AUTHENTICATION SYSTEM BASED ON A TELECOMMUNICATIONS DEVICE WITH A BRAIN ACTIVITY SENSOR
Patent Information
- Application Number
- DE502021005482
- Authority / Receiving Office
- DE · DE
- Patent Type
- Patents
- Current Assignee / Owner
- Priority Date
- 2020-04-02
- Filing Date
- 2021-04-01
- Publication Date
- 2024-11-07
- Estimated Expiration
- 2041-04-01
AI Technical Summary
Existing authentication methods require individuals to physically visit authorities or organizations, incurring effort and costs, and lack a secure remote authentication solution that is both efficient and secure.
A method and system utilizing a KLHA sensor to detect brain activity signals in response to familiar and foreign images displayed on a portable device, allowing for secure remote authentication by comparing the similarity of captured brain activity signals to authenticate individuals without the need for physical presence.
Enables secure, efficient, and user-friendly remote authentication, protecting against unauthorized access and ensuring privacy by not requiring storage of individual biometric characteristics, thus reducing the need for physical visits and minimizing data storage risks.
Description
Area
[0001] The invention relates to a method and system for remotely authenticating a person. background
[0002] On many occasions, individuals must reliably authenticate themselves to an authority or service. Examples include applying for an ID card renewal, applying for a passport, proving their identity to obtain a bank loan, and similar situations.
[0003] The need to appear in person at the premises of a government agency or private sector company is burdensome and costly for both the person who must appear and the agency or company before which the person presenting their ID must appear. Many governments are therefore attempting to digitize a growing portion of such processes ("electronic government"), and the private sector is also striving to make processes as efficient as possible at all levels.
[0004] Sometimes it is not possible to appear in person (e.g. due to physical limitations) or it would be disproportionately time-consuming (e.g. due to a stay abroad).
[0005] A secure remote authentication procedure that is sufficiently secure would therefore be beneficial for citizens and authorities, as well as consumers and service providers.
[0006] A well-known scientific publication is TOUYAMA HIDEAKI ET AL: "Non-target photo images in oddball paradigm improve EEG-based personal identification rates," ENGINEERING IN MEDICINE AND BIOLOGY SOCIETY (EMBC), 2013 34TH ANNUAL INTERNATIONAL CONFERENCE OF THE IEEE, August 25, 2008, pages 4118-4121, ISSN: 1557-170X, DOI: 10.1109 / IEMBS.2008.4650115. It describes a method for identifying test subjects using EEG measurements. The test subjects are shown various images while undergoing an EEG measurement. The test subjects are asked to focus particularly on images of a pre-arranged type. Depending on whether a test person focuses on an image or not, different EEG measurement curves result, which allow for improved identification of the test persons. Summary
[0007] The invention is based on the object of proposing an improved authentication method and a corresponding authentication system.
[0008] The objects underlying the invention are achieved with the features of the independent patent claims. Embodiments of the invention are specified in the dependent claims. The embodiments listed below can be freely combined with one another, provided they are not mutually exclusive.
[0009] In one aspect, the invention relates to a method for authenticating a person as specified in claim 1. The method comprises: Registration of a person with a server application, wherein the server application is instantiated on a server computer system, wherein the registration comprises capturing at least one known image of the one person within secure premises when the one person is present in the secure premises, wherein a known image is an image with an image motif that is known to the one person; reception of the at least one known image from the server application via a network connection by a client application, wherein the client application is instantiated on a portable telecommunications device that is assigned to the one person, wherein the telecommunications device is reversibly or irreversibly connected to a KLHA sensor, wherein the KLHA sensor is a contactless sensor for detecting brain activity;Generating a specification of a sequence of known images and foreign images by an authentication module, wherein a foreign image is an image with a motif that is unknown to the person, wherein the authentication module is part of the server application or the client application; Displaying multiple images on a display of the telecommunications device by the client application outside the secure space, wherein the displayed images include a sequence of one or more known images and one or more foreign images according to the specification;while the plurality of images of the person are displayed, detecting first and second brain activity signals, which are respectively triggered in the one person by viewing the images, by the KLHA sensor, wherein first brain activity signals are detected while a familiar image is displayed to the person, and wherein second brain activity signals are detected while an unfamiliar image is displayed to the one person; authenticating the one person as a function of the similarity of the detected first and second brain activity signals, taking into account the specification of the image sequence by the authentication module; and outputting an authentication result with respect to the one person. ;
[0010] The method for authenticating a person treats the person as successfully authenticated if the first brain activity signals are similar to each other, if the second brain activity signals are similar to each other, if the first brain activity signals are dissimilar to the second brain activity signals, and if the sequence of mutually similar first brain activity signals and mutually similar second brain activity signals corresponds to the sequence of the displayed familiar images and unfamiliar images.
[0011] Embodiments of the invention may have the advantage that a person can authenticate and register with an authority or service once in a secure manner. On this occasion, one or more familiar images of this person can be captured. For example, one or more facial images of this person can be taken with a camera within the secure premises, allowing the authority or service tothe other owner of the secure premises can assume with certainty that the facial images captured within the secure premises actually originate from the person who is currently in the secure premises and who has authenticated themselves during registration, for example by presenting appropriate documents (depending on the type of registration process, authentication in the secure premises can be carried out, for example, on the basis of biometric features, identity documents, birth certificates, registration certificates or other electronic or paper-based documents). As a result of the registration, a user profile can be created for the person and stored in a user register managed by the server application.
[0012] Authentication as a function of the similarity of the acquired first and second brain activity signals includes an analysis of the similarity of the acquired first and second brain activity signals.
[0013] The one or more acquaintance images of this person can be stored linked to this person's user profile, because the facial image is only a "acquaintance image" for this person and a "third-party image" for other people. The one or more acquaintance images captured during registration can be stored, for example, in a database accessible by the server application ("server-side storage"). Additionally or alternatively, the one or more acquaintance images captured during registration can also be stored in a local data storage of a telecommunications device assigned to this person. This also ensures that the acquaintance images are stored explicitly or implicitly linked to the profile and / or a device specifically assigned to the person.
[0014] The portable telecommunications device could be a smartphone, for example. If this person now wants to authenticate with the authority again after a few years without visiting the authority's premises, they can do so using the portable telecommunications device's client application.
[0015] For example, during the re-authentication of a person years after the original registration, a random image sequence can be generated by the client application or the server application, and the images of the image sequence can be displayed on the display of the person's telecommunications device. The image sequence can be, for example, a mixture of the person's known images captured during the registration process and several third-party images. Furthermore, the person can connect the KLHA sensor to the telecommunications device for the purpose of authentication. For example, the person can request the KLHA sensor for this purpose from the authority or the relevant authenticating authority and receive it by mail. According to some embodiments, however, it is also possible for the telecommunications device to include the KLHA sensor from the outset (for example, in the form of a smartphone camera).
[0016] While the client application displays this image sequence on the telecommunications device's display, the KLHA sensor captures brain activity signals and forwards them to the client application. The client application evaluates the received brain signals to determine whether the person currently viewing the display finds the subjects in the displayed familiar images familiar, or at least significantly more familiar, than the subjects in the displayed unrelated images, and whether the brain activity signals are significantly more similar when displaying images of the same type (either familiar or unrelated) than when displaying images of different types. Only if the analysis of the brain activity signals shows that the expected pattern of first and second brain activity signals corresponds to the displayed sequence of unrelated and familiar images is a person considered authenticated.
[0017] Embodiments of the invention can therefore have the advantage that a hardware extension for a telecommunications device can be used to securely authenticate a person to a government agency or other organization and to provide proof of identity for that person. It is not necessary for the person to re-enter a secure premises of the government agency or other organization for this purpose.
[0018] Embodiments of the invention may also be advantageous in that the remote authentication method reliably protects against a person other than the one who originally registered successfully authenticating remotely. For example, if the telecommunications device were to fall into the hands of another person, and this other person attempted to authenticate using the telecommunications device, for example, to apply for a new passport using their own facial image or to apply for a loan that would be paid into their own account, the authentication attempt would fail because the brain activity signals of this other, unauthorized person would not differ significantly when viewing the familiar images of the registered / authorized person from the brain activity signals measured when viewing other people's images.This is because the familiar images of the authorized person represent foreign images for the unauthorized person and the brain activity signals are generated automatically by the brain of the respective person without the person having any conscious influence on them.
[0019] Embodiments of the invention can therefore have the advantage of enabling a forgery-proof remote authentication process that frees both the person wishing to authenticate and the relevant authorities from the need for the person wishing to authenticate to visit specific authority premises. A further advantage can be the increased user-friendliness of the authentication process. For example, there is no need to remember a password.
[0020] Compared to other biometric authentication methods known from the prior art, embodiments of the invention can offer several advantages: The method is even more secure against an attacker attempting to falsely authenticate themselves as another person, because the method is not necessarily based on comparing an observed characteristic with a characteristic stored in a database, but rather on the direct comparison of brain activity signals generated in response to self-images or images of others. Furthermore, the method does not require the storage of individual person characteristics in a database, thus protecting the privacy of the person to be authenticated. This allows the method to be used even in situations where storing personal, biometric data is undesirable.By deleting the data, the privacy of the authenticating person can also be improved ("improved privacy").
[0021] According to embodiments, the method for authenticating a person is a method for remotely authenticating a person. Thus, while authenticating using this method, the person is not located within the secure premises used for registration. For example, while the authentication process is being performed, the person may be in their own home, in a hospital, abroad, or any other location where a network connection can be established between the client application on the portable telecommunications device and the server application on the server computer system. The network connection may, for example, be a mobile phone connection.
[0022] According to embodiments of the invention, the client application receives at least one acquaintance image during the registration of the person.
[0023] For example, the process of transmitting the identity photos captured during a person's registration to the client application of that person's telecommunications device can be performed once during or after the registration process, but not for each subsequent authentication process. This can be advantageous because it reduces data traffic over the network and also reduces the risk of unauthorized third parties reading the identity photos during transmission.
[0024] According to one embodiment, after receiving the at least one known image, the client application stores the at least one received known image in a protected manner in a non-volatile memory of the telecommunications device. For example, the storage can be in encrypted form and / or can be in a memory area to which only the processor of the telecommunications device has access, but not an external device. After successfully receiving and successfully storing the at least one known image in the local, non-volatile memory of the telecommunications device, the client application sends an acknowledgment of receipt regarding the at least one known image to the server application. In response to receiving the acknowledgment of receipt, the server application deletes all other existing copies of the known image.
[0025] Such embodiments can offer the advantage of increased data security or data protection, because at least one known image of a person is only stored on the server for a short time. The server-side storage only occurs until it can be transferred to a local storage device of the person's telecommunications device. After that, this potentially very sensitive personal data is deleted from the server. In the event that unauthorized third parties should intrude into the server computer system, the risk of obtaining known images of a specific person is therefore very low, or only a few people would be affected (namely, those who have just registered and whose known images have not yet been transferred to a telecommunications device of the respective person).
[0026] According to further embodiments, the client application receives the at least one known image during each authentication process for the individual. Each authentication process comprises: receipt of an authentication request for the individual from the client application by the server application; in response to receipt of the authentication request, identification of the at least one known image captured during the registration of the individual by the server application. For example, the step of identifying the at least one known image for the individual can comprise a search within an image database for one or more known images stored in association with a user profile of that individual.The image database can, for example, contain a plurality of facial images of persons registered with the server application. However, the facial images of other persons do not represent familiar images for that person, but can at most be used as third-party images in some embodiments. The authentication process also includes the server application sending the identified at least one familiar image to the client application via the network connection.
[0027] The fact that at least one known image is obtained anew from the server application during each authentication process can have the advantage that the client application uses less storage space on a non-volatile storage medium of the telecommunications device.
[0028] According to embodiments of the invention, the method comprises storing the at least one acquaintance image captured during the registration of a person in a protected manner (e.g., after encryption of the at least one acquaintance image by the server application) in a non-volatile memory of the server computer system by the server application. The at least one received acquaintance image is stored by the client application in a memory of the telecommunications device exclusively temporarily for the duration of the authentication process.
[0029] It is often sufficient for the identification images required for authentication to be stored temporarily in a volatile data storage device, such as the main memory, and only for the duration of the authentication process. Corresponding embodiments of the invention can have the advantage that in the event of loss of the telecommunications device, the identification images are not lost, since they are retrieved from the server application for each authentication process anyway. Thus, the identification images are better protected against loss. In the event of loss of the telecommunications device, the user does not have to re-register or appear in person at the authority again to have the identification images recorded again.
[0030] According to embodiments, the client application is configured to automatically delete the acquaintance images of a person stored in the volatile memory after the authentication result has been output.
[0031] Such embodiments can have the advantage that the acquaintance images are even more securely protected against loss or access by unauthorized persons.
[0032] According to embodiments of the invention, the at least one known image is a facial image of the one person.
[0033] This can have the advantage that it is self-evident for the authority or organization registering the person that the respective person's own facial image is known. In some embodiments of the invention, it may also be possible to capture familiar images that depict a subject other than a person's face, for example a pet, objects in the home or workplace, etc. However, these other subjects have the disadvantage that it is often not easy for the authenticating authority (e.g., authority) to determine whether a specific subject is definitely and reliably known or very familiar to a specific person. However, this can be assumed for the person's own facial image.
[0034] According to embodiments of the invention, the at least one known image is an image of an image motif captured by a camera or video camera that is demonstrably known to the person.
[0035] According to other embodiments of the invention, the at least one known image is a plurality of images captured by a camera or video camera of an image motif that is demonstrably known to a person. For example, the face of a person can be captured multiple times by a camera during the registration process, for example, from different angles and perspectives and / or using different light sources. Thus, a plurality of slightly different facial images of the same person are captured, each depicting or representing the same face slightly differently.
[0036] According to further embodiments of the invention, the at least one known image is a plurality of copies of an image captured by a camera or video camera of an image motif that is demonstrably known to the person. For example, a single facial image of the person can be captured during the registration process. This can be stored individually or as multiple identical copies. It is also possible for the copies of the one known image to be generated dynamically for a specific authentication attempt by the person.
[0037] The above-mentioned variants for generating multiple known images of a person during the registration process and / or only during the authentication process can be advantageous, as the security and reliability of the authentication process increases with the number of self-images and other images shown. For example, this can increase the length and complexity of the image sequence according to which the known images and other images of a person are displayed. The more images shown or the more brain activity signals recorded, the less individual interference signals can influence the overall result of the analysis of the brain activity signals and lead to false positive or false negative results.
