RU114194U1 - PERSONALITY IDENTIFICATION DEVICE - Google Patents

Abstract

A device for identifying a person by biometric dynamic characteristics of handwriting by converting primary physical quantities into biometric parameters, containing primary converters of physical quantities, an instrument part, characterized in that the device is made in the form of a writing instrument in which the primary transducers of physical quantities are located, the instrument part contains a pressure sensor associated with a device for primary conversion of a physical quantity, an acceleration sensor associated with a device for primary conversion of a physical quantity, a position sensor associated with a device for primary conversion of a physical quantity, an input-output device, an electronic computing device, a storage device for storing signals of primary converters of physical quantities, while the vector of biometric parameters is calculated by the formula:! ! where vi, - presented parameters; ! - discrete frequency spectrum.

Description

The utility model relates to computer technology and is intended to identify a person by individual biometric dynamic characteristics of handwriting, identified by the user playing a previously known passphrase / word or signature.

The well-known "Shift method for signature control" according to US patent N 4562592, MKI4 G06K 9/22, NCI 382-2, UDC 681.362. Publication 851231 T.1061 N 5 (VINIPI 111-10-86). According to this method, acceleration data and compression data are obtained and compared with standards (acceleration and compression). In this case, a shift of the compressed signals (presented and reference) is performed, the maximum point of the correlation function is found, and the shift corresponding to the maximum is stored. Further, it is precisely for this shift that the correlation value of the pairs of acceleration data (presented and reference) is controlled. This method, on the one hand, allows you to partially eliminate the uncertainty of the position and scale of the signature, however, it does not allow accurate correction of the time scales and the amplitude of the signature. In general, this method does not allow the identification of a person with a sufficiently high reliability.

The latter drawback is partially compensated by the “Method for comparing manual recording with a reference recording and application of the method” according to Swiss patent N 665915, MKI4 G06K 9/62, UDC 621.327. Publication 880615 N 11 (VINIPI 117-03-89). According to this method, the reference and presented signals are divided into separate fragments (for example, fragments of equal length) and each of these fragments is separately combined with the corresponding standard, and also individually scaled in time. The latter improves the probability of a correct solution, but still does not allow to obtain a sufficiently high reliability.

Also known is the "Method and device for identifying handwriting" according to Swiss patent N 659531, MKI4 G06K 9/62, G06D 7/00, UDC 681.327.12, publication 870130 N 2. This patent uses the maximum amount of initial information, analyzes the pen oscillation curve according to one axis - Y (t), the curve of the pen oscillation along the second axis - X (t) is analyzed, the curve of the pen pressure on the substrate - P (t) is analyzed. Three signals are digitized, entered into a computer or microprocessor and compared with the standards. The main disadvantage of this technical solution is the insufficiently high probability of a correct identification decision.

All of the above methods far from fully use the information obtained in the form of curves of pen and pressure fluctuations, which leads to insufficiently high reliability of the decision. Closest to the claimed is a device for implementing a method of identifying a person according to the signature on patent RU No. 2148274, G06K 9/22, G06K 9/62, G06F 15/18, publ. 04/27/2000 (prototype).

The essence of the known invention lies in the fact that they enter into the computer digitized vibrations of a pen that reproduces the signature of an identifiable person, and its pressure on the graphics tablet with subsequent determination of the beginning and end of the signature of an identifiable person, fragment the signature, scale each fragment of this signature, calculate differential and integral parameters of fragments of the mentioned signature, and also calculate the time intervals of fragments of this signature, decide identifying a person by comparing said calculated parameters and time intervals with their reference values, after determining said differential and integral parameters, re-scaling said calculated parameters and time intervals minimizing the average deviation of said calculated parameters and time intervals from their reference values, correlation coefficients of these calculated parameters and time intervals, and also estimate and these correlation coefficients at said deciding compare these estimates of correlation coefficients with their reference values.

The objective of the utility model is to increase the reliability of personal identification by the biometric dynamic characteristics of handwriting due to a more complete use of the source information.

