Structured patterns retrieval using a metric attractor network: application to fingerprint recognition. (English) Zbl 1400.94014
Summary: The ability of a metric attractor neural networks (MANN) to learn structured patterns is analyzed. In particular we consider collections of fingerprints, which present some local features, rather than being modeled by random patterns. The network retrieval proved to be robust to varying the pattern activity, the threshold strategy, the topological arrangement of the connections, and for several types of noisy configuration. We found that the lower the fingerprint patterns activity is, the higher the load ratio and retrieval quality are. A simplified theoretical framework, for the unbiased case, is developed as a function of five parameters: the load ratio, the finiteness connectivity, the density degree of the network, randomness ratio, and the spatial pattern correlation. Linked to the latter appears a new neural dynamics variable: the spatial neural correlation. The theory agrees quite well with the experimental results.
MSC:
| 94A08 | Image processing (compression, reconstruction, etc.) in information and communication theory |
| 62M45 | Neural nets and related approaches to inference from stochastic processes |
References:
| [1] | Amit, Daniel J., Modeling Brain Function: The World of Attractor Neural Networks (1989), Cambridge University Press: Cambridge University Press New York, NY, USA · Zbl 0709.92001 |
| [2] | Hertz, J.; Krogh, J.; Palmer, R., Introduction to the Theory of Neural Computation (1991), Addison-Wesley: Addison-Wesley Boston |
| [3] | González, Mario; Dominguez, David; Rodríguez, Francisco B., Block attractor in spatially organized neural networks, Neurocomputing, 72, 16-18, 3795-3801 (2009) |
| [5] | Baldi, Pierre; Chauvin, Yves, Neural networks for fingerprint recognition, Neural Comput., 5, 3, 402-418 (1993) |
| [6] | Rashid, M. M.; Hossain, A. K., Fingerprint verification system using artificial neural network, Inf. Technol. J., 5, 6, 1063-1067 (2006) |
| [9] | González, Mario; Dominguez, David; Rodríguez, Francisco B.; Sanchez, Angel, Retrieval of noisy fingerprint patterns using metric attractor networks, Int. J. Neural Syst., 24, 07 (2014) |
| [11] | Pérez, Patricio; Salinas, Dino, Storage of structured patterns in a neural network, Phys. Rev. E, 50, 4182-4186 (1994) |
| [12] | Biehl, Michael; Kühn, Reimer; Stamatescu, Ion-Olimpiu, Learning structured data from unspecific reinforcement, J. Phys. A: Math. Gen., 33, 39, 6843 (2000) · Zbl 0970.68134 |
| [13] | Monasson, R., Properties of neural networks storing spatially correlated patterns, J. Phys. A: Math. Gen., 25, 13, 3701 (1992) · Zbl 0791.68140 |
| [15] | Koyama, Shinsuke, Storage capacity of two-dimensional neural networks, Phys. Rev. E, 65, Article 016124 pp. (2001) |
| [16] | Nishimori, Hidetoshi; Whyte, W.; Sherrington, D., Finite-dimensional neural networks storing structured patterns, Phys. Rev. E, 51, 3628-3642 (1995) |
| [17] | Maltoni, Davide; Maio, Dario; Jain, Anil K.; Prabhakar, Salil, Handbook of Fingerprint Recognition (2009), Springer · Zbl 1027.68114 |
| [18] | Dominguez, D.; Koroutchev, K.; Serrano, E.; Rodríguez, F. B., Information and topology in attractor neural networks, Neural Comput., 19, 956-973 (2007) · Zbl 1118.68116 |
| [19] | Olshausen, Bruno A.; Field, David J., Sparse coding of sensory inputs, Curr. Opini. Neurobiol., 14, 4, 481-487 (2004) |
| [20] | Dominguez, David; González, Mario; Serrano, Eduardo; Rodríguez, Francisco B., Structured information in small-world neural networks, Phys. Rev. E, 79, 2, Article 021909 pp. (2009) |
| [21] | Derrida, B.; Gardner, E.; Zippelius, A., An exactly solvable asymmetric neural network model, Europhys. Lett., 4, 2, 167 (1987) |
This reference list is based on information provided by the publisher or from digital mathematics libraries. Its items are heuristically matched to zbMATH identifiers and may contain data conversion errors. In some cases that data have been complemented/enhanced by data from zbMATH Open. This attempts to reflect the references listed in the original paper as accurately as possible without claiming completeness or a perfect matching.
