Method for personal identification from handwritten text

FIELD: physics.

SUBSTANCE: method includes first creating a database from reference handwritten texts converted to digital in the form of patterns and matrices containing identification parameters in the form of average angles of projection of text, obtained by dividing the text into separate fragments, breaking down the picture of the text into elementary components and linearisation thereof;when analysing a new presented handwriting sample, creating a pattern and an identification matrix similar to the reference sample; comparing and making a decision to associate the presented handwritten text with one of the references.

EFFECT: high reliability of personal identification.

2 cl, 3 dwg

 

The invention relates to methods of reading and identification information that can be used to authenticate the identity of handwritten text and can be used in investigative practice and the security systems of government agencies and banks.

The task of establishing by the manuscript is in the handwriting identification task. The solution to this problem is the most common type of forensic examination.

Identification of a handwritten signature (and its dynamics play) is one of the oldest. Centuries ago began to use the signature to prove the authenticity of certain securities. Some authors [1, 2] consider appropriate separation to the problem of authentication of the person by facsimile signature in two independent tasks:

- identification only on static signatures, which is set in advance on the scanned document;

- identification by the dynamics of reproduction, that is, at the time of signing, with the possibility of observation of the individual characteristics of the process.

Both problems can be solved in parallel and independently. The first task is static, you just need to compare the acquired image of the pattern of a letter, which is available in the database. For modern information technologies it is large�but challenging. And the quality of its solution still leaves much to be desired, although achieved some significant results to date [1].

In the second task processed the data about the dynamics of reproduction of the drawing of the text (coordinates of the pen, vibrations, pressure, time intervals, etc.).

To identify the executor of the manuscript in his handwriting are the most important features of graphics and technical writing skills. Existing methods to solve the identification problem is based on the identification identificazione significant criteria [1, 3]. Therefore, the urgent task is the creation of computer methods to solve the identification task, taking on complex processing of handwritten material. The basic idea is that complex treatment extensive pokerowego material to highlight the few identificazione-significant criteria to facilitate the work of the expert is replaced by processing of a computer most full array of attributes without further differentiation identification of significance. The possibility of exhaustive search, which is available to the computer, allows you to automatically allocate significant signs. It is known that a person's handwriting contains a sufficient number of individual characteristics, allowing you to securely identify a person on individual characteristics of handwriting [1, 2, 3].However, a comparison of handwriting samples, held in investigative practice, specific to the user, requires a rather complicated techniques and experience of the researcher of forensic expert. Attempts to apply the experience gained in the field of forensic handwriting examination and graphology (methods of characterization by handwritten texts, developed by experts of the forensic science center of the Ministry of internal Affairs of Russia) for identification of signatories in information security systems do not allow to identify individuals in real time, in automatic mode, or to take a reliable decision on the admission of the subject to a protected system. So remains the problem of creating a fast and reliable identification system, allowing you to select the set of features that reliably separates different people, and to determine the values of these features.

A method of character recognition of the text information of the bitmap "Method of text recognition with application custom classifier", EN No. 2234126, G06K 9/66, publ. 10.08.2004. In this method an image of the character is recognized by using custom and/or non-configurable classifiers and contextual analysis. This method allows you to recognize a high-quality printed text and does not recognize handwritten texts because of the natural instability training�'s letters. This method does not allow for the identification of handwritten text.

The last drawback is partially compensated by the "Method of comparison of manual recording with a reference record and applications of the method" in patent Switzerland No. 665915, MKI 4 G06K 9/62, UDC 621.327. Publication 880615 No. 11 (WINAPI 117-03-89).

In this method, the reference and the texts submitted are divided into fragments and each of these fragments is separately combined with the relevant benchmark and scale individually. The latter improves the probability of correct decision, but it is not possible to obtain a sufficiently reliable identification.