[0038] According to embodiments of the invention, the registration of the one person further comprises: Providing at least one biometric sensor within the secure space, wherein the sensor is particularly designed as a sensor that can distinguish between people wearing face masks and people not wearing face masks; for example, thermal imaging cameras and / or sensors for the conductivity of facial skin can be used to determine whether the person currently registering is wearing a mask or not; detecting a biometric feature of the person by the biometric sensor; evaluating the detected biometric feature; if the evaluation of the detected biometric feature shows that the person is not wearing a mask, carrying out the registration and providing the at least one known image from the server application to the client application;If the evaluation of the captured biometric feature shows that the person is wearing a mask, abort the registration without capturing the at least one known image and / or without providing the at least one known image to the client application.
[0039] Corresponding embodiments of the invention can have the advantage that they can prevent a person from manipulating the registration process by using at least one known image, typically the facial image of the person registering, to represent the face of another person. For example, a person who already possesses a specific authorization card, e.g., an employee ID card with a photo, could authenticate themselves as an employee using a fingerprint and claim to have lost this card in order to have a new employee ID card issued. The person could have the goal of using this new employee ID card to gain access to the company premises for another person and, for this purpose, wear a face mask that reproduces the face of this other person. To prevent this, further security measures can be taken within the secure premises, e.g.,the aforementioned biometric sensors are used.
[0040] According to embodiments of the invention, the at least one known image captured in the secure space is captured by a security camera.
[0041] The security camera can, for example, be a camera with access protection. Access protection ensures that the camera can only be used by an operator after a prior authentication step with the security camera. Access protection can, for example, be password-based, based on a biometric feature, or other means that only allow access after an authentication process with the camera.
[0042] Additionally or alternatively, the security camera can be a camera with a signature key configured to sign the known images it captures. The server application has a signature verification key configured to verify the signature of images and, during the registration process, only accept those images as known images that have been verified to be signed by the security camera.
[0043] Such embodiments can have the advantage of further increasing security. For example, the secure premises within which the identity photo is captured may not only be a government agency premises, where employees and / or biometric sensors ensure that the identity photo actually depicts the person currently present in the secure premises for the purpose of registration. However, it is also possible that certain service providers, such as certain photo studios, take facial images or other identity photos of individuals for registration purposes on behalf of the authorities or with official approval, and capture the identity photo on their own premises (e.g., a photo studio). In this case, the photo studio is considered a secure premises.In such application scenarios, the risk that an identifiable image, for example a facial image of a person, is not authentic (i.e., depicting a person wearing a mask and / or morphed), is considerably higher. However, the risk that the identifiable image supplied by a photo studio is manipulated / inauthentic can be reduced by requiring the photo studio to use a security camera to capture identifiable images for security-critical registration processes. The authority can determine whether a security camera was used, for example, by looking at the signature on the identifiable images. Thus, in some cases, the capture of identifiable images can be delegated to service providers whose secure premises do not offer the same level of security as corresponding premises of public authorities, without significantly increasing the risk of tampering with the identifiable images.
[0044] For other applications (e.g. private sector applications) it is also possible for the registration process to take place in an environment in which a second person confirms the correctness of the registration process (so-called collaborative registration / enrollment).
[0045] According to embodiments of the invention, the authentication module (and thus the authentication method) treats the person as successfully authenticated if the first brain activity signals differ significantly from the second brain activity signals.
[0046] According to the invention, the authentication module (and thus the authentication method) treats the person as successfully authenticated if the first brain activity signals are similar to each other, if the second brain activity signals are similar to each other, if the first brain activity signals are dissimilar to the second brain activity signals and if the sequence of mutually similar first brain activity signals and mutually similar second brain activity signals corresponds to the sequence of the displayed familiar images and unfamiliar images.
[0047] Embodiments of the invention can have the advantage that the analysis of the first and second brain activity signals during the authentication process takes into account the relative nature of the first brain signals compared to the second brain signals, and preferably not a specific absolute signal or signal profile. The acquisition, storage, and evaluation of absolute, person-specific reference signals for the first and second brain activity signals is thus not required. The problem of such personal biometric reference data being lost can thus be avoided.
[0048] Furthermore, embodiments of the invention may have the advantage that it is not necessary for all persons to generate substantially identical first brain activity signals when viewing familiar images, nor for all persons to generate substantially identical second brain activity signals when viewing unfamiliar images. Rather, it may be sufficient to establish that viewing multiple familiar images reproducibly results in the generation of first brain activity signals that are significantly similar to one another, and that viewing multiple unfamiliar images also reproducibly results in the generation of second brain activity signals that are significantly similar to one another, whereas each of the first brain activity signals differs significantly from each of the second brain activity signals.
[0049] According to one embodiment of the invention, the displayed images contain a plurality of identified known images and a plurality of identified unidentified images. The authentication method treats the person as successfully authenticated if the first brain activity signals differ significantly from the second brain activity signals in a manner that is reproducible across a minimum number of first and second brain activity signals.
[0050] Embodiments of the invention may have the advantage that displaying multiple known images and multiple unknown images allows an authentication module to check whether the first brain activity signals differ from the second brain activity signals in a reproducible manner. If only a single first brain activity signal were recorded and compared with a single second brain activity signal, there would be an increased risk that the second authentication method would produce a false positive or false negative result due to measurement variances.
[0051] According to embodiments of the invention, the KLHA sensor includes a signature generated by a KLHA signature key. The client application includes a KLHA signature verification key corresponding to the KLHA signature key and is configured to permit the acquisition of the first and second brain activity signals and / or their use for authentication only if a verification of the KLHA sensor's signature with the KLHA signature verification key reveals that the KLHA sensor signature is valid.
[0052] According to embodiments of the invention, the images are displayed on a digital display of the telecommunications device. The use of a digital display has the advantage of greater device simplicity and the fact that such a display is already present on every smartphone, tablet computer, and notebook in use today. Digital displays in the form of, for example, LCD, LED, or OLED screens are available in all possible sizes and shapes.
[0053] According to embodiments, the authentication module is configured to determine the variability of various first brain activity signals and / or the variability of various second brain activity signals during the second authentication process. If the variability of the first and / or second brain activity signals exceeds a predetermined limit, the authentication module aborts the authentication process, preferably issuing an error message. The person has thus not authenticated themselves during the authentication process. This can be advantageous if, due to various interference signals or other unfavorable circumstances, the acquisition of brain activity signals is only successful with a quality that precludes reliable authentication.Typically, a severe disruption of the brain activity signals detected by the sensor would result in the person's authentication attempt being rejected during the authentication process, and the person would be considered "not authorized." In some cases, this can have negative consequences for the person, for example, if the authentication module locks the person's account after a certain number of failed authentication attempts or permanently deactivates the remote authentication option. According to embodiments of the invention, aborting the authentication process due to high signal variance does not lead to negative consequences for the authenticating person, since the reason for the authentication failure is rooted in the high signal variance.
[0054] According to embodiments of the invention, the authentication module generates a random sequence of one or more known images and one or more unknown images for each authentication process of the person. The specification of the image sequence is generated to reflect this random sequence.
[0055] The multiple images are displayed in such a way that the chronological sequence of the displayed images corresponds to this random order. The analysis of the recorded first and second brain activity signals is carried out taking the chronological sequence into account.
[0056] For example, the authentication module can be configured to display a sequence of six facial images to each person attempting to authenticate. The six facial images can, for example, always include a fixed number of known images and foreign images, for example, always two known images and four foreign images. However, it is also possible for the relative proportion of foreign images and known images to be randomly selected for each authentication procedure, as long as it is ensured that the multiple displayed images contain at least one known image, preferably several known images, and at least one foreign image, preferably several foreign images.
[0057] According to embodiments of the invention, the sequence of familiar images and unfamiliar images serves as a challenge in the sense of a challenge-response process for authenticating a person. The analysis of the recorded brain activity signals includes a check to determine whether the temporal order of the first and second brain activity signals reflects the order of familiar images and unfamiliar images encoded in the challenge.
[0058] Embodiments of the invention can have the advantage that particularly effective protection against replay attacks can be achieved. Furthermore, disadvantages associated with classic biometric authentication methods can be avoided and the advantages of biometric methods can be utilized at the same time. Classic biometric authentication data that has been recorded and used as a reference signal cannot simply be reset, which poses a problem if it is stolen. Embodiments of the invention, on the other hand, use a specific sequence of brain activity signals for authentication, whereby not one-time, static biometric characteristics of a person are used, but a dynamically generated challenge-response method in which the authentication module is dynamic and individual for each authentication attempt (e.g.The sequence of the displayed known and unfamiliar images is determined (using a random generator). If the known images of a person stored in a database are stolen, so that unauthorized third parties may turn them into "known images" through frequent viewing, the images can simply be replaced. Authentication is based on an analysis of the changes in certain brain activity signals that differ significantly between "known images" and "unfamiliar images." It is not absolutely necessary for a brainwave reference signal "characteristic" of a person to be recorded once and stored as a reference.
[0059] According to embodiments of the invention, the authentication module only treats the person as successfully authenticated if, in addition to a significant difference between the first and second brain activity signals, the following additional criteria apply: all first brain activity signals recorded during the display of one of the familiar images are identical or sufficiently similar to each other; all second brain activity signals recorded during the display of one of the unfamiliar images are identical or sufficiently similar to each other.
[0060] An advantage of the authentication method according to embodiments of the invention may be that it is not necessary for the first brain activity signals in different persons to have identical or nearly identical properties. The applicant has observed that even if the first brain activity signals triggered by viewing familiar images differ significantly from person to person, they are at least very significantly different from the second brain activity signals generated in that person when viewing unfamiliar images. Provided that the person to be authenticated is shown a sufficiently large number of unfamiliar images and familiar images, an analysis of the temporal sequence of the properties of the brain activity signals during the display time may be sufficient for a correct authentication decision.For example, at least two, preferably at least three known images and at least two, preferably at least five foreign images can be displayed to the person to be authenticated. The person's brain activity signals are continuously recorded while the various known images and foreign images are displayed and compared with the known temporal sequence of the display of the foreign images and known images. If it turns out that the first brain activity signals recorded while the person was viewing the known images all have identical or very similar properties, and that the second brain activity signals recorded while the person was viewing the foreign images all have identical or very similar properties, wherein the properties of the first and second brain activity signals are significantly different, the person is considered to have been successfully authenticated according to embodiments of the invention.
[0061] According to some embodiments, the first brain activity signals have properties that are the same or very similar for all people, and the second brain activity signals also have properties that are the same or very similar for all people, with the first and second brain activity signals of all people being significantly different. In this case, the analysis of the acquired first and second brain activity signals during the authentication process may also include a comparison of the acquired first and second brain activity signals with stored first and second brain activity reference signals.
[0062] According to one embodiment of the invention, the KLHA sensor is integrated into the portable telecommunications device by the manufacturer in a reversible or irreversible manner. For example, the portable telecommunications device can be a smartphone, with the smartphone camera being used as the KLHA sensor. The portable telecommunications device can be a standard smartphone, with the camera either permanently installed in the telecommunications device or installed in a replaceable manner. In some embodiments, the portable telecommunications device is a smartphone of a specific type or series, with the cameras of this smartphone type or series being protected against replacement by another camera in a mechanical, cryptographic, and / or other way.
[0063] According to other embodiments, the KLHA sensor is reversibly connected to the telecommunications device. For example, the KLHA sensor can be an electrical sensor that measures electrical brain activity signals without contact. Such sensors are currently not a common component of conventional portable telecommunications devices. However, it is possible to communicatively and / or mechanically couple electrical sensors to the portable telecommunications device via various interfaces, e.g., microUSB or Bluetooth, so that the measured values acquired by the KLHA sensor are transmitted via the communicative coupling to the telecommunications device and the client application instantiated thereon. To stabilize the connection, a holder, e.g.,a plastic-based holder may be used, which has a first recess which mechanically encloses the telecommunication device similar to a conventional mobile phone case, and which has a further recess to hold the KLHA sensor and hold it at a defined position relative to the telecommunication device.
[0064] For example, a person who wishes to remotely authenticate themselves with a government agency in order to obtain a new identity card with a current passport photo, but who cannot visit the agency's premises in person due to health restrictions, can submit a request to send the agency an "authentication system." The request can be made electronically, e.g., by email or web form, or by mail, specifying the mailing address of the person to whom the authentication system is to be sent. The agency then sends the authentication system to the specified address. The system can, for example, comprise the KLHA sensor and optionally also the aforementioned holder and / or cable for reversibly communicatively and / or mechanically connecting the KLHA sensor to the telecommunications device. The person can use the system as described herein for embodiments of the invention.For example, after a person has been successfully authenticated to the authentication module, the client application can be configured to capture one or more facial images of the person using the camera of the telecommunications device and to send the captured facial images to the server computer system operated by the authority with the server application via a cellular connection or another network connection, e.g. WLAN. Before or after transmission, a software function of the client application and / or the server application can test the quality of the facial images and reject the images and / or request the person to capture the facial images again if the quality is not sufficient. The server application can then forward the facial images to software for personalizing identification documents. The person can return the authentication system to a postal address of the authority after use.This means that a new ID card can be created for the person with a current facial image, which can be safely assumed to show the face of the person who has already registered with the authority in the past under a specific name and at a specific address.
[0065] Depending on the design, the brain activity signals captured by the KLHA sensor can be transmitted to the client application of the telecommunications device either wired (e.g., via a circuit board, a microUSB cable) or via a wireless interface (e.g., Bluetooth). From there, the signals are transmitted to the authentication module, which can be part of either the client application or the server application.
[0066] The authentication module can be configured to correlate the brain activity signals received by the KLHA sensor with additional measurement data. For example, the camera of the telecommunications device can detect the position and / or orientation of a person's head in the room and forward it to the authentication module. Changes in the position and / or orientation of the head can thus be correlated with signal changes. This can be helpful because, according to embodiments of the invention, the authentication module uses this additional measurement data and the detected correlations to filter out interference signals generated by movement of the head in the room during authentication, so that only or predominantly those measurement signals remain that are causally attributable to the viewing of external images and familiar images or to the observation of an image change.Sensor vibrations, such as those generated by the person's walking and turning movements in the room during the authentication process and transmitted to the sensor via mechanical coupling, can also be automatically detected with some sensor types, e.g., using the electrical sensor used as a KLHA sensor or another electrical, gyroscopic, or optical sensor in the telecommunications device. Additionally or alternatively, the camera of the telecommunications device can capture a series of facial images of the authenticating person over a period of time and forward these facial images to the authentication module.By analyzing the temporal series of facial images, the authentication module can, for example, detect whether and when a person blinks and / or whether and when the person experiences a clear emotion (joy, fear, shock, uncertainty), which can be identified, for example, by their facial expression. Brain activity signals associated with such emotions can overlay the brain activity signals generated by viewing the familiar and unfamiliar images and lead to errors. According to embodiments of the invention, the authentication module also automatically detects brain activity signals associated with strong emotions as interference signals and filters them out.
