Fingerprint matching from minutiae texture maps -com
Download as PPTX, PDF
1 like927 views
This document describes improvements made to the Fingercode fingerprint matching algorithm. The original Fingercode approach extracts a 640-byte feature vector from a fingerprint image using minutiae points, tessellation into sections, and Gabor filtering. The first improvement uses minutiae as reference points and local orientation. The second improvement adds section weighting and allows for orientation and localization variations. Experimental results on two fingerprint databases show the improved approach achieves error rates below 6%, outperforming other top algorithms. The improvements make the approach more robust to noise and errors while increasing computational complexity.
![Experimental Results
Matching performance
Location errors: [-7,+7]
Orientation variations:
FVC2002](https://image.slidesharecdn.com/fingerprintmatchingfromminutiaetexturemaps-com-131105040401-phpapp02/85/Fingerprint-matching-from-minutiae-texture-maps-com-22-320.jpg)
Fingerprint matching from minutiae texture maps -com
- 1. Fingerprint matching from minutiae texture maps
- 2. Contents The Original Approach--Fingercode Minutiae, Tessellation, Gabor Filter, Fingercode First Improvement Minutiae as reference points, local orientation Second Improvement Section weight, variations Performance and Evaluation
- 3. Background Fingerprint matching Widely used Give an yes-or-no answer fast and convenient Shortage for previous approaches Be of variable length Minutiae Ridge Endings Bifurcations
- 4. Fingercode a 640-byte feature vector The matching process calculate the Euclidean distance between the sensored fingercode and the template fingercode
- 5. Fingercode1 Reference Frame Determination reference point: Maximum curvature reference axis: local symmetry of ref point
- 6. Fingercode2 Tessellation A plate of 120 pixels is selected and divided into 80 section Nomalization
- 7. Fingercode3 3. Gabor Filter 8 directions Gabor filters
- 9. Fingercode4 Feature Vector Extraction standard deviation in 8*80 sections 640 bytes Finger 1 Finger 2
- 10. Why does it work... Each Garbor filter contains information about ridges and furrows both globally and locally. Band width is 20 pixels. Inter-ridge width is 10 pixels. So large variation means ridges and furrows along the direction.
- 11. Why does it work... Detected ref point can be 12 pixels away. Orientation can be 20 away Since fingerprint is "smooth", and we use statistical (rather than is-or-no) data.
- 12. First Adjustment Minutiae as reference points Fingercode at every minutiae Ref axis along the minutiae direction 1.Avoid pre-alignment risks 2.More robust
- 13. Second Adjustmdent Weighting: ADD
- 14. Problem? Wrong detection of ref point can make things disastrous Errors Location Errors Orientation Errors
- 15. Error Correction Orientation variation Localization Variation 7 Pixels
- 16. 1 2 3 4
- 17. Matching Input fingerprints Minutiae list Template fingerprints Minutiae list Compute the Distances for each pair 8 Directions 80 sectors Find the minimum one
- 18. Test Data Base Exploit databases from Fingerprint Verifictation Competition FVC2000 FVC2002 100 distinct fingers for each base 8 impressions for each finger
- 19. Experimental ResultsEER-curve DB1 FVC2000 d varied the minutiae orientations Compute FRR
- 20. Experimental ResultsEERs DB1 FVC2000 EERs estimated for different orientation variations ERR < 6%
- 22. Experimental Results Matching performance Location errors: [-7,+7] Orientation variations: FVC2002
- 23. Experimental Results Rank according to EER FVC2002 compared with other TOP 31 participants
- 24. Experimental Results Ref Points selection compared to fingerCode DB1-a Reasons Noise The ref point is close to border Scars near the ref points(DB3-a) DB2-a DB3-a DB4-a
- 25. Cons and Pros Pros Avoid pre-alignment risk Able to get correct result from poor picture. Cons Computaional expensiveness Assume 30 minutiae every picture, we have to compute 30*30=900 pairs.
- 26. END Thank you!