Evaluation of signature verification reliability based on artificial neural networks, Bayesian multivariate functional and quadratic forms
Ivanov A.I., Lozhnikov P.S., Sulavko A.E.

 

Penza Scientific and Research Electrotechnical Institute, Penza, Russia,
Omsk State Technical University, Omsk, Russia

Full text of article: Russian language.

Abstract:
An experimental comparison of various functional neural networks for signature verification is performed. A signature database for the realization of the computing experiment is built. It is confirmed that up to a certain point, the increase of the decision rule dimension reduces the probability of signature verification error, with an increase in the number of neurons in the network reducing the number of errors. A higher-dimension multi-dimensional Bayes functional with stronger inter-feature correlation is found to perform better. The best result for the signature verification is obtained using networks of Bayesian multidimensional functional, with false acceptance rate of FRR = 0.0288 and false rejection rate of FAR = 0.0232.

Keywords:
neural networks, network of quadratic forms, multi-dimensional Bayes functional, signature reproduction peculiarities, biometric features.

Citation:
Ivanov AI, Lozhnikov PS, Sulavko AE. Evaluation of signature verification reliability based on artificial neural networks, Bayesian multivariate functional and quadratic forms. Computer Optics 2017; 41(5): 765-774. DOI: 10.18287/2412-6179-2017-41-5-765-774.

References:

  1. The Global State of Information Security® Survey 2016. PricewaterhouseCoopers. Source: <http://www.pwc.com/ gx/en/issues/cyber-security/information-security-survey/do­wnload. html>.
  2. Ivanov AI, Lozhnikov PS, Samotuga AE. A technology to form hybrid documents [In Russian]. Cybernetics and Systems Analysis 2014; 50(6): 152-156.
  3. GOST R 52633.0-2006. Information protection. Information protection technology. Requirements to the means of high-reliability biometric authentication [In Russian]. Moscow: "Standartinform" Publisher; 2006.
  4. Lozhnikov PS, Sulavko AE, Eremenko AV, Volkov DA. Experimental evaluation of reliability of signature verification by quadratic form networks, fuzzy extractors and perceptrons [In Russian]. Information and Control Systems 2016, 5(84): 73-85. DOI: 10.15217/issn1684-8853.2016.5.73.
  5. Dodis Y, Reyzin L, Smith A. Fuzzy extractors: How to generate strong keys from biometrics and other noisy. EUROCRYPT 2004: 523-540. DOI: 10.1007/978-3-540-24676-3_31.
  6. Ahmetov BS, Ivanov AI, Funtikov VA, Bezjaev AV, Malygina EA. Technology of large neural networks usage for fuzzy biometric data conversion to access key codes: monograph. Almaty, Kazakhstan: "LEM" Publisher; 2014.
  7. Ivanov AI. Neural network protection of confidential biometric data and private cryptographic keys: A monograph [In Russian]. Penza: "PNIEI" Publisher; 2014.
  8. Galushkin AI. Synthesis of multi-level systems for pattern recognition [In Russian]. Moscow: "Energija" Publisher; 1974.
  9. Hinton GE. Training products of experts by minimizing contrastive divergence. Neural Comput 2002; 14(8): 1771-1800. DOI: 10.1162/089976602760128018.
  10. Hafemann LG, Sabourin R, Oliveira LS. Writer-indepen­dent feature learning for offline signature verification using deep convolutional neural networks. IJCNN 2016: 2576-2583. DOI: 10.1109/IJCNN.2016.77275212016.
  11. Kolmogorov AN. On the representation of a multivariate continuous function as a superposition of monovariate continuous functions [In Russian]. Doklady Akademii Nauk SSSR 1957; 114(5): 953-956.
  12. GOST R 52633.5-2011. Information protection. Information protection technology. The neural net biometry-code converter automatic training [In Russian]. Moscow: "Standartinform" Publisher; 2011.
  13. Ivanov A.I. Neural network algorithms for biometric personal identification [In Russian]. Moscow: "Radiotehnika" Publisher; 2004. ISBN: 978-5-93108-048-2.
  14. Ivanov AI, Lozhnikov PS, Kachajkin EI. Verification of authenticity for handwritten signatures using Bayesian-Hamming networks and quadric form networks. Information Security Questions 2015; 2: 28-34.
  15. Lozhnikov PS, Ivanov AI, Kachajkin EI, Sulavko AE. Biometric identification of handwritten images via correlation analog of Bayes' rule [In Russian]. Information Security Questions 2015; 3: 48-54.
  16. Ivanov AI, Lozhnikov PS, Serikova JuI. Reducing the size of training-sufficient sampling due to symmetrization of correlation relationships of biometric data [In Russian]. Cybernetics and Systems Analysis 2016; 52(3): 49-56.
  17. Bezev AV, Ivanov AI, Funtikova JuV. Optimization of the structure self-correcting bio-code, storing syndromes error as fragments hash-functions [In Russian]. UrFR Newsletter. Information Security 2014; 3(13): 4-13.

© 2009, IPSI RAS
Institution of Russian Academy of Sciences, Image Processing Systems Institute of RAS, Russia, 443001, Samara, Molodogvardeyskaya Street 151; e-mail: ko@smr.ru; Phones: +7 (846 2) 332-56-22, Fax: +7 (846 2) 332-56-20