The closest in technical essence as a prototype for the proposed method is "a Method of biometric authentication of the handwriting in the computerized access control system", patent RU No. 2469397 from 10.12.12, G06K 9/00, G01L 17/00, namely that converted to digital form oscillations of the pen, reproducing handwritten text, and its pressure is introduced into the computer, forming a matrix of quantized samples, the values of its elements calculated using two-dimensional discrete cosine transform matrix coefficients, decide to identify the person by comparing the mentioned calculated coefficients with their reference values in the database (the data�, in this re-introduced into the system user is identified by successive subtraction modulo elements of its matrix of coefficients of two-dimensional discrete cosine transform from the corresponding elements of the matrices in the database, and recognizes the user is considered incidental reference record, if this difference is minimal. The disadvantage of the prototype is the necessity of storing large amounts of data, and most importantly - the inability to identify the person on the "dead" handwritten text. Common drawbacks should also be considered as the need to carry out a comparison of the studied requirements of a large number of reference, which, ultimately, leads to a large delay and reduces the reliability of recognition. These drawbacks significantly limit the possibility of using the known methods in investigative practice for recognition of individual handwritten text information.

The problem solved by the present invention is to improve the reliability of identification by the peculiarities of handwriting by creating a view (template) handwriting, which can be preliminary forensic analysis more simple, accessible way on the computer. Thus created templates could�and would be used in investigative practice for the classification of the personalities, which these signatures, likely belonging to one type or the other, depending on the type of research. For example, to identify the likelihood of forgery of the signature.

The problem is solved by providing a method of identification by handwriting, in which after repeated scanning of the reference handwriting is exercised by the division of the text into separate fragments, allocate to each example of a fragment of a handwritten text line text path, then crushed a drawing of a piece of text on the unit cell, so that within a unit cell receive straight text path, then for all implementations of the standard text on each fragment for each unit of its cell calculates the average value and standard deviation of the slope angle, the calculated values form the identity matrix, the dimension of which is equal to the dimension of the fragment, and non-zero values which contain the value of average slope angles and root-mean-square deviations, remember these average values of orientation angles and root-mean-square deviations and use them when you identify a person by his handwriting; in the identification process presented handwriting exercise split scanned rice�ka studied the handwritten text into fragments, their scaling, allocation on each piece of handwritten text lines, text paths, the fragmentation pattern of the text fragment in the unit cell is similar to the fragmentation of the fragment of the reference text, the calculation for each elementary cell of a fragment of the values of the angle of the text, the formation by the calculated values of the angles of inclination of the matrix of angles, the dimension of which is equal to the dimension of the identity matrix fragment of benchmark, a comparison of these values of the angle of the text with the corresponding values of the average tilt angles of the fragment of the reference text from the database in the case of rotation of the template fragment reference signature for a complete matching of significant template cells adjust the angle of the trajectory of the studied text on each relevant cell within the angle of rotation of the template, while in the case of coincidence of the calculated values of angles of inclination with average tilt angles of the fragment of the reference text with a given probability the analyzed text is considered copyrighted, and in case of discrepancy between the calculated and reference values, the identification procedure is terminated; moreover, the method allows to classify imposed handwriting on the degree of authenticity with reference to classes of coincidence with the reference text. The stated set of essential features �allows you to create a computerized system of identification by signature.

An example of implementing this method is illustrated by drawings, in which Fig.1 shows a block diagram implementing the method, where 1 - scan standard handwriting, 2 - fragmentation of the standard, 3 - fragmentation of the fragment, 4 - calculation of average corners, 5 - formation of the template and the identity matrix, 6 - formatting the studied text, 7 - hanging of the standard template, 8 - affine transformation of the template standard, 9 - calculation of average angles of a fragment of the studied text, 10, formation of the matrix of average angles of a fragment of the studied text, 11 - decision-making on the identification of the fragment of the studied text; Fig.2 shows the result of the fragmentation pattern of a handwritten text image on the elementary cell (rectangular cells), each cell shows the average slope of lines of text, Fig. 3 education pattern of the reference text and the identification of the matrix of Fig. 4 - the results of the affine transformation of the studied text (for example, one fragment).

