WO2024091140A1 - Способ проведения аутентификации - Google Patents
Способ проведения аутентификации Download PDFInfo
- Publication number
- WO2024091140A1 WO2024091140A1 PCT/RU2023/000024 RU2023000024W WO2024091140A1 WO 2024091140 A1 WO2024091140 A1 WO 2024091140A1 RU 2023000024 W RU2023000024 W RU 2023000024W WO 2024091140 A1 WO2024091140 A1 WO 2024091140A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- user
- specified
- template
- key parameters
- trajectory
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Ceased
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/36—User authentication by graphic or iconic representation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/30—Authentication, i.e. establishing the identity or authorisation of security principals
- G06F21/31—User authentication
- G06F21/316—User authentication by observing the pattern of computer usage, e.g. typical user behaviour
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2221/00—Indexing scheme relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/21—Indexing scheme relating to G06F21/00 and subgroups addressing additional information or applications relating to security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F2221/2133—Verifying human interaction, e.g., Captcha
Definitions
- the invention relates to the field of computer technology, namely, to methods for authenticating users using graphic images and can be used, for example, as a CAPTCHA test (Completely Automated Public Turing test to tell Computers and Humans Apart - a fully automated public Turing test for distinguishing computers and people) or to unlock personal devices, equipment, doors, etc.
- CAPTCHA test Completely Automated Public Turing test to tell Computers and Humans Apart - a fully automated public Turing test for distinguishing computers and people
- CAPTCHAs have a number of flaws that allow bots to pass verification like a human.
- biometric technologies that are widespread today to solve this problem and develop specialized software is being explored.
- a method of implementing a CAPTCHA test is known from the prior art, which is based on showing the user simple text characters with noise and determining the reproduction of these characters by the user using known proximity metric algorithms (see patent US8978144B2, check. G06F21/30, publ. 03/10/2015).
- known proximity metric algorithms see patent US8978144B2, check. G06F21/30, publ. 03/10/2015.
- dynamic effects are added during demonstration (letters are shown one by one, particles are removed randomly, etc.).
- the main disadvantages of the known method are the complexity of implementation and the lack of reliability of the test, due to the relative simplicity of computer hacking of the displayed images.
- a method of implementing a CAPTCHA test is known from the prior art, according to which the user is shown a template of a graphic image and is asked to designate this image with one or more concepts, while authentication is carried out using the registered time of solving the test by the user (see patent US8752141 B2, cl. G06F7/04, published 10.06.2014).
- the main disadvantages of the known method are the inconvenience of its use and the lack of reliability due to the relative simplicity of computer hacking using modern image recognition algorithms.
- a method of user authentication is known from the prior art, according to which the Cartesian coordinates of the points of the line drawn by the user in the process of drawing the image and the drawing time are recorded, after which the set of key parameters of the generated image is compared with the corresponding set of key parameters of the reference image previously created by the authenticated user (see. patent US1 1238149B2, class G06F21/31, published 02/01/2022).
- the known method can be used in a system for unlocking portable computing devices.
- the main disadvantage of the known method is the insufficient reliability of distinguishing between computers and people.
- a method of authentication is known from the prior art, according to which a graphic image template is formed by combining several predefined elements, shown to the user, a set of coordinates of the image generated by the user is registered, a sufficient number of points are selected for comparison, they are compared with the corresponding set of template coordinates and a successful signal is issued.
- authentication if their difference does not fall outside the confidence range (see patent US9471767 B2, class G06F21/36, published 10/18/2016).
- the disadvantage of this known method is that it is estimated only how far the points of the drawn and demonstrated trajectory are located from each other friend, and in terms of drawing dynamics, it is only taken into account that the time it takes to generate an image for bots is longer than the drawing time for a human.
- the main disadvantage of the known method is its limited capabilities due to the small amount of registered data, which greatly simplifies hacking an authentication system that works using this method.