[0067] According to embodiments of the invention, the authentication module has access to an image database in which external images and known images are stored. The external images and known images are stored in the image database in such a way that at least the known images and optionally also the external images are each stored linked to a user profile. In contrast to the known images, the external images can also be stored linked to multiple user profiles. An known image for one user can serve as an external image for other users. Each user profile represents a person who has already registered with the server application and for whom one or more known images were recorded (e.g., during the registration process).
[0068] For example, the image database can contain known images, which are facial images of people who have already successfully registered with the server application and are considered trustworthy. For example, the facial images can be images captured in a secure environment, e.g., on the premises of an authority or a company that carries out the authentication process. This can prevent morphing attacks. Additionally or alternatively, the subjects of the known images can also be physical objects of any kind that are known only to a specific, authorized person to whom the known images in the database are also assigned, whereas these subjects are unknown or at least significantly less known to all or the vast majority of other people.For example, the employees of a company may belong to different work groups or departments, which selectively work at different locations and / or in different buildings or rooms depending on their group or department affiliation. Images that show objects within buildings or rooms in which only a specific group of people works are known images for the people in this group, and simultaneously foreign images for people who do not enter these buildings and rooms (provided the objects shown are not present in both buildings). According to a further embodiment, the known images assigned to a person can also be images of family members, pets, furnishings of the apartment occupied by the person, or of the person's vehicles.Preferably, however, the acquaintance pictures assigned to a person via their user profile are facial images or full-body images of that person (also called "self-image").
[0069] For example, the image database can be a database containing facial images of citizens of a state, employees of a company, or other members of any organization. For example, the person wishing to authenticate can enter their name during the initial authentication process, or the authentication module can identify in another way, such as through automatic facial recognition, which of a large number of registered persons (e.g., employees) wishes to authenticate. Based on this information, the authentication module searches the image database and identifies one or more familiar images stored in the database linked to the entered name or an automatically recognized identifier of this person. The familiar images can, for example, be facial images of this registered person. During the authentication process, it can, for example,It is about a person identifying themselves as the person ("trusted person") as whom they had previously registered with the authentication module.
[0070] The assignment of acquaintance images to personal profiles ("user profiles") can be realized, for example, via a secondary key of a relational database, which links user profiles of registered persons to one or more digital acquaintance images.
[0071] According to embodiments, the authentication module determines one or more third-party images for the person currently authenticating remotely. The third-party images can, for example, depict faces of other people who do not represent the person currently seeking authentication. The other people can be other citizens of the same or another state, or other employees or members of the organization. In this case, a specific facial image can be used by the authentication module both as a known image and as a third-party image, depending on which person currently seeks authentication.
[0072] In other embodiments, the external images stored in the database show people who are not members of the said company or organization and / or who are not registered with the server application, and who can therefore never serve as a known image in an authentication process for a registered person, since they do not depict the face of a registered person. Using facial images of non-registered persons as external images has the advantage of ensuring that the face depicted in an external image is certainly not known to the person authenticating. However, with regard to facial images of employees, a recognition effect in the brain activity signals can be discerned in individual cases, for example, if the face in question is a colleague the employee encounters daily.The use of facial images of persons external to the organization as third-party images can have the advantage that a fairly large number of different facial images are automatically available, e.g. in corresponding commercial image databases.
[0073] According to some embodiments, the external images may also be images depicting physical objects of any kind (people, animals, buildings, landscapes, vehicles, objects, etc.) for which the operator of the authentication module can assume with a sufficiently high degree of probability that no person registered as a trustworthy person with the authentication module and having a corresponding user profile has ever seen these physical objects, or at least has not seen them regularly over a longer period of time (e.g., several weeks, several months, or even several years). For example, the depicted objects may have been generated in a computer simulation, so that this alone rules out the possibility that the authenticating person has ever seen the corresponding motif before.In this case, the known images are preferably images depicting the same type of physical objects (e.g., animals of a certain species, buildings, landscapes, vehicles, objects, etc.), whereby the operator of the authentication module can assume with a sufficiently high degree of probability that these physical objects have been regularly seen by the trusted and registered person to whom the known image is assigned over a longer period of time (e.g., several weeks, several months, or even several years). This is typically always the case with self-portraits.
[0074] Preferably, the type of motif shown in the known images and the foreign images of the person during the authentication process is identical; only the degree of familiarity of the motif shown in the known images and foreign images for the authenticating user differs. For example, in first embodiments of the invention, foreign images are used that show the faces of strangers, and correspondingly, known images are used that are facial images and that are assigned to the user profile of the person wishing to authenticate. For example, in second embodiments of the invention, foreign images are used that show unusual objects such as, for example,Show vases or works of art that are generally unknown and are not located in the work premises where the person seeking authentication regularly works, and use corresponding known images of objects of the same type (vases, works of art, etc.) that are located exclusively in the work premises where said person regularly works. For example, in third embodiments of the invention, use is made of third-party images that show pets such as dogs or cats that are generally unknown (e.g., images of corresponding animals from other countries or distant locations), and use corresponding known images of physical objects of the same type (dogs, cats, etc.) that depict a pet of the person who successfully authenticated themselves in the first authentication method.
[0075] The use of identical image motifs for familiar images and unfamiliar images can have the advantage that differences in the first and second brain activity signals can be attributed exclusively to the different degree of familiarity and not to the difference in the displayed image motifs (face, house, work of art, animal, etc.).
[0076] According to embodiments of the invention, the analysis is performed exclusively on the basis of the first and second brain activity signals and on the basis of the display sequence of the familiar images and foreign images, and optionally additionally on the basis of typed (not individually collected) reference data. This method can be advantageous because it does not require the prior storage of sensitive, biometric, personal data. Consequently, the method can also be used in situations where the storage of personal, biometric data is undesirable. Furthermore, this protects the privacy of the authenticating persons. Compared to other biometric authentication methods known from the prior art, the method is even more secure against attackers because biometric data that does not need to be stored cannot be stolen.
[0077] According to embodiments of the invention, the method further comprises deleting the received brain activity signals after the authentication process. For example, the deletion can include all first and second brain activity signals and any copies created from them and can be carried out, for example, by the authentication software. This embodiment of the invention can be advantageous because it protects the privacy of the person to be authenticated. This allows the method to be used even in situations where storing personal, biometric data is undesirable. Furthermore, the method is even more secure against attackers than other biometric authentication methods known from the prior art, because biometric data that is not stored cannot be stolen. Electrical sensor
[0078] According to embodiments of the invention, the KLHA sensor is an electrical sensor.
[0079] According to embodiments of the invention, the electrical sensor is configured to detect the first and / or second brain activity signals by measuring charge changes and / or charge redistributions. Additionally or alternatively, the electrical sensor is configured to detect the amplitude of the first and / or second brain activity signals and / or the change in amplitude during a time interval.
[0080] In particular, this sensor can be designed to measure charge changes and / or charge redistributions in the vicinity of the sensor. The sensor can be designed, for example, as a sensor that includes an instrumentation amplifier or differential amplifier with high common-mode rejection, wherein the amplifier is coupled to measuring electrodes.
[0081] For example, the electrical sensor can be integrated into a special holder or be designed so that it can be reversibly integrated into such a holder. In addition to a recess or holding elements for the KLHA sensor, this holder also contains a recess and / or holding elements for the telecommunications device. The person can now hold the smartphone in front of their own face, e.g. at a distance of approximately an arm's length, whereby the client application is designed to output the objects currently detected by the camera via the display of the telecommunications device in order to give the person feedback as to whether their own face is currently being detected by the camera (if this is deemed necessary by the client application, e.g. to capture facial expressions for the purpose of noise suppression).Preferably, the person can freely move and rotate their head at least within a certain radius during the authentication process, and it is not necessary for the person to place electrodes on specific areas of the head so that brain activity signals can be recorded.
[0082] For example, the electrical sensor can contain or be coupled to a high-pass filter that rejects low-frequency signal components. For example, the high-pass filter can allow brain activity signals with a frequency of at least one heart rate to pass through unhindered, but attenuate or completely filter out signals with a frequency below one heart rate. This can reduce background noise.
[0083] The amplifier can make it possible to detect brain activity signals with an amplitude of 10 nanovolts or more. An electrical sensor with this level of signal amplification is many times more sensitive than conventional contact-based EEG technology, which can typically only detect brain activity signals with an amplitude of at least a few microvolts.
[0084] According to embodiments of the invention, the authentication module includes a data preprocessing module configured to detect interference signals and minimize them through automatic filtering, so that the first and second brain activity signals are provided as interference-minimized, "denoised" first and second brain activity signals and are further processed by another module of the authentication module. The interference signals can be caused, for example, by vibrations of the KLHA sensor, by the person's heartbeat, by changes in the position and / or orientation of the person's head in space, and / or by blinking movements or strong emotions of the person.
[0085] Examples of suitable high-pass filters and signal amplifiers are described in the international patent application WO 2017 / 189748.
[0086] Various sensors for contactless detection of brain activity signals, in particular electrical sensors, are known in the state of the art, but these have not been used for the authentication of persons.
[0087] For example, international patent application WO 2017 / 189748 describes various types of suitable sensors. The electrical sensor can, for example, be in the form of a contactless electrode, for example, a dry electrode that does not require contact gel and is not in physical contact with the person. Such sensors are also referred to as "biopotential sensors." The size and shape of this electrode can vary greatly depending on the application scenario. For example, the electrode can be in the form of several concentric rings or several parallel rods. Preferably, the electrical sensor is designed as an array of several electrodes, so that brain activity signals from several areas of the head reach different areas of the electrode with varying strengths.This allows the electrical sensor to capture additional information regarding the brain region generating a brain activity signal. For example, this array can contain 5-6 individual electrodes (a small number compared to the 30-60 contact electrodes in a conventional EEG). By using electrical sensors with multiple individual electrodes or a spatial extent of preferably more than 5 cm in two dimensions, the dependence of the signal strength on the distance and orientation of the head of the person to be authenticated can also be captured. This makes it possible, for example, to determine whether the person to be authenticated is actually looking at the display of the telecommunications device on which the external images orFamiliar images are displayed (and not, for example, on a secretly brought self-photo, the sight of which would enable a mask wearer to generate brain activity signals that are typical for viewing a familiar image, even though the officially shown familiar image (mirror image) only shows the mask.
[0088] Martin Oehler's doctoral thesis "Capacitive Electrodes for Measuring Bioelectrical Signals," published in 2009 at the Institute of Electrical Measurement and Fundamentals of Electrical Engineering, and particularly Chapter 3.2, describes further examples of electrical sensors suitable for the contactless recording of brainwave signals (and heartbeat signals), as well as examples of how the amplification and high-pass filter can be designed. An electrical sensor can, for example, comprise one or more electrodes and include the following elements: one or more electrodes, with the electrode surface, which may be insulated from the outside, detecting the actual signal. The electrode(s) is / are connected to the second important part, the electronics, which receives and amplifies the signal with high impedance. The third element of the electrical sensor is the shielding, also called a "filter," which can be either active or passive.
[0089] According to one embodiment of the invention, the function of the capacitive electrode is based on an extremely high-impedance measurement of the capacitively coupled brain activity signal ("biosignal"). The electrode surface and the electrode body form a coupling capacitor. The high-impedance measurement of the input signal results from the input characteristic, which is determined by the coupling capacitor and the input resistance of the circuit. This input behaves like an RC high-pass filter for the coupled signal.
[0090] According to embodiments of the invention, the authentication method comprises positioning the telecommunications device, including the KLHA sensor connected thereto, in spatial proximity to the head of the person to be authenticated. For example, the above-described mount for the telecommunications device and the KLHA sensor can be a mount designed to stand stably on a flat, horizontal surface and to accommodate and hold the telecommunications device in such a way that the display of the telecommunications device can be clearly seen by a person in spatial proximity to this mount or to this KLHA sensor.
[0091] Spatial proximity is preferably understood here as a distance of less than 100 cm, in some embodiments less than 70 cm, and in some embodiments less than 30 cm. These distances are particularly advantageous for electrical sensors; for optical sensors, the distance defining spatial proximity can also be greater, e.g., several meters, provided that the spatial resolution of the optical sensor allows image analysis-based detection of brain activity signals, such as pupil diameter changes.
[0092] The brain activity signals recorded by the electrical sensor can also be referred to as brainwave signals. They are directly caused by electrochemical activity in the brain cells and provide information about conscious and / or unconscious thought processes and brain activity at the time these signals are recorded by the sensor.
[0093] The brain activity signals recorded by an electrical sensor can in particular be signals in the frequency range (from 0.1 to 40 Hz).
[0094] According to embodiments, the electrical sensor is designed to detect the first and / or second brain activity signals by measuring charge changes or charge redistributions in the vicinity of the sensor. Proximity is preferably understood here as an area characterized by a distance of less than 100 cm from the electrical sensor. According to embodiments of the invention, the spatial proximity comprises an area with a distance of 1 cm to 100 cm from the sensor, in some embodiments from 1 cm to 70 cm from the sensor, in other embodiments from 1 cm to 60 cm. In some cases, the sensor is positioned such that it is at a distance of 15-60 cm from the person's head.The exact positioning may depend on the sensor type and / or how the person should move or position themselves to view the display during the secondary authentication process, and how much freedom of movement the person's head should be allowed during the authentication process.
[0095] The amplitude and / or the temporal amplitude profile of the first and / or second brain activity signals can be determined from the recorded signal.
[0096] According to embodiments of the invention, the electrical sensor is designed to detect the amplitude of the first and / or second brain activity signals and / or the change in the amplitude during a time interval.
[0097] According to a first, simple embodiment, the analysis of the recorded first and second brain activity signals comprises determining whether the amplitude of the first brain activity signals, i.e., the magnitude of the charge changes and / or charge redistributions in the vicinity of the sensor measured by the electrical sensor when viewing a familiar image, is higher than the amplitude of the second brain activity signals, i.e., the magnitude of the charge changes and / or charge redistributions in the vicinity of the sensor measured by the electrical sensor when viewing an unfamiliar image. If this is the case, the measured pattern of the first and second brain activity signals corresponds to the expected pattern, and the person is considered successfully authenticated.