The implementation of the proposed method is carried out in several stages, as shown in Fig. 1. Initially at the stage of pre-processing sample copyright text (text, passphrase, signature) is recorded repeatedly (more than 20-30), anonymously, in a familiar user environment (can be used murals in the Bank�their documents, the paintings in the books of account of the parish, in the library, forms, etc.) to create the database. This sample of handwritten text is clean known in investigative practice (1 in Fig.1). The text is divided (fragmented) into individual pieces (familiarity) (2 in Fig.1), the format of the fragment is determined by the size of the played character (height and width), in some cases, in the identification of signatures that can be one format. At the first presentation of the handwritten text, when you create a database of the author's text - the reference signature, its features (rounded, oblique, Gothic, etc.) determine the number of steps of crushing of the text (3 in Fig.1, in which the elements are linearized text path passing through the unit cell. As an example, the implementation of this step for one piece of text (uppercase letterconsider the simplest splitting the input lines, an example of which is shown in Fig. 2. On a drawing of a fragment of handwriting (signature) received by the scanning or other known method, is applied to the bounding mesh, dividing the text into separate unit cells, each unit cell (cell) of the grid contains a curvilinear trajectory elements of handwriting (Fig. 2A). Then the grid �Malchut as long while the curvilinear trajectory elements of text within the unit cell becomes almost linear (Fig. 2B). This fragmentation bounding mesh creates a template fragment with number of rowsand the number of columns. When you create a database fragment of the reference text, depending on its features, determine the number of steps of crushing and dimension of the template text fragment (the number of rows and number of columns). The template fragment as a set of unit cells through which the trajectory of handwriting (Fig.3b), can be represented in the form of adjacency matrix of the graph, the non-zero elements which define the unit cell through which the trajectory of the text. Active pixels, amenable to analysis, will be the point of intersection of the boundaries of the unit cell of the pattern and text paths passing through the cell of the template fragment. For each unit cell determine the magnitudeand(see enlarged image of the cell in Fig. 3A), and their attitudeis the tangent of the angle of inclination of the trajectory of the text, which is used as an identification parameter. Such statement by pixel, the set of points of the trajectory of the fragment signatures for�changing quite foreseeable, final setaverage angle of the line pattern of the trajectories of text on a selected fragment of the figure of the text. The bounding mesh, which is inscribed with the text with a default template (Fig.3b), and each piece of text is a set of cell - elements of the template, on the borders of which are active pixels. Passphrase (author signed) for the database is recorded repeatedly (more than 20-30) that allows you to give an interval estimate for each element of the template asby calculating the mathematical expectation of the angle of trajectory and standard deviation. To save memory when creating a database to calculate these statistics will use the following formula:

Here:vector identification parameters;

row, column cell of the template;

room implementation record when you create a database.

Thus, instead of pixel by pixel comparison, requiring a huge amount of time to create the database, will have a limited amount of templates for each fragment in the form of an identification matrix (Fig.3b), to a non-zero I�akah which will be the identification parameters of the reference text (characteristics grouping and scatterThe obtained representation of a fragment of handwriting allows you to compare signatures between them.

The authentication step is researched on the authorship of the signature is scaled and divided into individual pieces (6 in Fig.1), similar to the breakdown of the reference signature. For each fragment of the studied signatures sequentially one after the other can be loaded with the appropriate template fragment reference signature from the database (7 in Fig.1 and Fig.4), which is subjected to affine transformation, and compression (tension) in the two coordinate directions is subjected to a first pattern in General, further compression (tension) is subjected to a separate element of the template separately, so that in every element of the template has got the appropriate fragment studied the signature. Fig.4 shows conversion options template reference standard in the analysis of the studied fragment of the text. Fig.4A the standard template text is deformed with the rotation with the purpose of combining with a fragment of the studied text. Fig.4B with the same purpose for another fragment of the studied text held nonuniform tensile (compression) horizontally. Allowable tension (compression) in the horizontal and vertical center of the unit cell pattern of the Etalon is determined in accordance with the characteristic scatter

When combining the standard template and the segment calculates the values of angles of inclination of the trajectory of the text (9 in Fig.1) and their comparison with the average value of angles of the reference text (10 in Fig.1). In case of deformation of the template reference signature for a complete matching of significant cells (not zero) (Fig. 4A) mandatory operation is the operation of the reverse rotation, which is performed by using two-dimensional affine transformation, which adjusted the angle of the trajectory of the studied text on each relevant cell [4].