- the closest in technical essence to the proposed invention is a method of authentication, according to which the following steps are implemented: templates of graphic images are formed, one of the specified templates is shown to the user, the user is asked to depict the demonstrated template, a set of key parameters of the image generated by the user is registered, a set of key parameters of the generated image is compared user images with a corresponding set of key template parameters and issue a signal of successful authentication if their difference does not fall outside the confidence range (see patent US10657243B2, class G06F 21/36, published 05.19.2020). A randomly generated reference trajectory is used as an image template, and during the comparison it is determined how the characteristic (reference) points are located from each other on the trajectory.
- the disadvantage of this known method is the relatively high probability of hacking, since successful authentication is determined only by which object was drawn and in what total time.
- the technical problem is to eliminate the above shortcomings.
- the authentication method includes the following stages: (i) graphic image templates are generated, (ii) one of the specified templates is shown to the user and the user is prompted to depict the demonstrated template, (iii) a set of key parameters of the generated user of the image, (iv) compare the set of key parameters of the image generated by the user with the corresponding set of key parameters of the template and issue a signal of successful authentication if their difference does not fall outside the confidence range, and include in the specified set of key parameters, according to at least one time parameter characterizing the dynamics of the appearance of the trajectory along which the user generated the image, and at stage (i) for each template, the specified set of key parameters and the limits of the confidence range are determined by machine learning based on test images that were manually generated earlier authenticated users when demonstrating this template.
- the specified time parameter can be the average speed on the trajectory, the projection of acceleration onto the initial part of the trajectory, the average number of points on the trajectory for selected time intervals, time intervals during which the user formed selected parts of the image, the average and/or standard deviation of the length of time intervals between selected trajectory points.
- the specified set of key parameters may additionally include the area under the trajectory in projection onto the selected coordinate axis and the length of the trajectory.
- the specified confidence range limits may include an upper threshold value and a lower threshold value, and a successful authentication signal in step (iv) may be issued if the specified difference of key parameters of the user-generated image and the template is between the specified upper threshold value and the lower threshold value.
- the template is preferably shown with noise and/or animation.
- the successful authentication signal is preferably used as a CAPTCHA test pass signal or as an unlock signal.
- Figure 1 is a screenshot before starting step (ii) showing the graphic template
- Fig. 2 is a screenshot of step (ii) of displaying a graphic template on a mobile phone screen
- FIG. 3 shows a screenshot of step (ii) demonstrating the graphic image template on the screen of a tablet or personal computer
- Fig. 4 is a block diagram of the algorithm for demonstrating a graphic image template with noise, displaying it to the user and processing the data
- Fig. 5 is a block diagram of the authentication algorithm according to the proposed method. DETAILED DESCRIPTION OF THE INVENTION
- the user is asked to repeat the graphic image template in the form of a figure-drawing, and the information collected when constructing the image will be used for authentication.
- the proposed invention is based on the fact that the most difficult from the point of view of imitation (spoofing) for the purpose of hacking to ensure unauthorized access is the recreation of behavioral characteristics, in particular, the manner of drawing (biometric handwriting).
- the main advantage of biometrics is its uniqueness and multi-parameter nature, so parameters associated with human activity are almost impossible to repeat or fake.
- the proposed technical solution proposes to use during authentication not only the vectorized characteristics of the generated image, but also the vectorized characteristics of the image drawing process itself, obtained through the use of convolutional deep neural networks (such networks can be pre-trained on a huge volume of any images and then further trained on a replenished base data).
- the proposed authentication method includes the following main steps.
- Initial graphic templates may be loaded into the authentication system from an external source or proposed by previously authenticated users and are preferably simple monochrome objects (drawings).
- a set of key parameters is formed using machine learning based on test images, which is attached to the specified template.
- the test images used are images that were manually generated by previously authenticated users when demonstrating this template.
- both open platforms and users of the Yandex.Toloka platform can be used.
- the database can be replenished with both ready-made images and those offered by new users, as well as through the automatic generation of a template with further retraining of the neural network after showing it to real users and collecting information about a set of key parameters.