[0098] According to a second, more complex embodiment, the analysis of the acquired brain activity signals comprises a comparison of amplitude profiles of the first brain activity signals and of amplitude profiles of the second brain activity signals with each other and / or with amplitude profiles of first and second reference brain activity signals in order to determine whether the first brain activity signals differ significantly from the second brain activity signals.
[0099] According to embodiments, the first brain activity signals are identical or similar to the brain activity signals according to curve 902 in Figure 9 . Additionally or alternatively, the second brain activity signals are identical or similar to the brain activity signals according to curve 904 in Figure 9 .
[0100] The term "identical or similar" can specifically mean "identical within the respective measurement uncertainty of the sensor or so similar that a predefined minimum similarity criterion is met." The predefined minimum similarity criterion can, for example, be a predefined minimum value for a similarity measure (curve distance, integral of the difference area of the brain activity signal profiles over a period of time, etc.).
[0101] This means that for many people, a significant amplification of the signal (amplitude increase of approximately 100% relative to the initial state) can be observed within a period of approximately 100 to 2000 ms, in particular from approximately 300 to 1500 ms, and especially from 500 to 1000 ms after starting to view a familiar image, especially an image depicting the viewer's "own" face ("self-image"), whereas when viewing an unfamiliar image, this increase is absent or the signal strength even decreases within a period of approximately 500-700 ms after starting to view the unfamiliar image.
[0102] According to preferred embodiments, the analysis merely comprises a comparison of the amplitudes and temporal amplitude profiles of the first brain activity signals and second brain activity signals with each other and with the other brain activity signal type, taking into account the known temporal sequence of the display of the one or more familiar images and the one or more unfamiliar images. Brain activity signals recorded during the display of a familiar image are recorded as first brain activity signals. Brain activity signals recorded during the display of an unfamiliar image are recorded as second brain activity signals. If the person is truly the person depicted in the displayed familiar image, it can therefore be expected that all recorded first brain activity signals have identical or very similar properties (amplitude, amplitude curve over time, possibly other signal properties such as frequency).Likewise, it is expected that the recorded second brain activity signals will have identical or at least similar properties, although the properties of the first and second brain activity signals will differ significantly. However, if another person attempts to authenticate themselves as the person who originally registered with the authority, the sequence of foreign and familiar images generated by the authentication module will not generate the expected pattern (challenge) of second and first brain activity signals in that person's brain. This is because the familiar images will also be perceived as foreign images by this unauthorized person and lead to the generation of brain activity signals that appear similar for all images (second brain activity signals even where first brain activity signals were expected).
[0103] In this case of a fraudulent authentication attempt, the first brain activity signals have identical or similar properties to the second brain activity signals. In this case, the authentication module treats the remote authentication of the person as a failure.
[0104] According to embodiments, the analysis of the acquired brain activity signals comprises a comparison of the frequency and amplitude curves of the first brain activity signals and second brain activity signals with each other and / or with first and second reference brain activity signals in order to determine whether the first brain activity signals differ significantly from the second brain activity signals. For example, the authentication module can be configured to calculate a similarity score for a specific first (or second) brain activity signal with respect to a first (or second) reference brain activity signal, e.g., based on geometric distances between corresponding brain activity signal curves in a curve diagram. If the calculated similarity score, which can be calculated, e.g., as an inverse distance score of the compared signal profiles or signals, is below a predefined minimum threshold, the compared signals are considered to be significantly different. Optical sensor
[0105] According to embodiments of the invention, the KLHA sensor is an optical sensor, in particular an IR camera or an IR video camera or a camera or video camera for light in the visible wavelength range.
[0106] In some embodiments, the KLHA sensor is a camera integrated into the telecommunications device by the manufacturer. A camera integrated into the telecommunications device by the manufacturer is an example of a KLHA sensor that is irreversibly attached to the telecommunications device.
[0107] Such embodiments can have the advantage that it is not necessary to procure additional hardware and connect it to the telecommunications device. Rather, the optical sensor used as a KLHA sensor, in the form of a smartphone camera, is already an integral component of most telecommunications devices sold today.
[0108] According to other embodiments, the KLHA sensor is a separate component, for example, an external camera, which is connected to the telecommunications device specifically for the purpose of authentication (and usually in addition to the camera already integrated into the telecommunications device by the manufacturer). For example, it can be an external camera or video camera connected to the telecommunications device via a microUSB interface or a Bluetooth interface. In addition to this data exchange interface, the connection between the KLHA sensor, designed as an external camera, and the telecommunications device can also include a mechanical connection.The mechanical connection can be designed, for example, as a smartphone holder that includes one or more holding means (e.g., recesses, sliders, or gripping arms) designed to automatically mechanically fix a telecommunications device and / or the KLHA sensor. The smartphone holder can have an extension, for example, a 0.5 to 5 cm long stem, with the external camera attached to the end of the extension.
[0109] According to embodiments of the invention, the KLHA sensor configured as an optical sensor is configured to capture images or image sequences of the face or parts of the face of a person containing the first and / or second brain activity signals. The analysis of the first and second brain activity signals by the authentication module comprises an automatic image analysis of the images or image sequences. The image analysis comprises a computational determination of a pupil diameter of the person. Additionally or alternatively, the image analysis comprises a computational determination of a pupil diameter change profile (i.e., a change in diameter over time) of the person.
[0110] If the first authentication method uses a camera to capture biometric characteristics, the same or a different camera can also be used to capture brain activity signals optically.
[0111] According to embodiments of the invention, the method comprises machine learning method steps, e.g., in the course of a training phase, to generate a person-specific or generic classifier which is designed to learn the differences between the first and / or second brain activity signals (e.g., during the registration of a user), and / or in the course of a test or application phase of the trained classifier to distinguish between the first and / or second brain activity signals during an authentication process.
[0112] According to embodiments of the invention, the analysis of the recorded brain activity signals comprises: image processing to detect pupil diameter; and / or correlating properties of the first and second brain activity signals with the times at which familiar and unfamiliar images were displayed.
[0113] According to embodiments, the authentication module determines absolute pupil diameters and thus also differences in pupil diameter when viewing images of different types. The temporal progression of the pupil diameter when viewing a specific image is determined from the image sequences. The amplitude (pupil diameter) and / or the temporal amplitude progression of the first and / or second brain activity signals can be determined from the determined signal. An "amplitude progression" or "amplitude progression over time" is also referred to here as a "profile" or "amplitude profile." The amplitude can refer, for example, to brain activity signal intensities that occur in the signals acquired as raw data. In other embodiments, the amplitude refers to signals that were computationally derived from the raw data, e.g., to amplitudes of signals within a specific frequency range.Limiting the frequency range to selected ranges can have the advantage of reducing interference signals.
[0114] According to embodiments, the method comprises an automatic image analysis of the images or image sequences captured as the brain activity signals to determine a pupil diameter of a person depicted in the image. The determination of the pupil diameter can be performed by the authentication module or by another software program.
[0115] According to a first, simple embodiment, the analysis of the recorded first and second brain activity signals comprises determining whether the amplitude of the first brain activity signals, i.e., the pupil diameter when viewing a familiar image, is higher than the amplitude of the second brain activity signals, i.e., the pupil diameter when viewing an unfamiliar image. If this is the case, the measured pattern of the first and second brain activity signals corresponds to the expected, and the person is considered successfully authenticated.
[0116] According to a second, more complex embodiment, the analysis of the acquired brain activity signals comprises a comparison of amplitude profiles of the first brain activity signals and of amplitude profiles of the second brain activity signals with each other and / or with amplitude profiles of first and second reference brain activity signals in order to determine whether the first brain activity signals differ significantly from the second brain activity signals.
[0117] According to embodiments, the first brain activity signals are identical or similar to the brain activity signals according to curve 952 in Figure 11 . According to embodiments, the second brain activity signals are identical or similar to the brain activity signals according to curve 954 in Figure 11 .
[0118] This means that in many people, a significant increase in pupil diameter (by approximately 0.05 mm - 0.50 mm) can be observed within the first 1500 ms after beginning to view a familiar image, e.g. a "self-image", which is not observed when viewing an unfamiliar image with the same type of motif.
[0119] According to embodiments, the analysis of the recorded brain activity signals includes the implementation of an image processing process. This involves training algorithms using machine learning and deep learning methods that can independently detect differences in pupil diameter during the viewing of the unfamiliar and familiar images and thus perform a classification.
[0120] For example, the optical sensor can be a camera or video camera that takes one or more images of the face or parts of the face of the person to be authenticated in the infrared range or in the wavelength range of visible light and forwards them to the authentication module. The authentication module performs an image analysis, for example to recognize individual components of the eye area of the face, in particular the iris and / or pupil of one or both eyes of the person to be authenticated. By analyzing the position and / or size of a person's pupil in several images taken over time, a pupil diameter change profile over a time interval (e.g., 0-4 seconds after the start of the display of a known image or foreign image, in particular 0-4 seconds) of one or both eyes of this person can be determined by image analysis.
[0121] The applicant has determined that pupil diameter and the change in pupil diameter over time are expressions of brain activity, and in particular, they differ significantly depending on whether a person is shown a familiar image or an unfamiliar image. Thus, by automatically capturing and analyzing images of the eye area of people during a period in which the person is shown one or more familiar images and one or more unfamiliar images, it is possible to determine whether the pupil diameter(s) measured while viewing a familiar image differ significantly from the pupil diameter(s) measured while viewing an unfamiliar image, whereas the corresponding signals measured when viewing the same image type (familiar image or unfamiliar image) are identical or very similar.Thus, the said biometric features "pupil diameter" or "pupil diameter change profile" are suitable for authenticating the person on the basis of the authentication method according to embodiments of the invention.
[0122] The above-mentioned optically recorded brain activity signals provide at least indirect information about brain activity, because even if brain activity is not measured directly in the form of brain wave signals in these features, physiological parameters are recorded that depend in a direct and consciously uncontrollable manner on the brain activities triggered by viewing familiar and unfamiliar images.It should be emphasized that not every type of physiological parameter is suitable for authenticating a person, because if these physiological parameters do not, on the one hand, have reproducibly identical or very similar properties for at least one person when viewing images of the same type (either a familiar image or an unfamiliar image) and, at the same time, the signals when viewing images of different types have significantly different properties, they are not suitable to be used as a basis for the authentication process.
[0123] The applicant has observed that a significant increase in pupil diameter within the first 2000 ms, preferably within the first 1000 ms, after displaying an image (e.g., a magnification of at least 8%, preferably at least 15%) is an indication that the viewer is seeing a familiar image, in particular a self-image. If this pupil dilation does not occur, the image is considered to be foreign.
[0124] According to one embodiment, the eyelid opening is measured multiple times per displayed image, thereby detecting one or more blinking movements of the person. Preferably, the detected blinking movements serve to correct a detected brain activity signal. This can be advantageous because blinking can sometimes also disrupt other brain activity signals, such as the brainwave signals detected by an electrical sensor. Appropriate correction increases the accuracy of the prediction.
[0125] According to one embodiment, the external images are selected or their brightness is normalized such that they have a similar brightness level to the one known image, or to an average brightness level of several known images, or to the brightness level of the most recently shown known image. For example, a similar brightness level can mean that the mean value of all pixel intensities of the external image is no more than 25%, preferably no more than 10%, below or above the mean value of all pixel intensities of the known image or the aforementioned known images.
[0126] This can have the advantage of reducing or avoiding changes in pupil diameter that are due to adaptation to different brightness conditions and not to the reception of the foreign image / familiar image and that are to be considered as interference signals.
[0127] According to one embodiment, the image analysis is performed by a neural network or other machine learning software (e.g., a support vector machine). The machine learning software may, for example, have learned during a training phase to automatically recognize certain facial regions or eye regions, such as the iris and / or pupil, and optionally also to determine the size and / or position of the pupil (and / or, if necessary, the eyelid opening for signal correction purposes). The result of the image analysis accordingly consists of one or more numerical values that specify the pupil size and / or a pupil size change profile over a time interval. At least one or more of these numerical values are passed on as input to another software module of the authentication module.The additional module is configured to process such numerical values, each acquired when viewing familiar images, as first brain activity signals, and to process corresponding numerical values, each acquired when viewing unfamiliar images, as second brain activity signals. In the next step, the authentication module determines the similarity of the first brain activity signals among themselves (i.e., the similarity between multiple first brain activity signals acquired when viewing multiple familiar images), the similarity between the second brain activity signals among themselves (i.e., the similarity between multiple second brain activity signals acquired when viewing multiple unfamiliar images), and the similarity between the first brain activity signals and the second brain activity signals.For the latter comparison, various methods can be used, such as cluster analysis or a multi-step procedure in which a first mean is calculated from all first brain activity signals, a second mean is calculated from all second brain activity signals, and then the similarity of the first and second mean is calculated. By comparing the calculated similarities or deviations with predefined thresholds, the authentication module can automatically determine whether the first and second brain activity signals differ significantly from each other.
[0128] According to embodiments, the analysis of the acquired brain activity signals comprises the execution of a machine learning program.
[0129] For example, a supervised learning approach can be used to analyze temporal profiles of signal amplitudes (e.g., voltage of an electric field in the case of electrical sensors, pupil diameter in the case of optical sensors) generated by a person when viewing external or familiar images, and to learn to automatically recognize or predict the type of image viewed based on these profiles. For example, the machine learning program can be a neural network or a support vector machine. In particular, the neural network can be designed as a recurrent neural network. During training, the neural network learns how the signal amplitudes or pupil diameters should be determined based on temporal profiles of the signal amplitude (e.g., pupil diameter in micrometers).What signal amplitude profiles look like when viewing familiar and unfamiliar images and / or how they differ when viewing familiar and unfamiliar images.
[0130] According to embodiments, the raw signal detected by the electrical sensor is used and evaluated as the first or second brain activity signal. The raw signal can be the entirety of the brain activity signals detected within one or more frequency ranges or one or more discrete frequencies. According to other embodiments, the raw signal detected by the electrical sensor is first preprocessed, e.g., normalized, denoised, and / or filtered, to increase the signal quality and / or comparability of the signal, and this preprocessed brain activity signal is used and evaluated as the first or second brain activity signal.
[0131] Processing of image data beyond determining pupil diameter is not necessary, as no object classification is performed. Training can therefore be performed efficiently based on annotated data (e.g., pupil diameters or pupil diameter change profiles annotated with "when viewing a familiar image" or "when viewing an unfamiliar image").