When authenticating the identity of the expert-the criminalist makes handwritten text in the form of a signature or password to the scanner, the data on the figure of the text displayed on the monitor screen. The entered image is divided into fragments. The expert scales the signature, it can also change the position of the studied signatures, focusing on the position of the image on the monitor screen. On the stage of the research required signatures following steps are performed. From a pre-created database, which store the standards of handwriting in the form of an identification matrix containing the average tilt angles of the trajectories �Kapinovo text loaded the standards of handwriting in the form of an identification matrix. The study shall be presented for signature comparison consistently fragment by fragment.

Here a particular important role belongs to the expert, the monitor selects the standard text matching (similar) on the studied text. The entered image is divided into fragments similar to the procedure for creating master records (6 in Fig.1). Further there is a hanging template reference signature (7 in Fig.1) perform an affine transformation of the template pattern (scaling and rotation) to align the template reference with the trajectory of the studied handwriting (8 in Fig.1). Affine transformation of the template enables to adequately combine it meaningful (non-zero cells) with the path pattern of the studied text, for further comparisons of tilt angles. When combining the standard template presented with a fragment of handwriting calculate the angles of inclination of the trajectory of the text in each unit cell of the template and form a matrix of angles (9, 10 of Fig.1). The analyzed signature (11 in Fig.1) checked for authenticity in the following sequence:

- match the number of templates (identity matrix);

- match the dimensions of the templates (identity matrix).

When nesovpaden�and the number of templates, and the dimensions of the standard for at least one fragment of the studied signatures, the comparison stops and is taken from the database the following sample reference signature. When you run the above two conditions on the studied signature hinged template reference signature. In that case, when the pattern of the reference signature is placed entirely on the charged fragment of a signature on the image of the pattern stand out of the line of inclination of the text and calculate their angle of inclination, in accordance with the fragmentation of the image determine the value of the angle of the lines for each cell of the fragment. Then, the comparison of the values of the angles of the lines for each cell of the fragment is taken as an identity settings (11 in Fig. 1). Two possible cases:

- identification parameter of the investigated text for each unit cell template reference range;

- identification parameter of the investigated text for each unit cell template reference range;

- identification parameter of the investigated text for each unit cell template reference range;

- identification parameter of the investigated text for each unit cell template reference not pop�gives in the interval .

When deciding comparing the measure of closeness of the parameters against the specimen signature (password) with the reference. While there can be three cases of comparison:

1. If the format of the template of the reference signature (the dimension of the identity matrix,at least one fragment does not match the format of the studied signatures, we investigated the signature is not the author;

2. If the results of the affine transformation of the template of the standard scaling of the studied signatures do not allow to approach the study signature to the template of the reference signature, at least one fragment, we investigated the signature is not the author;

3. If the conditions under paragraph 1 and 2 and the average angles of inclination of the trajectory lines of text for each elementary cell of the template for each fragment are within the range of variation of the reference signature, then the study is the author's signature.

Assuming that the average values of the inclination angles of the trajectories of handwriting belonging to the entire set of fragments (reference recording, which is close to the normal law distribution, with a high degree of confidence it can be argued that the implementation of the provisions of clauses 1 and 2, the study is the author's signature. This allows you to divide the required signatures on �grades according to the degree of certainty. Detailed identification procedure can end positively if the presented pattern signatures coincide with the location of its features with the reference samples of handwriting. Otherwise, the user is prompted to repeat a sample of the signature.

The main advantage of the proposed method compared to known biometric methods of authentication is as follows:

- to identify the signature does not require expensive devices - card readers, biometric information;

the method facilitates the creation of the submission (template) handwriting, with which computers can be made to preliminary forensic analysis of handwriting, reduces the time of expert handwriting analysis and increases the reliability of the classification and identification;

- the method can precede and complement other means of identification used in investigative practice.

Literature

1. S. M. Gusakova, A. S. Komarov. Intelligent system to solve the identification problem in handwriting. Artificial intelligence and decision making, No. 4/2010, pp. 49-54.

