Fingerprints Synthesis Using Generative Adversarial Neural Network: Minutiae Structure Analysis
Abstract
The growing use of fingerprints drives the emergence of studies to enhance the technology, demanding higher quality fingerprint images to ensure good results. Due to the sensitivity of LGPD in Brazil since 2018, sharing fingerprint data is hesitant. In this context, this study creates realistic synthetic fingerprints using the StyleGAN-ADA neural network. The results are evaluated using the Earth mover's distance (EMD) metric to compare 2D distributions, and the Minutiae Histograms (MHs) method to map minutiae distributions. This approach succeeds with the FVC database, generating images that reflect realistic minutiae distributions according to the employed metrics.
References
Carsten Gottschlich, S. H. (2013). Separating the real from the synthetic: minutiae histograms as fingerprints of fingerprints, IET (Institution of Engineering and Technology) Biometrics 3(4): 291–301.
Goodfellow, I., Pouget-Abadie, J., Mirza, M., Xu, B., Warde-Farley, D., Ozair, S., Courville, A. and Bengio, Y. (2014). Generative adversarial nets, in Z. Ghahramani, M. Welling, C. Cortes, N. D. Lawrence and K. Q. Weinberger (eds), Advances in Neural Information Processing Systems 27, pp. 2672–2680.
Karras, T., Aittala, M., Hellsten, J., Laine, S., Lehtinen, J. and Aila, T. (2020). Training generative adversarial networks with limited data, 2: 37.
Planalto (2018). Lei nº 13.709, de 14 de agosto de 2018: Lei geral de proteção de dados pessoais (lgpd). URL: [link]
Zhang, H., Zhang, Z., Odena, A. and Lee, H. (2020). Consistency regularization for generative adversarial networks, International Conference on Learning Representations – to appear. URL: [link]