- the set of key parameters includes at least one time parameter characterizing the dynamics of the appearance of the trajectory along which the user formed the image (i.e., the drawing style).
- the specified time parameter may be:
- the set of key parameters may include non-dynamic parameters, such as
- the characteristics of objects are saved in vectorized form (the input is complexly structured “raw” object data, and the output is vector representations of the characteristics in the form of a set of key parameters).
- the vector data is obtained by running the input data through a pre-trained deep neural network (PDNN) using Deep Learning technology.
- PDNN pre-trained deep neural network
- the POGNS is initially trained by a human by manually labeling pre-assembled objects into similar classes, after which the POGNS is stored for direct vectorization of the input data.
- the specified confidence range limits may include both an upper threshold (when the generated image parameters are too far from the template - for example, more than 30% difference) and a lower threshold value (when the generated image parameters, on the contrary, are too similar to the template and were likely copied by computer hacking - for example, less than 3% difference).
- IP address, etc. can be analyzed for comparison with possible use by the robot (use of additional open and commercial databases).
- Step (ii) showing one of the specified templates to the user.
- one of the available templates is selected and shown to the user on the device screen (see Figs. 1, 2, 3).
- the template is shown with noise, i.e., the overlay of random noise (see Fig. 4), and to simplify subsequent analysis, it is shown in the form of an animation showing which trajectory the proposed template should be drawn along.
- the user is asked to depict (repeat) the demonstrated pattern using the touch screen of a mobile device, laptop (touchpad) or computer mouse.
- Stage (iii) registers a set of key parameters of the user-generated image.
- the system dynamically records the process of the emergence of the trajectory (coordinates of points on the touch screen and the astronomical or relative time of their appearance) along which the user draws the image, and transmits the received data to a remote server for subsequent processing and identification of the values of key parameters.
- Step (iv) compares and completes authentication.
- the set of key parameters of the user-generated image is compared with the corresponding set of key parameters of the template.
- the preferred option is to use the cosine distance as a metric of proximity between sets of key parameters (vectors).
- a successful authentication signal is issued if the difference between these parameters does not fall outside the confidence range (i.e., is between the upper and lower threshold values). Otherwise, the procedure is repeated or an error notification is issued in accordance with the algorithm in Fig. 4.
- the successful authentication signal can be used as a signal for passing a CAPTCHA test (for example, to provide access to a website) or as a signal for unlocking personal devices, equipment, doors, etc. (see figure 5).
- the proposed method can significantly increase the reliability of authentication both when used as a CAPTCHA test and when used as an unlock key.
- Testing of the proposed method was carried out using a deep machine learning algorithm like a neural network with more than 300 layers and 1 million parameters. In total, more than 21,000 biometric samples were collected during testing. On average, each user repeated 30 shapes from a set of more than 700 shapes and generated samples of biometric handwriting for each image with coordinate-time characteristics for training deep neural networks. Datasets were collected using JavaScript in the form of drawing images, as well as a list of X, Y coordinates and time every few milliseconds.