The problem is solved in that, in contrast to the known device for identifying a person by signature, by converting primary physical quantities to biometric parameters, containing primary converters of physical quantities, the instrument part, according to a utility model, the inventive device is different in that it is made in the form of a writing instrument, which are the primary converters of physical quantities, the instrument part contains a pressure sensor associated with the primary conversion device a physical quantity, an acceleration sensor associated with a device for primary conversion of a physical quantity, a position sensor associated with a device for primary conversion of a physical quantity, an input / output device, an electronic computing device, a storage device for storing signals of primary converters of physical quantities, the vector of biometric parameters calculated by the formula:

,

where v _i - presented parameters;

- discrete frequency spectrum.

The technical result achieved by the proposed utility model is to more fully use the source information, increase the accuracy of changes in the initial controlled parameters and increase the total number of controlled parameters by attracting secondary controlled parameters obtained from primary parameters due to a new set of features, which allows us to attribute the proposed Utility model for technical solutions that meet the condition of patentability “novelty”.

The utility model is illustrated by schematic images: Fig. 1 shows a structural diagram of a biometric user identification system according to the dynamics of handwriting reproduction; figure 2 - parameters measured during the reproduction of handwriting.

A device for identifying a person according to the biometric dynamic characteristics of handwriting contains: a device 1 for collecting information in the form of a writing instrument; located in the information collection device 1, a pressure sensor 2, an acceleration sensor 3, a position sensor 4; devices 5, 6, 7 of the primary conversion of physical quantities, an input-output device 8; electronic computing device 9; storage device 10; system power device 11.

The device operates as follows.

The user using the information collecting device 1 reproduces the handwriting (by means of a previously known passphrase / word or signature). In this case, the sensors 2, 3, 4, located in the device for collecting information, convert non-electrical quantities into electrical signals. The pressure sensor 2 converts the tensile-compression strain along the length into a frequency signal. The acceleration sensor 3 registers acceleration projections on the three axes X, Y, Z. The position sensor 4 is designed to measure the rotation angle and tilt angles of the information collection device, since the user can rotate the information collection device when re-entering the passphrase / word or signature, which will lead to change the orientation of the axes of the acceleration sensor. The signals coming from sensors 2-4 are processed using devices 5, 6, 7 of the primary conversion of physical quantities and electronic computing device 9 and are transmitted to memory 10. In electronic computing device 9, data is processed, the formation of a vector of biometric informative parameters and a biometric standard / sample personality. The biometric standard / sample of personality can be transferred to external devices through the device 8 input-output. When a user is registered in the system, a biometric standard of personality is formed. Since dynamic human images have significant variability, the formation of a biometric standard requires several examples of realizations of the same sample. The formation of the biometric standard is carried out on an electronic computing device 9, the generated standard is stored in the database of the device 1 for collecting information. When the user enters a previously known password phrase / word or signature, the individual biometric dynamic characteristics of the handwriting are reflected in the frequency structure of the functions of pressure, angle, orientation and acceleration, and the task of biometric analysis is reduced to extracting this information.

The obtained discrete signals of biometric parameters, using a discrete Fourier transform, are converted into a discrete frequency spectrum according to the formula:

Further information processing is carried out by moving from the signal spectrum to the energy spectrum of the implementation.

The energy spectrum of the implementation is calculated by the formula:

The resulting spectrum is considered as an N-dimensional vector of informative biometric parameters that form the vector of biometric parameters:

Vector V is the source for the subsequent identification / authentication procedure.

In the identification mode, the vector of controlled biometric parameters obtained from the presented image is formed and compared with the biometric standard from the database in the information collection device. The verification program (comparison) gives the percentage of similarity of the vector with the standard. The result of comparing the presented sample with the standard is the conclusion about the successful / unsuccessful procedure. This system is intended for personal identification.

Industrial applicability of the proposed device is provided by well-known technologies in computer technology.

Claims (1)

A device for identifying a person according to the biometric dynamic characteristics of handwriting by converting primary physical quantities to biometric parameters containing primary converters of physical quantities, an instrument part, characterized in that the device is a writing instrument in which primary converters of physical quantities are located, the instrument part contains a pressure sensor associated with the device of the primary conversion of physical quantities, the acceleration sensor associated with the device m of the primary conversion of a physical quantity, a position sensor associated with the device of the primary conversion of a physical quantity, an input-output device, an electronic computing device, a storage device for storing signals of primary converters of physical quantities, the vector of biometric parameters calculated by the formula:

where v _i

- presented parameters;

- discrete frequency spectrum.