[0132] The annotated training data used for training can be recorded and annotated, for example, using a protocol so that the model can be trained and verified.
[0133] According to embodiments, the analysis of the acquired brain activity signals comprises the detection of a correlation of significantly different brain activity signals at image type change times, wherein at an image type change time a known image shown in the display is replaced by an unfamiliar image or vice versa.
[0134] According to embodiments, the method includes a step of enabling and / or executing a function of the server application only if the individual has successfully authenticated. The function may, for example, consist of receiving current facial images of this individual from the client application and forwarding them to software for personalizing a new identity document for the individual. Further embodiments of various KLHA sensor types
[0135] According to embodiments, a reversible connection between the KLHA sensor and the telecommunications device includes a microUSB interface and / or a Bluetooth interface, as well as, optionally, a mechanically stabilizing connection between the telecommunications device and the KLHA sensor. The mechanical connection can be designed, for example, as a plastic holder.
[0136] In some embodiments, the KLHA sensor is signed and the client application is configured to verify the signature when connecting the KLHA sensor to the telecommunications device.
[0137] This can be advantageous because only KLHA sensors that have been manufactured by the operator of the server computer system, e.g., a government agency or a company, or that have been approved as trustworthy and provided with a corresponding signature can be used for the authentication process according to embodiments. This precludes the use of a manipulated sensor that, upon detection of the challenge, automatically generates a signal sequence matching the challenge in order to deceive the authentication module.
[0138] In some embodiments, the KLHA sensor includes a KLHA signature key configured to sign the first and second brain activity signals. The client application includes a KLHA signature verification key corresponding to the KLHA signature key. The client application is configured to use the KLHA signature verification key to verify the signatures of the first and second brain activity signals and to further process the first and second brain activity signals for the purpose of authenticating the individual and / or to forward them to the server application for evaluation only if the signatures of the brain activity signals are valid.
[0139] Corresponding embodiments of the invention can have the advantage that it is not possible for malware possibly installed on the telecommunications device to "intercept" the challenge and automatically generate a sequence of first and second brain activity signals corresponding to the challenge, replacing actually measured signals with these artificially generated signals and thereby enabling an unauthorized person to authenticate as someone else. This is prevented by the authentication module only processing signals with a valid signature after performing a signature check, since only signed signals can reliably be signals that the KLHA sensor has measured.
[0140] According to embodiments of the invention, the KLHA sensor is configured to capture images or image sequences of the face or parts of the face of a person that contain the first and / or second brain activity signals. The analysis of the first and second brain activity signals by the authentication module comprises an automatic image analysis of the images or image sequences. The image analysis comprises a computational determination of a pupil diameter of the person. Additionally or alternatively, the image analysis of the images or image sequences comprises a computational determination of a pupil diameter change profile of the person.
[0141] According to embodiments of the invention, the first brain activity signals are identical or similar to the brain activity signals according to curve 952 in Figure 11 ;
[0142] According to said and / or further embodiments, the second brain activity signals are identical or similar to the brain activity signals according to curve 954 in Figure 11 .
[0143] According to embodiments of the invention, the analysis of the acquired brain activity signals comprises one or more of the following steps: Image processing to detect pupil diameter; and / or correlation of properties of the first and second brain activity signals with the times at which familiar and unfamiliar images were displayed.
[0144] According to embodiments of the invention, the telecommunications device includes a camera. The at least one known image captured during registration is a facial image of the at least one person. If the authentication result indicates that the person has successfully authenticated, the method further comprises: Displaying a GUI of the client application on a display of the telecommunications device, wherein the GUI includes a request to bring the face of the person into the detection range of the camera of the telecommunications device; capturing a current facial image of the person (e.g., at the instigation of the client application); and sending the current facial image from the client application to the server application via the network connection.
[0145] Corresponding embodiments of the invention can have the advantage that, following successful remote authentication of a person, the available equipment can also be used to capture a current facial image. The method inherently ensures that the captured image is a current facial image of the person who has just successfully authenticated. Thus, the person does not need to visit a government agency or an authorized photo studio to provide the agency with a current facial image in a forgery-proof manner.
[0146] According to one embodiment, the method comprises a further step of displaying the captured current facial image via the display of the telecommunications device outside the secure space and capturing brain activity signals while displaying the current facial image. The facial image is only forwarded from the client application to the server application and / or is only accepted by the authentication module as the current facial image of the person who has just authenticated if the captured brain activity signals are significantly similar to the first brain activity signals and significantly dissimilar to the second brain activity signals measured during the authentication process. This increases the certainty that the current facial image actually depicts the face of the person who has just successfully authenticated, and not, for example,a face depicted in a photograph held up to the camera of the telecommunications device after authentication.
[0147] According to embodiments, the method comprises storing and using the current facial image of the one person as the at least one known image of the person instead of the at least one known image captured during registration of the one person.
[0148] Additionally or alternatively, the method comprises providing the current facial image of the one person from the server application to software for personalizing identity or security documents in order to personalize an identity or security document for the one person.
[0149] Corresponding embodiments can have the advantage that a person only needs to authenticate and register once within the secure premises, for example, to obtain an identity card or other identity document with their own facial image for the first time after reaching adulthood. To then obtain a new ID card based on a new and current facial image several years later, the person does not need to appear in person again at the authority's secure premises, neither to authenticate themselves nor to provide a current facial image to the authority issuing the ID card in a secure and reliable manner. Embodiments of the invention can enable remote authentication and, at the same time, to capture a current facial image using a camera of the telecommunications device and transmit it to the server application via the client application.Preferably, the captured current facial image of the person is displayed again, and during this time, the person's brain activity signals are recorded via the KLHA sensor. Only if the captured brain activity signals are similar to the first brain activity signals and dissimilar to the second brain activity signals is the current facial image considered trustworthy and used as the basis for reissuing the identity document. Otherwise, the current facial image is not used for reissuing the identity document. For example, the authentication module can be part of the client application and omit forwarding the image to the server application, or the authentication module can be part of the server application and omit forwarding the facial image to the personalization software.
[0150] In a further aspect, the invention relates to a method for producing an identity or security document for a person. The method comprises authenticating the person according to a remote authentication method according to claim 1, which is described for this purpose in embodiments of the invention that include capturing a current facial image of the person. Furthermore, the method comprises producing an identity or security document that is uniquely assigned to the person, wherein the current facial image of the person, which was transmitted to the server application, is used as a security element of the identity or security document.
[0151] For example, the current facial image can be printed on or into the document or it can be used to create a hologram that is applied to the security document.
[0152] In another aspect, the invention relates to a system for authenticating a person. The system is configured to perform an authentication method according to claim 1 according to one of the embodiments described herein.
[0153] The system includes a server computer system with a server application, a portable telecommunications device and an authentication module.
[0154] The server application contains a register with the profiles of several people registered with the server application. Each profile of a registered person is stored and linked to at least one known image. The at least one known image is an image captured within a secure space while the person was present in the secure space. The known image of a person is an image with an image motif that is familiar to that person.
[0155] The portable telecommunications device is reversibly or irreversibly connected to a KLHA sensor. A client application is instantiated on the telecommunications device, which is interoperable with the server application. The KLHA sensor is a contactless sensor for recording brain activity. The telecommunications device is assigned to a single person.
[0156] The authentication module is part of the server application or the client application.
[0157] The client application is configured to receive the at least one known image from the server application via a network connection. The authentication module is configured to generate a specification of a sequence of known images and foreign images. A foreign image is an image with a motif that is unknown to the person. The client application is configured to display multiple images on a display of the telecommunications device outside the secure space. The displayed images include a sequence of one or more known images and one or more foreign images according to the specification.
[0158] The KLHA sensor is configured to detect first and second brain activity signals, which are respectively triggered in the one person by viewing the images, while the plurality of images are displayed to the person, wherein first brain activity signals are detected while a familiar image is displayed to the person, and wherein second brain activity signals are detected while an unfamiliar image is displayed to the one person.
[0159] The authentication module is designed to authenticate the one person as a function of the similarity of the detected first and second brain activity signals, taking into account the specification of the image sequence, and to output an authentication result with respect to the one person.
[0160] The method for authenticating a person treats the person as successfully authenticated if the first brain activity signals are similar to each other, if the second brain activity signals are similar to each other, if the first brain activity signals are dissimilar to the second brain activity signals, and if the sequence of mutually similar first brain activity signals and mutually similar second brain activity signals corresponds to the sequence of the displayed familiar images and unfamiliar images.
[0161] For example, the output of the result may be that the authentication module sends a message to the server application that the person has successfully authenticated.This can in turn cause the server application to perform further steps, for example to send a message to the client application to cause the client application to capture a current facial image of the person, to display the captured facial image on the display of the telecommunications device, meanwhile to measure brain activity signals of the person with the KLHA sensor, to compare these brain activity signals with the first and second brain signals captured during the authentication process, and then, if the brain activity signals obtained when viewing the current facial image are similar to the first brain activity signals and dissimilar to the second brain activity signals, to forward the captured current facial image to the server application for further processing.Further processing may, for example, include using the current facial image to reissue an identity document for that person.
[0162] According to embodiments, the KLHA sensor is an electrical sensor for detecting a brain signal without direct contact of the head with the electrical sensor.
[0163] According to other embodiments, the KLHA sensor is an optical sensor.
[0164] According to one embodiment, a training dataset is generated by measuring brain activity signals during the viewing of several familiar images and images of others, for example, during a person's registration with the authenticating entity. A person-specific classifier is trained on these images. The trained classifier is capable of classifying brain activity signals specific to this person into those that are the result of viewing a familiar image and those that are the result of viewing an image of others. The person-specifically trained classifier can be stored, for example, as part of this person's user profile.
[0165] Several embodiments of the invention have been described with reference to a public authority as an example of an authenticating authority. However, instead of a public authority, a private organization, e.g., a company, an association, or any other entity to which a user wishes to authenticate, can also act.
[0166] A " identity document " (also called "ID document") is a physical object, for example, a paper-based or plastic-based card with a chip, that proves the identity or a property or right of the holder of the identity document. An identity document can be, for example, an identity card, a passport, an employee ID card, or a personalized valuable document (for example, a credit card). Identity documents often contain security features such as facial images, holograms, check digits, cryptographic keys, or the like.
[0167] Under a " Display device" "Display" refers here to a device designed to display time-varying information such as images or characters. In particular, the display device can be a screen or other form of electrically controlled display. However, it can also be a combination of a projector and a projection surface. The display device can be a standalone device or part of a device, e.g., part of a terminal.
[0168] Under a "Sensor",A sensor, also known as a detector or (measuring) sensor, is a technical component or device that can qualitatively or quantitatively detect certain physical and / or chemical properties (especially impedance, electric fields, electromagnetic fields, optical signals, charge changes and / or charge redistributions in the environment, etc.). These quantities are detected using physical or chemical effects and converted by the sensor into a processable electrical signal.
[0169] Under a " electrical sensor" is understood here to mean a sensor which is designed to measure charge changes and / or charge redistributions in its environment, wherein the environment is the spatial sensitivity range of the sensor.
[0170] Under a " optical sensor"is referred to here as a sensor which is designed to capture one or more images of a physical object, e.g. a face or facial region. The images can in particular be digital images. The optical sensor can be, for example, a camera or a video camera. The optical sensor can, for example, be designed to selectively capture light of a specific wavelength range, e.g. infrared light (IR light) or light in the wavelength range visible to the human eye (white light, daylight). The captured images are recorded chemically or electronically. The captured images are preferably digital images which are stored electronically and / or provided directly to an authentication module.
[0171] Under a "contactless brain activity sensor (KLHA sensor)"This refers to a sensor that can receive one or more signals emitted by a human without requiring contact between the sensor and the human during reception. The received signals provide information, either directly (brain wave signals measured by a contactless electrical sensor) or indirectly (pupil diameter, pupil diameter change profile), about the brain activity of the person from whom the signals are received at the time of reception.
[0172] Under a " Brain activity" This refers to the physiological activity of a multitude of neurons in a person's brain. Physiological activity corresponds to changes in the electrical state of the neurons, which are responsible for the brain's information processing.
[0173] Under a " Brain activity signal"This refers to a signal emitted by a person that is directly or indirectly caused by brain activity. For example, the electrical state changes and potential changes of the neurons distributed throughout the head add up according to their specific spatial arrangement and can be directly measured as so-called "brain wave signals" using an electrical sensor. A brain activity signal can also be any other signal emitted by a physiological process in a human body that is caused by brain activity and can preferably not be consciously controlled or can only be controlled to a limited extent. For example, adjusting the iris to a certain pupil diameter depends on unconscious neural processing in the human brain, which depends, for example, on what a person is currently seeing.an image of themselves or an image of another person (external image).
[0174] Under a " first brain activity signal" This refers to a brain activity signal that is recorded while a person is looking at a familiar image.
[0175] Under a " second brain activity signal" is understood here as a brain activity signal which is recorded while a person is looking at an image of another person
[0176] Under a " Brainwave signal" is understood here as a brain activity signal which results from the superposition of a large number of electrical state changes, in particular potential changes, of the neurons distributed over the entire head of a person according to their specific spatial arrangement.
[0177] Under a " facial image" is understood here to mean an image, in particular a digital image, which depicts the face or a facial area of a person.
[0178] A "digital image"is a data set in which image content is represented and stored. In particular, the digital image can be a data set in which the content of an image is represented by integers. In particular, the digital image can be a raster graphic.
[0179] Under a " confidant" is understood here to mean a person who has registered in advance with the authentication module and / or the server application and who, after successful authentication of the person with the authentication module, is treated as trustworthy by the person and the server application, at least with regard to certain selected functions of the server application.
[0180] Under a " "familiar image"is understood here to be an image with an image motif that is familiar to the person, provided that the person is the person as whom the person is authenticating. For example, during the registration process of a person in secure premises, the operator of a server computer system creates a secure, trustworthy link between a profile of this one person and one or more known images for this person. During later authentication processes, in which an unknown person attempts to authenticate themselves from a distance as the person already registered, this person is shown a - preferably random - sequence of known images and foreign images, whereby the known images are the known images that were linked during the registration process to the profile of the person as whom the person just authenticated is claiming to be.If it is actually another person, i.e. an attempt at deception, this will be recognized because the brain activity signals generated when looking at the familiar images are similar to brain activity signals generated by this person when looking at unfamiliar images.