2. Boll Ores, etc. Guide to biometrics. /Boll Ores, 'connell Jonathan X., Pankanti Sharath, Ratha, Nalini K., Sir Andrew W. //Moscow: Technosphere, 2007. - 368 p. (first paragraph on p. 48).

3. STANDARDS A. B. Basic methods�s, used for handwriting recognition, initial release 24.04.2002 found 18.02.2008 on the Internet at http://www.recognition.mccme.ru/pub/RecognitionLab.html/methods.html

4. R. Hartshorn. Foundations of projective geometry. M.: Mir, 1970. - 161 p.

1. Method of identification by handwriting, including the digitization of the detected and reference texts, their division into separate fragments and combining each of the studied fragments with the corresponding reference, the comparison and determination of the coincidence of the detected and reference samples of the text, characterized in that after repeated scanning of the reference handwriting is exercised by the division of the text into separate fragments, allocate to each example of a fragment of a handwritten text line text path, then crushed a drawing of a piece of text on the unit cell so that within a unit cell receive straight text path, then for all implementations of the standard text on each fragment for each unit of its cell calculates the average value and standard deviation of the slope angle, the calculated values form the identity matrix, the dimension of which is equal to the dimension of the fragment, and non-zero values which contain the values of the average tilt angles and the root-mean-square deviation�tions, remember these values of the average orientation angles and root-mean-square deviations and use them for identification in her handwriting; in the identification process presented handwriting exercise split scanned picture studied the handwritten text into fragments, their scale, the allocation for each example of a fragment of a handwritten text line text path, the fragmentation pattern of the text fragment in the unit cell is similar to the fragmentation of the fragment of the reference text, the calculation for each elementary cell of a fragment of the values of the angle of the text, the formation by the calculated values of the angles of inclination of the matrix of angles, the dimension of which is equal to the dimension of the identity matrix fragment reference, comparing these values of the angle of the text with the corresponding values of the average tilt angles of the fragment of the reference text from the database, and in the case of rotation of the template fragment reference signature for a complete matching of significant template cells adjust the angle of the trajectory of the studied text on each relevant cell within the angle of rotation of the template, while in the case of coincidence of the calculated values of angles of inclination with average tilt angles of the fragment of the reference text with a given probability of the studied text �read author, and in case of discrepancy between the calculated and reference values, the identification procedure is terminated.

2. A method according to claim 1, characterized in that allows you to categorize imposed handwriting on the degree of authenticity with reference to classes of coincidence with the reference text.



 

Same patents:

FIELD: medicine.

SUBSTANCE: invention relates to field of microscopic examination of tissue and cells. Method contains stages at which: object, containing biological material and including multitude of adjacent layers and peculiarity of interest, extending through said multitude of adjacent layers are provides; picture of section of the first of said multitude of adjacent layers, cut from subject is provided; picture of section of the second of said multitude of adjacent layers, cut from object, is generated; delay between cutting sections of the first and the second adjacent layers of said multitude of adjacent layers is selected from any of hours, days, weeks, months, years; area of interest is determined on picture of the second layer section on the basis of area of interest on picture of the first layer section, area of interest is determined in section of the second layer on the basis of area of interest on picture of the second layer section and material is extracted from area of interest in the second layer section.

EFFECT: increased accuracy of extraction of material from object in the field of biology, histology or pathology.

16 cl, 11 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to means of processing three-dimensional images. The method includes selecting a primary volumetric section and a secondary volumetric section which is adjacent to the primary volumetric imaging section; determining one or more z-axis alignment parameters (310); determining one or more x,y-axis alignment parameters (314); applying (316) the one or more z-axis parameters (310) and the one or more x,y- axis parameters (314) to shift the position of the secondary volumetric section to align it with the primary volumetric section.

EFFECT: shorter time for forming final images when aligning volumetric sections of image data.

14 cl, 16 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to forming a 3D model of vessels of a region of interest of an object. The system comprises at least one image data acquisition device; a data processing unit; a display device; and an interface unit; wherein the image data acquisition device is configured to acquire image data of a region of interest of an object; wherein the processing unit is configured to determine probability values for predefined features in the image data for each picture element; determine the most accurately corresponding predetermined features based on denoted features; calculate object data based on the determined features; and use the calculated object data for additional processes; wherein the display device is configured to display probability values for each picture element of the image data for interaction purposes; wherein the interface unit is configured to denote features in the displayed probability values via user interaction.