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Computer Hardware Design (AREA)
- Physics & Mathematics (AREA)
- Software Systems (AREA)
- Data Mining & Analysis (AREA)
- Social Psychology (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Collating Specific Patterns (AREA)
- Lock And Its Accessories (AREA)
Abstract
Description
Claims
Priority Applications (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| EP23771754.1A EP4610854A1 (en) | 2022-10-26 | 2023-02-02 | Authentication method |
| US18/282,092 US20250036740A1 (en) | 2022-10-26 | 2023-02-02 | Method of Authentication |
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| RU2022127829A RU2815478C1 (ru) | 2022-10-26 | Способ проведения аутентификации | |
| RU2022127829 | 2022-10-26 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2024091140A1 true WO2024091140A1 (ru) | 2024-05-02 |
Family
ID=90831523
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/RU2023/000024 Ceased WO2024091140A1 (ru) | 2022-10-26 | 2023-02-02 | Способ проведения аутентификации |
Country Status (3)
| Country | Link |
|---|---|
| US (1) | US20250036740A1 (ru) |
| EP (1) | EP4610854A1 (ru) |
| WO (1) | WO2024091140A1 (ru) |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2003123560A (ru) * | 2003-07-24 | 2005-02-27 | ООО "Крейф" (RU) | Способ и устройство идентификации пользователя по подписи |
| US20080092245A1 (en) * | 2006-09-15 | 2008-04-17 | Agent Science Technologies, Inc. | Multi-touch device behaviormetric user authentication and dynamic usability system |
| US20100115610A1 (en) * | 2008-11-05 | 2010-05-06 | Xerox Corporation | Method and system for providing authentication through aggregate analysis of behavioral and time patterns |
| US20130243242A1 (en) * | 2012-03-16 | 2013-09-19 | Pixart Imaging Incorporation | User identification system and method for identifying user |
| US8752141B2 (en) | 2008-06-27 | 2014-06-10 | John Nicholas | Methods for presenting and determining the efficacy of progressive pictorial and motion-based CAPTCHAs |
| US8978144B2 (en) | 2010-06-22 | 2015-03-10 | Microsoft Corporation | Automatic construction of human interaction proof engines |
| US9471767B2 (en) | 2014-08-22 | 2016-10-18 | Oracle International Corporation | CAPTCHA techniques utilizing traceable images |
| KR20180048121A (ko) * | 2016-11-02 | 2018-05-10 | 충남대학교산학협력단 | 사용자의 키보드 및 마우스 입력 행위 패턴을 이용한 인증 방법 및 그 방법을 구현하는 프로그램을 기록한 기록매체 |
| US10657243B2 (en) | 2017-03-02 | 2020-05-19 | Jingxia YUAN | Variation analysis-based public turing test to tell computers and humans apart |
| US11238149B2 (en) | 2019-01-28 | 2022-02-01 | Joseph Carlo Pastrana | Computerized user authentication method that utilizes the Cartesian coordinate system to verify a user's identity |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US8776173B2 (en) * | 2011-03-24 | 2014-07-08 | AYAH, Inc. | Method for generating a human likeness score |
-
2023
- 2023-02-02 WO PCT/RU2023/000024 patent/WO2024091140A1/ru not_active Ceased
- 2023-02-02 US US18/282,092 patent/US20250036740A1/en active Pending
- 2023-02-02 EP EP23771754.1A patent/EP4610854A1/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2003123560A (ru) * | 2003-07-24 | 2005-02-27 | ООО "Крейф" (RU) | Способ и устройство идентификации пользователя по подписи |
| US20080092245A1 (en) * | 2006-09-15 | 2008-04-17 | Agent Science Technologies, Inc. | Multi-touch device behaviormetric user authentication and dynamic usability system |
| US8752141B2 (en) | 2008-06-27 | 2014-06-10 | John Nicholas | Methods for presenting and determining the efficacy of progressive pictorial and motion-based CAPTCHAs |
| US20100115610A1 (en) * | 2008-11-05 | 2010-05-06 | Xerox Corporation | Method and system for providing authentication through aggregate analysis of behavioral and time patterns |
| US8978144B2 (en) | 2010-06-22 | 2015-03-10 | Microsoft Corporation | Automatic construction of human interaction proof engines |
| US20130243242A1 (en) * | 2012-03-16 | 2013-09-19 | Pixart Imaging Incorporation | User identification system and method for identifying user |
| US9471767B2 (en) | 2014-08-22 | 2016-10-18 | Oracle International Corporation | CAPTCHA techniques utilizing traceable images |
