A mathematical model of changes of parameters of handwritten patterns depending on the signers states

A E Samotuga
2018 Journal of Physics, Conference Series  
Perspectives of subjects' psychophysiological state identification using dynamic biometric features P S Lozhnikov, A E Sulavko, R V Borisov et al. -Mathematical and information maintenance of biometric systems Z Boriev, S Sokolov, A Nyrkov et al. -This content was downloaded from IP address on 13/04 Abstract. Authors of some studies point out the need of taking into account the state of employee to reduce the risk that he will cause damage to organization including information
more » ... ng information resources. An assessment of psychophysiological state is possible on the basis of the dynamic parameters of a signature. Etalons creation of handwritten signature for each state significantly complicates the formation of an etalons base of subjects. The purpose of the study is to provide a mechanism for the formation of subjects etalons in different states without the need to collect of handwritten signature in each states. To achieve the goal we should solve the problem of finding the regularity of the signature parameters changes while changing the states of signers. We develop a model of changing signature parameters depending on the state based on the discovered regularities. We compared the accuracy of the obtained model by estimating the probabilities of erroneous decisions obtained in the experiment of states recognition using etalons obtained in a natural way and using a model. To carry out the research, we used the apparatus of probability theory and mathematical statistics. This included using the multidimensional Bayes functional and Bayes hypothesis formula. As a result the author describes the model of biometric etalons of handwritten signatures of the user. The model assumes the use of formulas and calculated coefficients for calculating the distribution of signature features characteristics in the changed state from the values of the signature features in the normal state. The introduction of this model will significantly simplify the procedure for obtaining the etalons in changed state The mentioned above model can find applications in systems of dynamic biometric identification.
doi:10.1088/1742-6596/1050/1/012076 fatcat:i7k5nepyynadvhmz6lx7g3efdq