EFFECT: high accuracy of forming a 3D model of vessels of a region of interest of an object.

14 cl, 11 dwg

FIELD: medicine.

SUBSTANCE: invention refers to diagnostic visualisation tools. A nidus detection system comprises a segmentation unit of an anatomical first presented image of the region, a detection unit of high tracer accumulation according to the functional second presented image, a classification unit of high tracer accumulation in accordance with a position with regard to the anatomical structures, an accumulation detection unit, which analyses the segmented regions to identify normal and abnormal regions, an attenuation unit of high tracer accumulation on the functional second presented image on the basis of the results obtained by the classification unit, the above attenuated regions correspond to the anatomical structures, which are identified as normal; there are also provided an identification unit of the high accumulation as one of the potential pathological change, and a regulation unit configured to compare metabolic activity of the non-attenuated high-intensity regions to metabolic activity of the regions identified by the accumulation detection unit as normal. A diagnostic visualisation system for implementing a method for diagnostic visualisation comprises an anatomical image scanner, a visualisation scanner by positron tomography and said nidus detection system.

EFFECT: using invention enables the qualitative assessment of pathological changes and reducing the time of identification.

15 cl, 5 dwg

FIELD: agriculture.

SUBSTANCE: invention relates to animal breeding, and in particular, to devices and methods for determining body condition state (BCS) of animals. Additional technical results are providing automatic, effective, quick, safe, easy to use and low-cost determining the quantitative indicator of body condition of the animal. The device for determining the quantitative indicator of the body condition of the animal (50) comprises a system (51) of a three-dimensional camera and a device (52) of image processing. The system (51) of the three-dimensional camera is directed at the animal and is provided for recording three-dimensional image of the animal. The device (52) of image processing is connected to the system (51) of the three-dimensional camera and is provided for forming a three-dimensional representation of the surface of the animal part from the three-dimensional image for the statistical analysis of the surface and for determining the quantitative indicator of the body condition state of the animal on the basis of statistically analysed surface.

EFFECT: improved accuracy and reliability, as well as elimination of errors in determining the quantitative indicator of the body condition of animal.

19 cl, 20 dwg, 1 tbl

FIELD: physics, video.

SUBSTANCE: invention relates to camera control means. The method comprises acquiring first information used to control a first area specified within a full image captured by a camera unit; acquiring second information used to control a second area specified within the full image; controlling mechanical movement of the camera unit based on the first information input; acquiring an image of the first area from the full image captured by the camera unit and extracting an image of the second area from the first area based on the second information.

EFFECT: high range of the obtained image.

18 cl, 20 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to image processing devices and methods. An image processing device is disclosed. The device comprises a collection unit configured to collect image data showing an image which includes a given object. The device also comprises a forming unit configured to form a contour signal showing part of a contour included in the image. The device also comprises a detecting unit configured to detect, based on the contour signal, a characteristic direction of the contour for each of a plurality of divided regions obtained by dividing the image. The device also includes a determining unit configured to determine the type of object, and a correction unit configured to correct image data according to a correction method corresponding to the type of object.

EFFECT: high accuracy of determining contour detection in an object.

10 cl, 44 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to means of analysing perfusion images. The system comprises a subsystem (7) for identifying part of the outermost area surrounding a central damage area which is the area of interest, a subsystem (1) for calculating a plurality of frequency distributions of intensity values of the area of interest on images; a perfusion information extractor (2) for extracting information relating to perfusion from the plurality of frequency distributions.

EFFECT: high accuracy of extraction of perfusion information from an image.

13 cl, 2 dwg

Perfusion imaging // 2541126

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to rendering perfusions. The method comprises steps of: executing, by a data analyser, computer-executable instructions which select, without interacting with the user, a processing protocol from an electronic protocol storage based on imaging data corresponding to a patient; processing, by the data analyser, functional imaging data for a subject using the selected processing protocol in a first processing mode, wherein the selected image data processing protocol includes at least two processing steps; and performing, using a computer processor, validation of the processed data when executing the selected processing protocol; changing, using the computer processor, the processing mode from the first processing mode to a second processing mode based on the validation, wherein the data analyser is configured to process functional imaging data in the second processing mode.