[0181] The motifs of the known images and foreign images determined and displayed by the authentication module preferably belong to the same type. A motif type can be, for example, a type of physical object, e.g., a face, the human body, an object of a specific category, a vehicle, a watercraft, a building, a species of animal, particularly a pet. The known images of a trusted person can be, for example, a work building or pet, or the face of a family member, while the foreign image can be an unknown building of a similar size or function, an unknown individual of the same pet species, the face of a stranger, etc.
[0182] Preferably, a known image is an image where it is known, or the operator of the authentication module can assume with sufficient probability, that the motif of the known image is known only to the trusted person and no other person, or at least no other person who is sufficiently likely to attempt unauthorized authentication. For example, objects in a room to which only a small group of people has access may well be known to all members of this group of people and be used as the motif of a known image. However, all of these people are considered authorized to access this room. For the object to be used as the motif of a known image, it is only necessary that it does not appear in other rooms in which other people work who should not be granted access to the protected room.
[0183] Familiarity is to be understood here in the sense of a high degree of familiarity, which arises from repeated viewing of the motif of the familiar image over a longer period of time, preferably through at least daily or at least weekly viewing over a period of several months or years.
[0184] Under a " External image " is understood here as an image depicting a physical object of which it is known or can be assumed with a sufficiently high probability that the depicted physical object is unknown to the authenticating person or at least significantly less known than to the trusted person, provided that the authenticating person is actually the person as whom they wish to authenticate.
[0185] For example, an external image shown to a specific person may depict the face of another person who lives on a different continent, who lived a long time ago, or who, for other reasons, cannot be assumed to have ever been seen by the person seeking authentication. The "unknown" nature of the subject of an external image should preferably be understood as meaning that, with a certain minimum probability, no person likely to use the authentication procedure has ever seen the subject depicted in the external image, for example, another person.
[0186] Under a " significant similarity"Distinctiveness of data values such as brain activity signals is understood here in particular to mean a situation in which two signals are considered to be identical or very similar within the scope of the measurement accuracy. Taking into account the measurement inaccuracy of the sensor type and sensor model, "significantly similar signals" can therefore be signals whose difference score lies below a predefined threshold, whereby the threshold is chosen such that difference scores of brain activity signals generated by essentially identical or very similar processes in the person's head lie below this threshold. Identical or nearly identical brain activity signals are examples of signals that are significantly similar to one another.
[0187] Under a " significant dissimilarity" of data values such as brain activity signals is understood here to mean that the compared data differ so greatly from each other that it is very unlikely that the differences are due to metrological or other variances (measurement errors). When measuring signals (e.g., with optical or electrical sensors), it is inevitable that slight fluctuations in signal strength will occur.The resulting variance, also referred to as measurement error or "background noise," varies in magnitude and nature depending on the sensor type and model. Therefore, it is not possible to specify an absolute limit for all conceivable types and models of sensors, at which level signal differences are still different within the measurement uncertainty, and when they are, however, "significantly different" ("significantly dissimilar"), i.e., so different that the differences can no longer be attributed to the usual measurement uncertainty. However, experts are familiar with how to deal with measurement inaccuracies when processing and comparing measured signals. By selecting suitable limit values, e.g.a minimum distance between two signal amplitudes or a minimum area calculated as the difference between the areas under two amplitude profiles of the two signals over time, the expert can define from which difference threshold two brain activity signals are considered "significantly different" and when the signals, on the other hand, are considered "identical or similar within the scope of the respective measurement accuracy".
[0188] Taking into account the measurement uncertainty of the sensor type and sensor model, "significantly dissimilar signals" or "signals that are significantly different" can be signals whose dissimilarity score lies above a predefined threshold, whereby the threshold is chosen such that dissimilarity scores of brain activity signals generated by essentially different processes in the person's head (e.g., when viewing images of others on the one hand and familiar images on the other) lie above this threshold.
[0189] Under a " Authentication module " is understood here as software that is designed to carry out at least some partial steps of an authentication process in order to authenticate a specific person. The software can be designed as a component of the client application or the server application.
[0190] Authentication is the verification of a person's claimed property (claim) by a verifying entity (e.g., an authentication module). The person provides data to the verifying entity for this purpose, thus performing authentication. The word "authentication" here refers to the entire process of providing the information (authentication) as well as the verification of the authenticity of this data. Authenticating a person with regard to the claimed property of authenticity, which can, for example, be granting "existing access authorization" or "membership in a group of people that has certain rights to access certain functions," allows the authenticated person to perform further actions. The person is then considered authentic or "successfully authenticated."
[0191] Under a " Challenge" is understood here as a challenge for use in a challenge-response protocol. A challenge is a set of one or more data values which is known to a first party and which is transmitted to a second party in a specific way to enable the second party to authenticate itself to the first party by proving knowledge of the challenge or by proving a predefined processing of the challenge.
[0192] Under a " safe space"Secure premises" is understood here to mean a premises within which the collection, storage and / or processing of personal data is considered sufficiently secure with regard to a specific purpose due to organizational and / or technical precautions. For example, the premises of a public authority for the capture of a person's facial image for the purpose of issuing or reissuing an identity or security document may be considered a "secure premises" because the employees there are appropriately trained and optionally also use certain technical sensors to ensure that the captured facial image actually depicts the person the person claims to be.In some cases, it is also possible for a private photo studio, which is deemed sufficiently trustworthy by the authority and is accordingly authorized and qualified to take facial images for document creation, to be considered a secure premises. However, the home or workplace of the person applying for the identity document is not generally considered a secure premises. Preferably, the "secure premises" is also particularly protected from a technical perspective. For example, the secure premises may only be open to persons who have previously identified themselves and / or may contain technical devices, particularly computers, designed to check whether personal data collected and / or processed to personalize a document originates from a data source considered secure (e.g.,from a camera or photo studio that can prove its trustworthiness, e.g. by means of an electronic signature of the facial image.
[0193] Under a " Server computer system" is understood here to be a distributed or monolithic computer system that is connected to one or more other data processing systems ("clients"), e.g. telecommunications devices, via a network and provides data and / or services to these clients.
[0194] Under a " Server application " is understood here as a software application that is instantiated on the server computer system.
[0195] Under a " Registration"This refers to the process of entering information into a directory, e.g., creating a new user profile for a person in a user profile database. Preferably, the registration process requires authentication of the person for whom the new user account is to be created, in secure premises of the organization that manages the user profile database.
[0196] Under a " Client application " is understood here as a software application that is instantiated on the portable telecommunications device and that is interoperable with the server application (i.e., able to exchange data, in particular user data and commands, in such a way that the respective counterpart can correctly process and interpret this data).
[0197] Under a " portable telecommunications device"This term refers to a data processing device whose size and weight allow a person to carry the telecommunications device with them, and which is designed to exchange data with other data processing devices via a network connection, in particular a mobile radio connection. In particular, the telecommunications device can be notebooks, tablet computers, and mobile phones, in particular so-called "smartphones." Smartphones are mobile phones with extensive computer functionality and connectivity. Internet access is generally possible either via a mobile broadband connection provided by the mobile phone provider or via Wi-Fi. Short description of the drawing
[0198] Embodiments of the invention are described below with reference to the drawing. In the drawing, Fig. 1 shows an exemplary flow diagram of an embodiment of the method according to the invention for the registration and subsequent authentication of a person; Fig. 2 shows a flow diagram of a variant of the method according to Figure 1 , according to which the acquaintance images are stored on the client side during registration; Fig. 3 a flowchart of another variant of the method according to Figure 1, according to which the familiar images captured during registration are stored on the server side and transmitted to the client application in response to an authentication request; Fig. 4 a block diagram of an authentication system with a server-side authentication module; Fig. 5 a block diagram of an authentication system with a client-side authentication module; Fig. 6 an authentication system which, following successful authentication of a person based on an image sequence, captures a current facial image of the person and displays it on a display in order to verify the authenticity of the captured facial image; Fig. 7 a sequence of displayed familiar images and foreign images and corresponding first and second brain activity signals; Fig. 8 a diagram of a neural network for analyzing and classifying the brain activity signals; Fig.Fig. 9 shows a plot with two brain activity signals in the form of measured charge redistributions in the sensor environment when viewing an external image and a known image; Fig. 10 shows a block diagram of another authentication system; and Fig. 11 shows a plot with brain activity signals in the form of measured pupil diameters when viewing an external image and a known image.
[0199] Figure 1 illustrates an exemplary flowchart of an embodiment of the inventive method for registration and subsequent authentication of a person.
[0200] The method according to embodiments of the invention may be advantageous because it enables a person who, for health or other reasons, is unable to visit certain premises of an authority or other organization that are considered secure in order to authenticate themselves, to authenticate themselves remotely in a convenient yet secure manner.
[0201] For this purpose, the person must first register once with the authority or with another organization that operates an authentication system according to embodiments of the invention. The registration step 102 involves the person visiting a premises of the organization that are considered secure in order to authenticate themselves there and to provide the organization with one or more known images. For example, the person can identify themselves to an authority using various documents, ID cards, birth certificates, biometric or other features in order to receive an identity card containing this person's facial image upon reaching the age of majority. During the registration process, a user profile is created for the registered person or, if this user profile already existed, is updated. One or more known images of this one person are captured in the secure premises.Typically, these are one or more facial images or full-body images of that person. However, in individual cases, they may also be images of other objects that the organization can be certain, based on the circumstances, are very familiar to that one person, but unknown or significantly less familiar to virtually all other people.
[0202] The secure premises could, for example, be a room within a government agency. However, it is also possible that the secure premises actually consist of a combination of different premises. For example, a government agency may allow certain trusted photo studios to create passport photos in digital form and transmit them to the agency on behalf of that individual. In such cases, the location where the passport photos are captured and the location where the individual is authenticated during registration with the organization may be different. Nevertheless, the photo studio is also considered a secure premises because organizational and / or technical measures have been taken in advance to ensure that the captured passport photos actually belong to the person currently registering. For example, the organization may generate a transaction number for the purpose of registering the individual.The individual can then go to the photo studio with this transaction number and have digital facial images created. The digital facial images are then signed by the camera and / or software used by the photo studio. The photo studio then sends the signed digital facial images, along with the transaction number, to the organization where the registration is being processed. The organization verifies both the transaction number and the signature of the images and can therefore be sure that the photo actually depicts the face of the person wishing to register, and is not a "morphed" image created on the computer based on the facial images of two or more people.
[0203] The captured acquaintance images are stored on the server and / or client side, depending on the embodiment. The acquaintance images are stored in such a way that they are linked to the user profile of the person during whose registration they were captured. It is sufficient that a single acquaintance image, for example, a facial image, is captured during registration. This can later serve as a basis for generating multiple acquaintance images in the form of multiple copies of this single acquaintance image. However, it is also possible that multiple acquaintance images are captured during registration, for example, multiple variants of a facial image that exhibit slight variations due to slightly different facial positions and / or slightly different facial expressions.
[0204] In a next step 104, the known images and optionally also several third-party images are sent from the server application to a client application via a network, for example the Internet. The client application is instantiated on a portable telecommunications device of the person who registered in step 102. The portable telecommunications device can in particular be a smartphone. The targeted sending of the known images and optionally also the third-party images to the client application of only the telecommunications device that is also assigned to this person can be achieved, for example, by generating a user name and password for the user profile created during registration, and the person downloading a so-called client app and logging in there using this user name and password combination.The client application can then query the server application as to whether there are any known and / or external images for this person and automatically download the corresponding images in the background.
[0205] Step 104 can be implemented very differently depending on the embodiment. For example, the familiar images can be transferred once from the server application to the client application during or immediately after registration, where they are then stored locally and can be used repeatedly for remote authentication. In this case, the familiar images can optionally be deleted on the server side. Alternatively, the familiar images can also be stored primarily on the server side and transferred remotely to the telecommunications device for each subsequent authentication. Details of corresponding embodiments are discussed in more detail in other figures.
[0206] In a next step 106, an authentication module generates a (preferably random) sequence ("challenge") of several known and unknown images of the person. The authentication module can be part of the client application or part of the server application. If it is part of the server application, the specification of the sequence is transmitted to the client application over the network.
[0207] The sequence contains a mixture of foreign images and known images in a specific, preferably random, sequence that is newly generated for each authentication, the sequence being known to the authentication module. A random sequence of known and foreign images is an image sequence in which the chronological order of the known images and foreign images, and optionally also the number of known images and / or foreign images, is randomly selected. According to one embodiment, the sequence of known images and foreign images is randomly generated anew by the authentication module for each authentication process.
[0208] The multiple known images may be copies of a single known image. The multiple external images may be copies of a single external image. However, they may also be mixtures of different known images and / or different external images.
[0209] Depending on whether the authentication module is implemented on the server or client side, if the image sequence is generated on the server side, a specification of the image sequence and, optionally, the corresponding external images and known images must be transmitted from the server application to the client application over a network. This step is not necessary if the image sequence is generated on the client side (i.e., by the client application of the portable telecommunications device) and the corresponding known images and external images are already available there.
[0210] Once the person has been registered in the secure premises and one or more acquaintance images have been captured for that person and stored on the client and / or server side, the technical requirements are in place for the person to be able to remotely authenticate themselves to the organization one or more times from now on, as described below.
[0211] For example, several months or years may have passed since the person's registration (step 102) before the person is remotely re-authenticated to the organization according to the following steps 108-114. For example, after a few years, the ID card issued during registration may expire and need to be reissued. However, for health or other reasons, the person cannot visit the secure premises of the authority and therefore requests remote authentication. For example, in response to receiving such a request, the authority or other organization may send a contactless sensor for recording brain activity signals (KLHA sensor), for example, a special camera or an electrical sensor, to the desired address specified by the person.The KLHA sensor is designed to be irreversibly or reversibly connected to the person's portable telecommunications device, allowing it to forward recorded brain activity signals to the telecommunications device's client application. After the person has received the KLHA sensor and connected it to their own telecommunications device, the next steps can be performed. In some embodiments, it is also possible to use the camera already built into the telecommunications device by the manufacturer as an optical KLHA sensor. In this case, shipping the KLHA sensor by mail is not necessary.
[0212] In step 108, the sequence of known images and foreign images, which was generated for the purpose of authenticating the person, is displayed on a display of the telecommunications device so that the person can see the image sequence.
[0213] During the display of the image sequence, in step 110, the KLHA sensor records the brain activity signals generated by the person while viewing the sequence of unfamiliar and familiar images. The KLHA sensor may also comprise multiple sensors. Where reference is made to a single sensor in this application, the possibility of using multiple sensors is implicitly included.
