350 rub
Journal Neurocomputers №9 for 2016 г.
Article in number:
The using of existing handwriting techniques for writer identification
Authors:
S.D. Kulik - Ph.Dr. (Eng.), Senior Research Scientist, National Research Nuclear University «MEPHI» (Moscow). E-mail: sedmik@mail.ru D.A. Nikonets - Ph. Dr. (Eng.), Senior engineer-programmer, Company Informcontact Consulting CJSC (Moscow). E-mail: denis_n@mail.ru
Abstract:
Questioned document examination, particularly forensic handwriting examination, is one of the types of forensic human identifi-cation. One of the most frequent problems in handwriting expert practice is the task of handwritten text writer verification. Verifi-cation refers to the task when expert is presented with two handwritten samples and it is necessary to determine whether they were made by the same performer. The technique for verification of the writer of handwritten document, developed by our re-search group together with the experts of the Forensic Center of the Ministry of Internal Affairs of the Russian Federation is also represented. The major issues, solved during the development of this technique, are described. One of the current, but not solved problems in forensic handwriting examination is developing a software system for the writer identification (identification search). Identification refers to the task when handwritten sample is presented and it is neces-sary to determine by which of the known (contained in the database) writers this document is performed. Currently, handwriting technique for writer identification has not been implemented. In the presented paper the possibility to use an existing handwriting technique for writer verification for identification of the writer of the handwritten text, the problems and possible solutions, options for the modernization of existing and develop-ment of new handwriting techniques are discussed. The paper also concludes that for the development of the new handwriting techniques, statistical methods, which are more resistant to the dependence of the handwriting features, such as neural networks, must be used. The use of such statistical methods will improve the quality of decisions made by the handwriting techniques.
Pages: 64-70
References

 

  1. Sudebno-pocherkovedcheskaja ehkspertiza: obshhaja chast: teor. i metod. osnovy / Pod nauch. red. V.F. Orlovojj: Gosudarstvennoe uchrezhdenie RFCSEH pri Minjuste Rossii. Izd. 2-e, pererab. i dop. M.: Nauka. 2006. 544 s.
  2. Kulik S.D., Nikonec D.A., Tkachenko K.I., Lukjanov I.A. Metody i sredstva povyshenija ehffektivnosti informacionnykh sistem (nejjronnye seti, kriminalistika, formirovanie faktograficheskikh dannykh, morfologicheskijj analiz). T 1. Kriminalistika (Dep. v VINITI 05.05.2011, № 206-132011). M.: Radiotekhnika. 2011. 300 s.
  3. Nikonec D.A. Nejjrosetevye tekhnologii dlja reshenija zadach kriminalistiki // Nejjrokompjutery razrabotka primenenie. 2009. № 4. S. 26-45.
  4. Kulik S.D., Nikonec D.A. Primery ispolzovanija nejjrosetevogo algoritma v metodikakh ehksperta-pocherkoveda // Nejjrokompjutery razrabotka primenenie. 2009. № 9. S. 61-85.
  5. Kulik S.D., Nikonec D.A. Avtomatizacija klassifikacionno-diagnosticheskikh pocherkovedcheskikh issledovanijj s pomoshhju nejjronnykh setejj // Informacionnye tekhnologii. 2012. № 1. S. 70-75.
  6. Averjanova T.V., Belkin R.S., Korukhov JU.G., Rossinskaja E.R. Kriminalistika: Uchebnik dlja vuzov / Pod red. prof. R.S. Belkina. M.: NORMA. 2001. 990 s.
  7. Belkin R. S. Kriminalisticheskaja ehnciklopedija. Izd. 2-e dop. M.: Megatron XXI. 2000. 334 s.
  8. Kulik S.D., CHelyshev M.M., Levickijj A.B., Bazhakin G.A., Belousova O.D., Murashova O.S., Kolesova E.JU. Metodika verojatnostno-statisticheskojj ocenki sovpadajushhikh chastnykh priznakov pocherka v propisnykh bukvakh russkogo alfavita: Spravochnoe posobie. M.: VNII MVD SSSR. 1990. 260 s.
  9. Nikonec D.A., Kulik S.D. Problema ispolzovanija sushhestvujushhikh pocherkovedcheskikh metodik dlja organizacii identifikacionnogo poiska ispolnitelja rukopisi // Tezisy dokl. XIV Vseross. nauch. konf. «Nejjrokompjutery i ikh primenenie». M.: MGPPU. 2016. S. 77-78.
  10. Kulik S.; Nikonets D. Forensic Handwriting Examination and Human Factors: Improving the Practice through Automation and Expert Training // Proc. 2016 Third International Conference on Digital Information Processing, Data Mining, and Wireless Communications (DIPDMWC). P. 221-226.
  11. ZHenilo V.R., Vinkova O.M., Naumova V.V., Poljakova A.V. O sozdanii chastotnogo atlasa individualnykh artikuljacionnykh osobennostejj proiznesenija zvukov rechi // Sb. trudov XVIII Mezhdunar. konf. «Informatizacija i informacionnaja bezopasnost pravookhranitelnykh organov». 19-20 maja 2009 g. M.: Akademija upravlenija MVD Rossii. 2009. S. 356-368.
  12. Srihari S.N., Su C. Computational Methods for Determining Individuality // Proc. Int. Workshop on Computational Forensics, Washington DC. Springer LNCS 5158, 2008. P. 11-21.
  13. Sekejj G. Paradoksy v teorii verojatnostejj i matematicheskojj statistike. M.: Mir. 1990. 240 s.
  14. Kaye D.H. Beyond uniqueness: the birthday paradox, source attribution and individualization in forensic science testimony // Law, Probability and Risk. 2013. V. 12 (1). P. 3-11.
  15. Page M., Taylor J., Blenkin M. Uniqueness in the Forensic Identification Sciences - Fact or Fiction? // Forensic Sci. Int-l. 2011. Vol. 206. P. 14-18.
  16. Gittelson S., Biedermann A., Bozza S., Taroni F. The database search problem: A question of rational decision making // Forensic Sci. Int-l. 2012. V. 222. P. 186-199.
  17. Puri M., Srihari S.N., Tang Y. Bayesian Network Structure Learning and Inference Methods for Handwriting // Proc. Int. Conf. Document Analysis and Recognition, Washington DC, Aug. 2013. P. 1320-1324.
  18. Srihari S.N. Evaluating the Rarity of Handwriting Formations // Proc. Int. Conf. Doc. Anal. Recog. ICDAR, 2011. P. 618-622.
  19. Aitken C., Taroni F. Statistics and the evaluation of evidence for forensic scientists. 2nd ed. John Wiley & Sons, 2004.
  20. Srihari S.N. Statistical Examination of Handwriting Characteristics using Automated Tools. US DoJ Report. 2013. 85 p.