SCIENCE CHINA Information Sciences, Volume 62 , Issue 1 : 019104(2019) https://doi.org/10.1007/s11432-018-9512-1

Learning discriminative and invariant representation for fingerprint retrieval

More info
  • ReceivedApr 27, 2018
  • AcceptedMay 30, 2018
  • PublishedDec 19, 2018


There is no abstract available for this article.


This work was supported by National Natural Science Foundation of China (Grant No. 61333015).


Appendixes A–D.


[1] Maltoni D, Maio D, Jain A K, et al. Handbook of Fingerprint Recognition. Berlin: Springer, 2003. Google Scholar

[2] Gago-Alonso A, Hernández-Palancar J, Rodríguez-Reina E. Indexing and retrieving in fingerprint databases under structural distortions. Expert Syst Appl, 2013, 40: 2858-2871 CrossRef Google Scholar

[3] Xinjian Chen , Jie Tian , Xin Yang . A new algorithm for distorted fingerprints matching based on normalized fuzzy similarity measure. IEEE Trans Image Process, 2006, 15: 767-776 CrossRef ADS Google Scholar

[4] Tang Y, Gao F, Feng J, et al. FingerNet: an unified deep network for fingerprint minutiae extraction. In: Proceedings of IEEE International Joint Conference on Biometrics, 2017. 108--116. Google Scholar

[5] Krizhevsky A, Sutskever I, Hinton G E. ImageNet classification with deep convolutional neural networks. In: Proceedings of Annual Conference on Neural Information Processing Systems, 2012. 1097--1105. Google Scholar

[6] He K, Zhang X, Ren S, et al. Delving deep into rectifiers: surpassing human-level performance on imagenet classification. In: Proceedings of the IEEE International Conference on Computer Vision, 2015. 1026--1034. Google Scholar

[7] Wang J, Li Y, Miao Z. Learning deep discriminative features based on cosine loss function. Electron Lett, 2017, 53: 918-920 CrossRef Google Scholar

[8] Song D, Feng J. Fingerprint indexing based on pyramid deep convolutional feature. In: Proceedings of IEEE International Joint Conference on Biometrics, 2017. 200--207. Google Scholar

  • Figure 1

    (Color online) The distribution of learned deep convolutional features $\boldsymbol~f_c$ (without normalization) in fingerprint training data. A two-dimensional feature $\boldsymbol~f_c$ is learned by the DCNN with ten fingerprint classes. (a) Activation functions; (b) the distribution of features with rectified linear unit (ReLU); (c) the distribution of features with power function.


Contact and support