[0214] The authentication module knows the time periods at which familiar images and foreign images were displayed. The authentication module receives the brain activity signals recorded by the KLHA sensor during the display of a familiar image as so-called "first brain activity signals" from the KLHA sensor. Brain activity signals recorded by the KLHA sensor during the display of a foreign image are received by the authentication module as so-called "second brain activity signals."
[0215] The sensor is provided in such a way that the position, distance and orientation of the sensor relative to the head of the person wishing to authenticate themselves is suitable to capture the brain activity signal to be captured in the respective case with sufficient strength.
[0216] If the KLHA sensor is an electrical sensor, for example, which can be mailed by the organization to which the person wishes to authenticate themselves, this transmitter can be sent to the person together with a special holder. The holder can be made of wood, plastic, cardboard, or another material or composite. The holder can include one or more recesses and / or retaining elements that mechanically fix and hold both the portable telecommunications device and the KLHA sensor. Preferably, the holder also includes one or more stand elements ("feet") so that the holder, including the telecommunications device and the KLHA sensor, can be securely positioned on a flat surface, such as a tabletop.The client application can be configured to indicate to the user how to position themselves or their face relative to the sensor while viewing the image sequence on the display of the telecommunications device. This ensures that the person positions their head at a reasonable and, as far as possible, constant distance from the KLHA sensor. According to some embodiments, particularly with electrical KLHA sensors, the distance between the head and the KLHA sensor is less than 15 cm, preferably less than 10 cm, with no contact between the head and the KLHA sensor.
[0217] If the sensor is an optical sensor, for example an external camera integrated by the manufacturer or provided by the said organization and usable with the telecommunications device, the distance between the head and the KLHA sensor can also be greater so that, in particular, the person's eye area can be fully optically captured.
[0218] In step 112, the authentication module performs an authentication process. During this authentication, the authentication module checks, for example, whether the chronological order of the first and second brain activity signals corresponds to the temporal pattern of familiar images and unfamiliar images encoded in the specification of the image sequence ("challenge"). In doing so, the authentication module checks, for example, whether the same or a very similar first brain activity signal was always received at the times when familiar images were shown, and whether the same or a very similar second brain activity signal was always received at the times when unfamiliar images were shown, with the first and second brain activity signals being significantly different from each other. Only if all of these criteria are met will the person be treated as successfully authenticated by the authentication module.
[0219] According to some embodiments, the sequence of known images and foreign images is generated such that it contains a minimum number of known images and / or foreign images, e.g., at least 3 or at least 5 images each. During the evaluation of the first and second brain activity signals, the authentication module checks whether the first brain activity signals are similar to each other, whether the second brain activity signals are similar to each other, and whether the first brain activity signals are dissimilar to the second brain activity signals. In some embodiments, in which the challenge contains a large number of images, e.g.,at least 10 images in total, the authentication module can also have a certain tolerance for individual false results, for example a maximum of one brain activity signal that looks more similar to the other signal category (first or second brain activity signal) than the category to which it should belong based on the shown image. In some embodiments, the requirements for signal category fidelity (reproducibility) are absolute, i.e., there must be no false deviation from expected and observed signal similarities. In other embodiments, e.g., those where very long signal sequences are used and / or the security requirements are lower, the requirements for reproducibility can be lower and allow one or a few errors.
[0220] In step 114, the authentication module outputs the results of the first and second authentication methods. For example, the authentication module can be instantiated on the client side and communicate the authentication result to the server application, so that the server application performs certain functions depending on the transmitted result and / or releases these functions for use by a successfully authenticated person. Additionally or alternatively, the authentication result can be displayed to the person on the display of the telecommunications device.
[0221] Figure 2 a flow chart of a variant of the procedure according to Figure 1 , after which the acquaintance pictures are stored on the client side during registration.
[0222] The process steps 102, 104, 106-114 essentially correspond to those already described for Figure 1described steps and will therefore not be repeated here. Figure 2 In the process variant described, the acquaintance image(s) captured during registration are sent from the server application to the client application via a network connection during or after registration. For example, it can be part of the process according to Figure 2 It may be that the person who registers in step 102 also downloads the client application to his or her own smartphone during the registration process, instantiates it there, and downloads the captured acquaintance images and preferably also one or more third-party images during the registration process.
[0223] In step 150, the one or more known images of the person, and optionally also one or more third-party images, are stored in a non-volatile storage medium of the portable telecommunications device. Preferably, the one or more known images of the registered person are then completely deleted from the server computer system. This can have the advantage of particularly high data protection. The known images are personal data, for example, facial images. Centrally storing the facial images of a large number of people carries the risk that, in the event of unauthorized persons penetrating the server computer system, personal data of a large number of people could be stolen. To prevent this, high security measures can be implemented on the server side, for example.An alternative to this could be not to store all the acquaintance pictures permanently on the server at all, but to only store the acquaintance pictures locally in the data storage of the portable telecommunications devices of the persons registered with the server.
[0224] In step 156, the authentication module receives an authentication request from the client application. If the authentication module is part of the server application, the authentication request is received over a network. Otherwise, the authentication module is part of the small application and therefore contains the request from another module of the same application. The subsequent steps 106-114 are performed in response to the authentication request based on the images stored on the client side. It is therefore not necessary for the known images and optionally also the foreign images to be transferred from the server computer system to the telecommunications device for each new authentication request. This can have the advantage of reducing the amount of data to be transferred and thus accelerating the authentication process.
[0225] Figure 3shows a flow chart of another variant of the procedure according to Figure 1 , according to which the acquaintance images collected during registration are stored on the server side and transmitted to the client application in response to each future authentication request.
[0226] The process steps 102, 104, 106-114 essentially correspond to those already described for Figure 1 described steps and will therefore not be repeated here. Figure 3In the method variant described, the known image(s) captured during registration is / are stored in a non-volatile memory of the server computer system during or after registration in step 160. For example, the server computer system can contain a user profile database in which a plurality of user profiles are stored. Each of the user profiles represents a person who has registered with the operator of the server computer system. In addition, the server computer system can contain an image database that contains a plurality of known images and third-party images, wherein at least each of the known images is assigned to only a single user profile. The user profiles and images can be stored in different or similar databases.The database can be, for example, a relational database, a directory within a directory tree, or another data structure for organized data storage. Preferably, the user profiles and / or acquaintance images and / or third-party images are stored on a non-volatile storage medium of the server computer system.
[0227] In step 162, the server application receives an authentication request. The authentication request includes an identifier for the person, e.g., a personal ID, a combination of name and date of birth, or other data that makes the person and their associated user profile uniquely identifiable for the server application.
[0228] For example, the authentication module can be instantiated on the server side and receive the authentication request over the network. In response to receiving the authentication request, the server application uses the person ID to identify the user profile representing this person, as well as one or more server-side acquaintance images associated with this user profile. The at least one identified acquaintance image and, optionally, one or more external images are transmitted by the server application to the client application via a network connection in response to the authentication request in step 104.
[0229] The further steps 106-114 are performed in response to the authentication request based on the images transmitted by the anti-client application for a specific authentication request. Preferably, the transmitted images are stored only temporarily on the client side, for example, only for the duration of the authentication process, and then deleted. Therefore, for each renewed authentication request, the known images and optionally also the foreign images are retransmitted from the server computer system to the telecommunications device in response to a corresponding authentication request. This can have the advantage of reducing memory consumption on the client side.Given that storage space on portable devices is often scarce and expensive, and that the memory requirements can be considerable, particularly when using a large number of known images and external images, this embodiment can have the advantage of lower client-side memory consumption.
[0230] Figure 4 shows a block diagram of an authentication system 200.1 with server-side authentication module.
[0231] The system 200.1 includes a server computer system 202. The server computer system can be designed as a monolithic computer system or as a distributed computer system, for example, as a cloud computer system. A server application 205 is instantiated on the server computer system, which, according to the Figure 4shown embodiment includes an authentication module 204. The server computer system 202 includes a data store 207, which can consist of one or more physical data storage media. Preferably, the data store 207 is designed as a non-volatile data store. The data store 207 includes a database with the user profiles of a large number of users who have securely registered with the server application 205. In addition, the data store 207 includes an image database 206 with several third-party images 208, 210, 212, 214 and several acquaintance images 230, 232. Each acquaintance image is assigned to a maximum of one of the user profiles. The assignment is based on the fact that, during the person's registration, one or more acquaintance images are captured specifically by and / or for this person and stored in a link with this person's user profile.The capture of a known image for a specific person 224 takes place within a secure room 262 and preferably with technically secured image capture systems 203, for example cameras.
[0232] According to one embodiment, the camera used to create a known image for a specific person during the person's registration contains a signing key. The camera is also configured to use this signing key to sign the images captured by the camera. The server application contains a signature verification key corresponding to the signing key and uses this key to verify the signature of the images captured by the camera and to use the images as known images of the person only if the signature is valid.
[0233] According to one embodiment, the authentication system 200.1 also includes the camera 203. According to other embodiments, the camera is not included in the system 200.1.
[0234] The system also includes a portable telecommunications device 201 belonging to the person 224 who previously registered with the server application 205. The telecommunications device includes a camera 242 and a client application 240 that is interoperable with the server application 205 and can exchange data via a network connection. The telecommunications device is communicatively and preferably also mechanically coupled to a KLHA sensor 218. The KLHA sensor can be an optical sensor, for example. In this case, the sensor is either designed as an additional external special camera that can be subsequently connected to the telecommunications device 201, or the optical KLHA sensor 218 and the camera 242 integrated into the telecommunications device 201 by the manufacturer are the same camera.
[0235] According to other embodiments, the KLHA sensor 218 is an electrical sensor that was delivered, for example, by mail to the person 224 for the purpose of remote authentication. The sensor can be subsequently coupled to the telecommunications device via appropriate interfaces and holding devices, with the coupling preferably being reversible. This has the advantage that the person 224 can return the KLHA sensor to the operator of the server computer system 202 after use.
[0236] The telecommunications device includes a temporary data storage 209, for example, a working memory. According to embodiments of the invention, this is used to temporarily store the known images, which are normally permanently stored on the server side, locally on the telecommunications device for the purpose of remotely authenticating the person. For example, a client-side copy of a portion of the image database 206 can be temporarily stored, wherein the portion of the image database includes one or more external images and selectively only the known images of the person to whom the telecommunications device is assigned. After completion of the authentication process, the temporarily stored images can be deleted again by the client application. Furthermore, the telecommunications device 201, which can be embodied, for example, as a smartphone, includes a display 216, for example, a touchscreen.The client application 240 is configured to generate a graphical user interface (GUI) 219 that guides the person through the authentication process.
[0237] The following will briefly outline how this Figure 4 illustrated authentication system 200.1 can be used according to embodiments of the invention to remotely authenticate an already registered person and also, after successful authentication, to immediately transmit a current known image, for example a facial image, of this person to the server in order to use this image for future authentication attempts and / or to personalize documents for this person.
[0238] First, the person 224 registers in the secure room 262 as already described. A camera 203 captures one or more familiar images, which are linked to a user profile of this person and stored in a non-volatile data storage 207 of the server computer system 202. For example, the familiar image 230 is captured during registration and stored in the image database 206 on the server side in the data storage 207.
[0239] After many years, the person 224 wishes to authenticate themselves again with the server application outside of the secured premises, for example in their own home, for example to use a specific service or to reissue an ID card. To enable this, the person 224 requires a KLHA sensor 218 that can be coupled to their own smartphone. This sensor can be requested, for example, by mail from the operator of the server application or, if it is an optical sensor, can already be included in the smartphone in the form of the camera 242. The person can interact with a GUI 219 of the client application 240 to initiate the sending of an authentication request 233 to the server application 205. The request contains an identifier of the person 224, i.e., a data value that allows the server application to identify one or more familiar images assigned to this person within the image database 206.In response to the authentication request, the authentication module 204 generates a specification 234 of a random sequence consisting of several external images and several identified known images of this person 224. The specification of the sequence, and optionally also the sequence of the external images and known images, is transmitted by the server application in response to the authentication request 233 to the client application via a network connection. The client application then generates a GUI 219 that instructs the person 224 to position their head and the KLHA sensor 218 in a specific manner relative to each other and to concentrate on the images shown below. If the person confirms that they have understood the instruction, the client application 240 displays a sequence of external images and known images according to the transmitted sequence specification 234 via the display 216.While the image sequence is displayed, the KLHA sensor continuously and contactlessly records brain activity signals from person 224. Optionally, camera 242 can record additional data from the person, for example, whether the eyes are actually focused on display 216. The brain activity signals recorded while displaying familiar images are referred to as first brain activity signals. The brain activity signals generated while displaying unfamiliar images are referred to as second brain activity signals. The recorded first and second brain activity signals are transmitted over the network to the authentication module in the same chronological order in which they were recorded. Since the authentication module is instantiated on the server side in the example shown here, the brain activity signals are sent over the network.The authentication module determines the similarities and dissimilarities of the first and second brain activity signals to determine whether the similarities and dissimilarities of the brain activity signals ("response") correspond to the random sequence of known images and unfamiliar images ("challenge"). If the pattern of the brain activity signals corresponds to the expected pattern based on the specified sequence, the person is considered successfully authenticated. This can result in a series of further steps. For example, the client application can then generate a GUI 219 informing the person that they have successfully authenticated and in which the person is prompted to hold their face in a specific position relative to the camera 242 so that the camera can capture a facial image of the person.Optionally, the client application can check the correct distance between the face and the camera and, if necessary, generate and output corrective instructions. If the person 224 has correctly positioned their face relative to the camera 242, the client application triggers the capture of a current facial image 232 by the camera 242. The current facial image can then be transmitted directly to the server application via the network connection, for example, to replace the known images for the person 224 captured during registration and / or to use the current known image 232 to personalize an identity document or security document for the person 224. Optionally, a further verification step can be performed before the step of transmitting the current known image 232, so that the current known image 232 is only transmitted to the server application 205 if the current known image 232 is considered valid.The verification step may include the captured current known image 232 being displayed again by the client application 240 via the GUI of the person 224, during which time brain activity signals of the person 224 are captured by the KLHA sensor 218. The authentication module (which may be presented on the client side or the server side) compares the brain activity signals captured during the display of the current known image 232 with the first and second brain activity signals captured during the display of the image sequence. Only if the brain activity signal of the current known image is similar to the first brain activity signals and dissimilar to the second brain activity signals is the current known image 232 considered valid and trustworthy. This can help to rule out the possibility that the person puts on a mask during the capture of the current known image in order to generate an identity document with a forged image.