EFFECT: reduced interaction with the user and faster processing of perfusion imaging data.

14 cl, 3 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to organ segmentation. The method comprises steps of: selecting (210), by a user, a surface model of an organ; selecting (220), by the user, a plurality of points on the surface of an image of the organ; and transforming (230-290) the surface model into a plurality points on the image in order to change the outline of the surface model for more accurate conformity to the outline of the organ, wherein the transformation includes a step of interpolating the plurality of points to determine intermediate points between the selected plurality of points on the image of the organ, and wherein the interpolation includes a step of determining corresponding intermediate points on the surface model.

EFFECT: high accuracy of organ segmentation.

15 cl, 2 dwg

FIELD: information technologies.

SUBSTANCE: method of biometric authentication of users consists in that a database of identification parameters of handwriting examples of users with access to an IT system is created in advance, then the user entering the access control system is suggested to make any writing of handwriting example, the received analogue signal is converted to digital form, forming an array of quantized counts, by values of its elements an array of coefficients is calculated, the elements of which are compared with the respective array elements of the registered users, and the user to be identified is deemed incidental to the reference record if this difference is minimum. After the analogue signal is converted into a digital form, an analogue signal jitter, the ratio of number of values of the analogue signal jitter exceeding the first threshold value to the total number of values of the analogue signal jitter are determined, the calculated ratio is compared with the second threshold value, in case the calculated relation doesn't exceed the second threshold value, the array of coefficients is calculated by means of two-dimensional discrete cosine conversion, otherwise the handwriting example is written again.

EFFECT: reduction of probability of error of the first sort of user authentication when the user accessed the system, whose handwriting example parameters are available in the access control system database, is denied by the access control system.

2 dwg

FIELD: information technology.

SUBSTANCE: authentication device is used to store biometric data of a monitored individual by sensing a sequence from predefined handwritten words written by the monitored individual on a graphics pad. Each word is presented to the monitored individual before the individual repeats the word and the time allocated for the individual to write said word is limited. Further the biometric authentication device is trained to recognise said handwritten words and fragments thereof written by the monitored individual. Data input and training are carried out until first- and second-type error probability forecast falls to a given value for a predefined number of words. In authentication mode, verification words are presented to the monitored individual in a random sequence and the individual handwrites each verification word just presented recently. The obtained biometric data are then compared with reference data and an evaluation is made of the similarity of said words. The sequence of verification words presented to the individual is increased until a decision is made on the authentication result with the probability of first- and second-type errors, close to a predefined value.

EFFECT: high reliability of automated monitoring of enforcement of house arrest and avoiding the need to use hand and leg bracelets.

4 cl, 1 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to computer engineering and can be used in video analysis and processing systems. The apparatus for automated facial recognition when processing a group photograph comprises a unit for searching for regions containing faces, a decision unit, wherein the output of a unit for localising facial regions (1) is connected to the input of a unit for separating the image into a background and an object (2), the output of which is connected to the first input of an image normalisation unit (3); the output of a control unit (9) is connected to the second input of the image normalisation unit (3), the output of which is connected to the input of a unit for generating vectors of facial features (4), the output of which is connected to the input of a buffer unit (5), the output of which is connected to the first input of a comparator unit (6); the output of a delay unit (7) is connected to the input of a unit for generating vectors of facial features (8), the output of which is connected to the second input of the comparator unit (6), the output of which is the data output of the apparatus.

EFFECT: automated facial recognition when processing a group photograph.

1 dwg

FIELD: physics.

SUBSTANCE: device comprises P units for selecting a maximum signal, P units for calculating an activation function, P groups of multiplier units, P groups of decoders and P groups of units for generating membership function values.

EFFECT: high accuracy of recognition when recognising objects in conditions without or with strong distortion of separate areas of an image, broader functional capabilities.

1 dwg

FIELD: physics, computer engineering.

