Related Experiment Videos
Fingerprint warping using ridge curve correspondences.
Arun Ross1, Sarat C Dass, Anil K Jain
1Lane Department of Computer Science and Electrical Engineering, West Virginia University, PO Box 6109, Morgantown, WV 26506, USA. arun.ross@mail.wvu.edu
IEEE Transactions on Pattern Analysis and Machine Intelligence
|January 13, 2006
Summary
This study introduces a new method to correct nonlinear deformation in fingerprint matching using ridge curves. This technique improves fingerprint recognition accuracy by better aligning distorted fingerprint images.
Area of Science:
- Biometrics
- Computer Vision
- Pattern Recognition
Background:
- Fingerprint matching systems are hindered by nonlinear deformation during image acquisition.
- This deformation distorts critical fingerprint features like minutiae points and ridge curves.
Purpose of the Study:
- To develop a technique for estimating and correcting nonlinear deformation in fingerprint images.
- To improve the accuracy of fingerprint matching by addressing feature distortion.
Main Methods:
- A novel method estimates nonlinear distortion using ridge curve correspondences.
- The thin-plate spline (TPS) function models the deformation.
- An average deformation model is created from multiple impressions of the same finger.
- An index is proposed to select the optimal deformation model.
Main Results:
- The proposed ridge curve-based deformation model achieves better image alignment than minutiae-based models.
- Experimental results show improved matching performance with the incorporation of the deformation model.
- The study utilized 1,600 fingerprints from 50 individuals over two weeks.
Conclusions:
- The developed technique effectively estimates and corrects nonlinear fingerprint deformation.
- Utilizing ridge curve correspondences for deformation modeling enhances fingerprint matching accuracy.
- This approach offers a significant improvement for biometric security systems.