Direct Feature Point Correspondence Discovery for Multiview Images: An Alternative Solution When SIFT-Based Matching Fails

  • Jinwei XuEmail author
  • Jiankun Hu
Conference paper
Part of the Lecture Notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering book series (LNICST, volume 177)


3D fingerprint identification is an emerging biometric authentication method, which is powerful against spoofing attacks. For actualizing 3D fingerprint identification, this paper develops a 3D fingerprint minutiae cloud reconstruction technique based on 2D multiview touchless fingerprint images. This technique provides a practical solution for 2D minutiae matching for multiview fingerprint images, when traditional feature correspondence finding based on 2D SIFT (Scale Invariant Feature Transformation) feature points fails. In this case, developing a new 2D feature point correspondence establishment algorithm is necessary to cover the deficiency of the SIFT-based technique. In this paper, minutiae, a type of detailed ridge-valley features in fingerprint images, are utilized for the correspondence discovery. Furthermore, differential evolution, an efficient evolutionary computing framework is employed to directly infer the possible correspondence of minutiae sets from the different posed fingerprint images. Our experiments demonstrate that the proposed direct 2D feature point correspondence discovery strategy is able to handle the cases when the SIFT-based matching fails. To further illustrate the advantages of the proposed algorithm, 3D fingerprint minutiae cloud construction is conducted based on the feature correspondence discovered by the proposed algorithm. The experiments on 2D different posed fingerprint image matching and 3D fingerprint minutiae cloud construction show that the proposed algorithm can be used as an alternation when SIFT-based matching fails.


3D fingerprint reconstruction Feature matching Minutiae cloud Differential evolution 


  1. 1.
    Liu, F., Zhang, D.: 3D fingerprint reconstruction system using feature correspondences and prior estimated finger model. Pattern Recogn. 47, 178–193 (2014)CrossRefGoogle Scholar
  2. 2.
    Choi, H., Choi, K., Kim, J.: Mosaicing touchless and mirror-reflected fingerprint images. IEEE Trans. Inf. Forensics Secur. 5, 52–61 (2010)CrossRefGoogle Scholar
  3. 3.
    Liu, F., Zhang, D., Song, C., Lu, G.: Touchless multiview fingerprint acquisition and mosaicking. IEEE Trans. Instrum. Meas. 62, 2492–2502 (2013)CrossRefGoogle Scholar
  4. 4.
    Kumar, A., Kwong, C.: Towards contactless, low-cost and accurate 3D fingerprint identification. IEEE Trans. Pattern Anal. Mach. Intell. 37, 681–696 (2015)CrossRefGoogle Scholar
  5. 5.
    Hartley, R., Zisserman, A.: Multiple View Geometry in Computer Vision. Cambridge University Press, Cambridge (2004)CrossRefzbMATHGoogle Scholar
  6. 6.
    Storn, R., Price, K.: Differential evolution - a simple and efficient heuristic for global optimization over continuous spaces. J. Global Optim. 11, 341–359 (1997)MathSciNetCrossRefzbMATHGoogle Scholar
  7. 7.
    Qin, A., Huang, V., Suganthan, P.: Differential evolution algorithm with strategy adaptation for global numerical optimization. IEEE Trans. Evol. Comput. 13, 398–417 (2009)CrossRefGoogle Scholar

Copyright information

© ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2017

Authors and Affiliations

  1. 1.Marketing Analytics and Insight, Data2DecisionsSydneyAustralia
  2. 2.School of Engineering and Information TechnologyThe University of New South WalesCanberraAustralia

Personalised recommendations