| KR20180048121A (ko) * | 2016-11-02 | 2018-05-10 | 충남대학교산학협력단 | 사용자의 키보드 및 마우스 입력 행위 패턴을 이용한 인증 방법 및 그 방법을 구현하는 프로그램을 기록한 기록매체 |
| US10657243B2 (en) | 2017-03-02 | 2020-05-19 | Jingxia YUAN | Variation analysis-based public turing test to tell computers and humans apart |
| US11238149B2 (en) | 2019-01-28 | 2022-02-01 | Joseph Carlo Pastrana | Computerized user authentication method that utilizes the Cartesian coordinate system to verify a user's identity |
Non-Patent Citations (2)
| Title |
|---|
| MARAKHTANOV ALEXEY; PARENCHENKOV EVGENY; SMIRNOV NIKOLAI: "Mouse Dynamics Analysis Using Machine Learning to Prevent Account Stealing in Web Systems", 2022 31ST CONFERENCE OF OPEN INNOVATIONS ASSOCIATION (FRUCT), FRUCT OY, 27 April 2022 (2022-04-27), pages 167 - 173, XP034120970, DOI: 10.23919/FRUCT54823.2022.9770926 * |
| STRINGHAM, EDWARD P.: "Private Governance: Creating Order in Economic and Social Life", 2015, OXFORD UNIVERSITY PRESS, pages: 105 |
Also Published As
| Publication number | Publication date |
|---|---|
| EP4610854A1 (en) | 2025-09-03 |
| EP4610854A8 (en) | 2025-11-12 |
| US20250036740A1 (en) | 2025-01-30 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11736478B2 (en) | Device, system, and method of user authentication based on user-specific characteristics of task performance | |
| US10877568B2 (en) | Three-dimensional in-the-air finger motion based user login framework for gesture interface | |
| US10586036B2 (en) | System, device, and method of recovery and resetting of user authentication factor | |
| Fierrez et al. | Benchmarking touchscreen biometrics for mobile authentication | |
| Thomas et al. | A broad review on non-intrusive active user authentication in biometrics | |
| Tolosana et al. | BioTouchPass2: Touchscreen password biometrics using time-aligned recurrent neural networks | |
| US10164985B2 (en) | Device, system, and method of recovery and resetting of user authentication factor | |
| US10579784B2 (en) | System, device, and method of secure utilization of fingerprints for user authentication | |
| US10476873B2 (en) | Device, system, and method of password-less user authentication and password-less detection of user identity | |
| US9571490B2 (en) | Method and system for distinguishing humans from machines | |
| US9202035B1 (en) | User authentication based on biometric handwriting aspects of a handwritten code | |
| JP2016526210A5 (ru) | ||
| JP2016526210A (ja) | 通信デバイスを用いた個人の動作ベース識別認証 | |
| Hashia et al. | On using mouse movements as a biometric | |
| Sun et al. | A 3‐D hand gesture signature based biometric authentication system for smartphones | |
| CN114730371A (zh) | 使用已登记生物特征数据集检测基于生物特征的认证系统中的敌对实例 | |
| Wang et al. | Improving reliability: User authentication on smartphones using keystroke biometrics | |
| Sharma et al. | Behavioral Biometrics: Past, Present | |
| Lu et al. | Global feature analysis and comparative evaluation of freestyle in-air-handwriting passcode for user authentication | |
| RU2815478C1 (ru) | Способ проведения аутентификации | |
| Wang et al. | Towards DTW-based unlock scheme using handwritten graphics on smartphones | |
| Rahman et al. | Web user authentication using chosen word keystroke dynamics | |
| Saini et al. | Authenticating mobile phone user using keystroke dynamics | |
| WO2024091140A1 (ru) | Способ проведения аутентификации | |
| EA045306B1 (ru) | Способ проведения аутентификации |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| WWE | Wipo information: entry into national phase |
Ref document number: 18282092 Country of ref document: US |
|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 23771754 Country of ref document: EP Kind code of ref document: A1 |
|
| WWE | Wipo information: entry into national phase |
Ref document number: 2023771754 Country of ref document: EP |
|
| NENP | Non-entry into the national phase |
Ref country code: DE |
|
| ENP | Entry into the national phase |
Ref document number: 2023771754 Country of ref document: EP Effective date: 20250526 |
|
| WWP | Wipo information: published in national office |
Ref document number: 2023771754 Country of ref document: EP |