SUBSTANCE: invention relates to a method of identifying and classifying an object. The method comprises the following steps: detecting an object using at least one physical detector disposed on said object; using the output signal of the detector, by forming a threshold value, and an analysing device to detect at least one object defined as a singly-connected region having defined physical properties which uniquely distinguish the object from other objects; identifying and/or classifying the object from the output signal based on predetermined properties; using the output signal for the object to deduce multiple different physical features; based on the selected physical features, associating the object with at least one of N predetermined base classes; ordering the N base classes in a predetermined sequence into an N-dimensional vector V, which is associated with the object, wherein elements v1…vN of the vector V indicate the identity of the object to the corresponding base class; depending on the vector V, associating the object with an arbitrary class selected from a reference database, wherein if the object belongs to the corresponding base class, the vector element v1…vN is assigned a binary value "1", otherwise the binary value "0" is assigned.

EFFECT: faster identification and classification of objects by predefining N base classes ordered by an N-dimensional vector V.

2 cl, 2 tbl

FIELD: physics.

SUBSTANCE: database of users allowed to access a system is formed from sampled and quantised readings of handwriting samples of a matrix of two-dimensional discrete Haar transform coefficients. When allowing a new user into the system, a matrix of two-dimensional discrete Haar transform coefficients similar to the reference sample is formed from the sampled and quantised readings of handwriting samples of that user. Comparison is carried out and a decision is made to associate the obtained record to one of the reference records.

EFFECT: high reliability of authorised access in systems requiring use of personal identification means.

2 cl, 4 dwg

FIELD: information technology.

SUBSTANCE: device comprises the pretreatment units of first and second images, the recording units of first and second images, the conversion units of first and second images into a color space YIQ, the enhancing units of the real component of first and second images, the image forming units as a result of rotation of the first and second image, the units of image forming in changing the angle of inclination of the first and second images, the units of storage of simulated images for the first and second images, the unit of application of the method SIFT, the calculation unit of quantity of equal descriptors, the unit of storage of the found pair of duplicates.

EFFECT: ensuring the ability to compare the descriptors applied to the task of searching image duplicates.

5 dwg

FIELD: physics.

SUBSTANCE: device additionally includes a register of criteria codes, a unit of memory of criteria codes, a decoder of criteria codes and a unit of result memory.

EFFECT: increased efficiency of a device due to reduced quantity of requested criteria of recognition for instances, when the result becomes available in advance by the current situation of recognition.

5 dwg, 2 tbl

FIELD: radio engineering, communication.

SUBSTANCE: image recognition device contains a multichannel switch, an ADC, a marker associativity coefficient memory unit, a logical AND unit, a shift register unit, a control unit, a logical OR element, an address register, an address selection memory unit, a buffer register.

EFFECT: device performance improvement.

4 dwg, 4 tbl

FIELD: textiles, paper.

SUBSTANCE: invention relates to a paper sheet processing device. A paper sheet processing device in accordance with the present invention comprises: a unit of image formation, made with the ability to capture an image of a paper sheet and to generate an image of a paper sheet; an identification unit made with the ability to identify symbol of each digital position included in the serial number from the serial number area of the image of the paper sheet; an output unit made with the ability to output an image of the part corresponding to the serial number part of the paper sheet image, when there is a digital position which symbol can not be identified by the identification unit; a display unit made with the ability to display each symbol identified by the identification unit, and the image output from the output unit; and the input unit made with the ability to receive input of each symbol corresponding to the digital position which symbol can not be identified by the identification unit.

EFFECT: improving performance in identification of the symbol.

13 cl, 12 dwg

FIELD: automated recognition of symbols.

SUBSTANCE: method includes following stages: tuning, forming symbols models, recognition, recording background model together with background of read image, separating model of registered background from elementary image of background, combining for each position of symbol of model of letters and/or digits with elementary displaying of appropriate background, forming of combined models, comparison of unknown symbols to combined models, recognition of each unknown symbol as appropriate symbol, combined model of which is combined with it best in accordance to "template comparison" technology.

EFFECT: higher efficiency.

10 cl, 10 dwg

Up!