group No. - 15

2. Aim of the Project
To match a Fingerprint image with a one already stored in the database.
A fingerprint image essentially consists of a set of minutiae on the plane.
Minutiae are the terminations and bifurcations of ridge lines in a fingerprint image.
A new approach towards fingerprint recognition is to match the distribution and orientation of such points.

3. Motivation behind it……
Finger-print recognition is used in various systems for Verification, Identification etc.
Recognizing manually can be very time consuming and costly.
There are systems already in use which use similar technology and a lot of research is going on to improve the technique.

4. Algorithm
This particular method of fingerprint matching consists mainly of six stages ….
Image Enhancement,
Ridge extraction
Binarization
Thinning
Minutiae extraction
Post processing.

5. Ridge Detection
As alluded earlier, the objective of the ridge detection algorithm is to separate ridges from the valleys in a given fingerprint image.
A more reliable property of the ridges in a fingerprint image is that the gray level values on ridges attain their local minima along a direction normal to the local ridge orientation.

6. Image Enhancement and Binarization
Removing noise and sharpening the ridges using various filters. eg. Gabor Filter
Making a binary image from the enhanced image.
Ridges in black color on a white background.

7. Thinning
The objectives of this step is to obtain a thinned image using morphological filters on binary images.
All the ridges are only 1- pixel thick.

8. Minutiae Detection
Once the thinned ridge map is available, the ridge pixels with three ridge pixel neighbors are identified as Ridge bifurcations and those with one ridge pixel neighbor are identified as Ridge endings.

9. Building a minutiae skeleton
Set of distances between ridge bifurcating and ridge ending minutiaes.
Distribution of minutiaes.
Orientation of minutiaes.

10. Matching the details …
Comparing the obtained skeleton and minutiae score with the other image.
There can be many ways to match the details obtained.
One approach can be using a skeleton structure of minutiae points.

11. Overall Process Image Enhancement and Ridge Detection Sensor
Binarization
Thinning
Fingerprint Database
Matching
Minutiae Extraction
Result

12. This is all what we proposed
how we do it …..

13. Enhancement
We have achieved Appreciable enhancement using Gabor filters.
Gabor filters have both frequency selective and orientation selective properties, so it enhance the image and preserves the true ridge/valley structures.
Binarization of image using threshold values.

14. Thinning Reducing width of ridges to a ‘single’ pixel.
Morphological thinning.
Doing erosion till no further change occurs in the image.
Removing pixels not satisfying m-connectivity to eliminate multiple paths.

15. Minutiae Detection Next step is to detect Minutiae in the image.
Mask of 3x3 is made to check the neighbours of a pixel.
Then another mask having binary values is used to check the direction of the ridge at that point.
All the minutiae’s have 2 properties :-
Type of minutiae ( termination or bifurcation )
Orientation
We have achieved quite efficient results in detecting all the minutiae points.
Removal of False minutiae points.

16. Matching of minutiae sets
Algorithms Used
Relative Distance Matching :– Finding the nearest similar minutiae and giving a score depending upon the distance.
Using Quad Tree :– exploiting distribution property.
Image Mapping :– sliding one thinned image upon another and finding the max. score obtained.
Each algorithm having a different threshold score for matching.
Results
Matching Criterion:–
( match score > threshold score ) - Appreciable match
( match score < threshold score ) - Non - match

17. Results

18. Matched Images
Match Score = 145 Threshold = 130 (Accepted)

19. Non Match
Match Score = 110 Threshold = 130 (Rejected)

20. Constraints Rotation Variant. Quality of images should be good.
Difficulties …
High efficiency needed as the fields of application are related to security.

21. Future Work We can improve the results by -
Enhancing the image by completing ridges.
Considering average ridge thickness.
A more robust algorithm for minutiae matching.

22. Applications … Fingerprint Matching Fingerprint Verification
Identifiers, Personal Identification Systems, Tracing criminals.
Fingerprint Verification
Secure access, digital signatures etc.

23. Workbed References … Platform – Windows
Tools – Microsoft Visual c++ , Matlab and Matlab addin for MS VC++.
Image Input - Scanner
References …
[1] A. K. Jain, L. Hong, S. Pankanti, R. Bolle, “An identity authentication system using fingerprints”, Proceedings of the IEEE, 85(9)(1997)
[2] A. K. Jain, A. Ross, S. Prabhakar, “Fingerprint matching using Minutiae and Texture Features”.
[3] P. Bhowmick, A. Bishnu, B. B. Bhattacharya , M. K. Kundu, C. A. Murthy, T. Acharya, “Determination of Minutiae Scores for Fingerprint Image Applications”.
[4] Dario Maio and Davide Maltoni “Direct Gray-Scale Minutiae Detection In Fingerprints”.

24. Thx