[0240] Figure 5 shows a block diagram of an authentication system 200.2 with client-side authentication module 204.
[0241] The individual system elements essentially correspond to the elements of the system according to the embodiment shown in Figure 4 and will not be explained again here. Rather, only those aspects will be discussed here in relation to which the Figure 5 system shown in Figure 4 system shown.
[0242] The server computer system 202 includes a data memory 235, which can be embodied as a volatile or non-volatile data memory. The data memory 235 comprises an image database 206, in which third-party images and / or acquaintance images of various persons registered with the server application are stored. At least the acquaintance images of the persons are only briefly, i.e., temporarily, recorded. The storage period preferably extends only to the period until the recorded acquaintance images could be successfully transmitted to a client application 240 of the portable telecommunications device 201 of the person 224. For example, the server application can be configured to transmit the acquaintance images of the person 224 recorded during registration to the client application 240 via a network connection during or immediately after registration.The client application is configured to store the known images of the person 224 and, optionally, also the third-party images in a local data storage 236 after receiving the known images and, optionally, also the third-party images. The local data storage 236 is preferably configured as a non-volatile data storage. After the images have been successfully stored on the data storage 236, the client application sends a message regarding the successful storage to the server application 205. Receipt of the message causes the server application to delete the known images of the person 224 from the data storage 235. As a result, after the person has been registered, the sensitive known images are only available on the client side. Preferably, but not necessarily, in embodiments with purely client-side storage of the known images, the authentication module is also a component of the client application.
[0243] In order to remotely re-authenticate the person months and years after successful registration, the client application 240 generates a specification 234 of a sequence of foreign images and known images specifically for the current authentication attempt. The sequence is preferably random and is generated anew by the authentication module for each individual authentication attempt. The client application displays a sequence of foreign images and known images based on the local image database 206 and according to the sequence specification 234 via a GUI 219. The following steps can be performed as for Figure 4described. A transmission of known images from the server to the client for each individual authentication process is not possible and not necessary, since the known images are stored on the client side. In some embodiments, however, it is possible for the external images to be stored on the server side and, in this case, to be transmitted again for each authentication process. This can have the advantage that if the remote authentication method is used very frequently, there is no habituation effect to a limited number of locally stored external images. A much larger number of external images can generally be stored on the server side, making it possible to present the user with external images that the user has not seen before during each authentication process.This ensures that the difference between the first and second brain activity signals remains consistently high even with frequent use of the authentication method according to embodiments of the invention.
[0244] Figure 6 shows an authentication system 200 which, following a successful authentication of a person 224 based on an image sequence, captures a current facial image 232 of the person and displays it on a display in order to verify the authenticity of the captured facial image.
[0245] First, person 224 is shown a sequence 602 of known images and foreign images, randomly generated by the authentication module for a current authentication process of person 224. The sequence begins with a first foreign image 208. This is followed by another foreign image 212, then a known image 504, then another foreign image 210, and finally another known image 504.
[0246] The client application is configured to display the individual images sequentially on the display 216 of the person 224 according to this randomly generated sequence. For example, the client application 240 can be configured to display each of the images 208, 212, 504, 210, 504 for two seconds and, after the two seconds have elapsed, to immediately display the next image in the sequence without a pause. For example, at time t1, the client application displayed the stranger image 208, 2 seconds later, at time t2, another stranger image of a different stranger, and again 2 seconds later, at time t3, an acquaintance image 504 of the person 224 who currently wishes to authenticate. The display 216 shows the acquaintance image of the person displayed on the display at the current time (between t3 and t4).When a further 2 seconds have elapsed after time t3, the familiar image 504 will be replaced by the further foreign image 210, and after a further 2 seconds at time t5 by the familiar image 504. While the image sequence 602 is displayed to the person 224, a corresponding sequence of first and second brain activity signals is recorded by the KLHA sensor and forwarded to the authentication module for further analysis and authentication of the person.
[0247] The display duration of each image typically ranges from one to five seconds, preferably between two and four seconds. Pauses may be introduced between the display of different images, typically shorter than five seconds.
[0248] The authentication module is configured to analyze the brain activity signals acquired by the KLHA sensor 218 during the display of various images to determine whether the one or more first brain activity signals received during the display of familiar images are significantly dissimilar to several second brain activity signals received during the display of unfamiliar images, wherein brain activity signals acquired during the display of facial images of the same type must preferably be identical or very similar. Depending on the type of acquired brain activity signals, this analysis can vary considerably. For example, the analysis can include determining and comparing the amplitude and / or amplitude profile over time ("signal profile") of a measured first and second signal. The signal can be an electrical signal (measured, for example,in volts) or a signal captured optically and by image analysis (measured, for example, in mm pupil diameter or pupil diameter change). Person 224 is considered to have been successfully authenticated to the authentication module if the authentication module determines that the first brain activity signals are significantly different from the second brain activity signals, wherein the sequence of the first and second brain activity signals corresponds to the sequence of familiar images and unfamiliar images, wherein the brain activity signals received when displaying images of the same type are preferably identical or very similar.
[0249] It is therefore not necessary in every case, for example, that the initial brain activity signals be identical or nearly identical in all people. However, the applicant has determined that at least a significant dissimilarity in the brain activity signals between familiar images and unfamiliar images is reproducibly present in virtually all people. Thus, according to embodiments of the invention, an authentication method can be provided that combines the advantages of biometric authentication methods with the advantages of challenge-response authentication methods.
[0250] After the person has successfully authenticated himself to the authentication module, further steps can be taken for the purpose of capturing a current acquaintance image 232, such as with regard to, for example, Figure 4already described. To ensure that the current known image is an image of the person's actual face and not an image of a mask worn by the person, the current known image 232 captured by the camera 242 is shown on the display 216 of the person 224. Meanwhile, the person's brain activity signals are captured by the KLHA sensor 218 and compared with the first and second brain activity signals captured when the sequence 206 was displayed. If the person is not wearing a mask and the current known image 232 is thus valid, it is expected that the captured brain activity signals are identical or similar to the first brain activity signals and simultaneously dissimilar to the second brain activity signals. If the expectation is correct, the current known image 232 is considered valid.This ensures that current acquaintance images, which can be used, for example, to personalise current identity documents, can actually be attributed to a specific person without the person having to visit the premises of a public authority, where it is normally the task of specially trained staff to ensure that a facial image used to personalise documents matches the person's real face.
[0251] Figure 7 shows a sequence 714 of first and second brain activity signals evoked by displaying the image sequence 602 in the person 224.
[0252] The display of the external image 208 generates a second brain activity signal 704. The display of the external image 212 generates a second brain activity signal 706. The display of the known image 504 following the image 212 generates a first brain activity signal 708. The display of the external image 210 generates a second brain activity signal 710. The display of the known image 504 following the image 210 generates a first brain activity signal 712. The dashed arrows here do not represent points in time, but rather time windows, for example 2-second time windows, during which the corresponding images are displayed and the brain activity signals corresponding to them are measured.
[0253] If the person 224 who currently wants to authenticate themselves as a specific person already registered with the server system is truly who they claim to be, it can be expected that all first brain activity signals 708, 712 are very similar or identical to one another, that all second brain activity signals 704, 706, 710 are very similar or identical to one another, that the temporal sequence of the first and second brain activity signals corresponds to the temporal sequence of the familiar images and unfamiliar images, and that the first and second brain activity signals are significantly different. In this case, the authentication module treats the person 224 as successfully authenticated. Otherwise, the authentication module treats the person as unauthenticated.
[0254] A practical application scenario according to an embodiment of the invention is described below: For example, the authentication method is used to remotely authenticate a citizen for the purpose of renewing an identity document. The population of the state in which the person is registered as a "citizen" comprises 80 million people, referred to here as "M1-M80,000,000." Ten facial images of each citizen are stored in a database, i.e., for citizen M1, the facial images B M1.1 , B M1.2 , ..., BM 1.10 , and for citizen M2, the facial images BM 2.1 -BM 2.10 . The first citizen M1 wishes to remotely authenticate himself and initiates an authentication method according to embodiments of the invention.
[0255] In the process, the authentication module selects a random sequence of two known images and four foreign images from the image database, e.g., according to the sequence B M44.7 , B M102.2 , B M1.3 , B M15.1 , B M1.8 , B M16.7 . In this sequence, images B M1.3 and B M1.8 are "known images" of citizen M1. The images of the other citizens M44, M102, M15, and M16 are foreign images of citizen M1, but can each serve as known images when these other citizens authenticate themselves. If the citizen currently authenticating is indeed citizen M1, the following signal sequence is to be expected, where the abbreviations "HAS1" stand for "first brain activity signal" and "HAS2" stand for "second brain activity signal": HAS2, HAS2, HAS1, HAS2, HAS1, HAS2.If the sequence of the recorded first and second brain activity signals corresponds to the signal sequence expected according to the sequence of facial images shown, citizen M1 has successfully authenticated himself.
[0256] Citizen M55 then wishes to authenticate himself and initiates another authentication process. During this process, the authentication module selects a random sequence of two known images and four foreign images from the image database, e.g., according to the sequence B M55.2 , B M14.7 , B M55.6 , B M232.2 , B M15.1 , B M46.7 . In this sequence, images B M55.2 and B M55.6 are "known images" for citizen M55. The images of the other citizens M14, M232, M15, and M46 are foreign images for citizen M55, but can each serve as known images if these other citizens authenticate themselves. If the citizen currently authenticating is indeed citizen M55, the following signal sequence is to be expected: HAS1, HAS2, HAS1, HAS2, HAS2, HAS2.
[0257] However, this example only serves to illustrate the technical principle. For data protection reasons, only facial images of people or citizens who have explicitly consented to this use would be used as third-party images, or computer-generated facial images that do not correspond to the face of a real person would be used.
[0258] Figure 8 shows a diagram of a neural network for analyzing and classifying brain activity signals. After a training phase, the neural network is able to determine whether and to what extent the first and second brain activity signals are similar to each other. It has been found that neural networks, especially recurrent neural networks 804, are particularly suitable for determining the similarity of brain activity signals.
[0259] For example, the machine learning software and / or the authentication module may have, in addition to the neural network 804, a preprocessing module 802 and an evaluation module 806. The preprocessing module 802 is configured to represent the signal detected by the KLHA sensor 218 in such a way that the information contained in the signal can be evaluated by the authentication module and / or the neural network. For example, the profile of temporal changes in amplitude (measured charge changes and / or charge redistributions in the vicinity of an electrical sensor) and optionally also frequency of the signal measured by an electrical sensor can be represented as a multidimensional vector, wherein, for example, the vector may include a plurality of elements (e.g.200 elements for a vector that specifies the signal properties of a signal measured over two seconds with a temporal resolution of 10 ms), wherein each element contains an amplitude value valid at that time and a frequency measured at that time in the vector. If the KLHA sensor is an optical sensor, the preprocessing module 802 can be configured to extract a plurality of features from the face image or eye area image captured by the optical sensor, such as pupil diameter and / or temporal profile of pupil diameter changes, and to represent these features in a suitable form, for example, a meaning also defined as a vector with a plurality of numerical values.In a training phase, the machine learning software 800 learned, based on an annotated training data set, to recognize which variability in signal properties can and must be accepted without two compared brain activity signals being considered significantly dissimilar. Biometric signals are never 100% identical. However, by training machine learning software on a data set of multiple brain activity signals, which were elicited once by viewing familiar images and another time by viewing unfamiliar images, and which are annotated accordingly, it is possible to ensure that the trained machine learning software can reliably recognize whether deviations in the properties of the brain activity signals are relevant or irrelevant with regard to the question of whether a familiar image or an unfamiliar image is being viewed.
[0260] The evaluation module 806 aggregates the outputs of the neurons in the output layer of the neural network and prepares the result (authenticated or not) so that it can be interpreted by the receiver. For example, the result can be output to a human as a natural language text, either acoustically or visually. Alternatively or additionally, the authentication result can also be output to a software or hardware component, which, depending on the authentication result, enables or disables a specific hardware or software function.
[0261] According to another embodiment not shown here, a convolutional neural network (CNN) is used to classify brain activity signals into similar and dissimilar signals. It can, for example, include a first convolutional layer with six feature maps whose kernel size is 5. The first convolutional layer is a first maxpooling layer with a scale size of 2. A second convolutional layer with 12 feature points has a kernel size of 3. The neural network also includes a second maxpooling layer with the same scale size as the first maxpooling layer. Finally, there is a fully connected layer with 108 feature points that calculates the score estimate of the input. In the training stage, the learning rate, loss size, and learning epoch are set to, for example, 1, 100, and 500.
[0262] Figure 9shows a plot 900, i.e., a diagram, with a first brain activity signal 902, which was measured by an electrical sensor during the viewing of a familiar image, and with a second brain activity signal 904, which was measured by an electrical sensor during the viewing of an unfamiliar image. The brain activity signals shown here are brainwave signals that were recorded contactlessly in the form of charge changes and / or charge redistributions in the vicinity of the electrical sensor.
[0263] In neurology, different frequency bands are generally distinguished when measuring brain waves: The sub-δ band with a frequency range of 0.15 to 0.5 Hz, the δ band with a frequency range of 0.5 to 3.5 Hz, the θ (theta) band with a frequency range of 3.5 to 8 Hz, the α band with a frequency range of 8 to 13 Hz, the β band with a frequency range of 13 to 30 Hz, as well as frequency components above the β band (i.e. above 30 Hz).
[0264] According to embodiments, the brain activity signals are measured in one or more of these bands.
[0265] In some embodiments, the α band provides additional information about a person's state of alertness. Frequency components above this α band (β band and above) can provide information about impairment of mental functions by drugs or centrally acting medications. The theta band, located directly below the α band, provides other types of information that can correlate with, among other things, the state of alertness, ingested substances, or illnesses. Specific stimuli, such as facial images, often trigger brainwave signals in the theta band and other bands. According to embodiments of the invention, a "multi-band" signal analysis, i.e., an analysis of the properties of signals across a very broad frequency range, is performed.
[0266] According to further embodiments, the brain activity signals are measured without differentiation into individual bands. Rather, the signal amplitude is recorded over a wide frequency band, which is essentially determined only by the sensitivity of the electrical sensor.
[0267] Figure 10 shows a block diagram of a further authentication system 200.3 according to an embodiment of the invention. The structure of the authentication system 200.3 essentially corresponds to the structure of the authentication systems 200.1, 200.2, which are exemplified by the Figures 4 and 5 Person 224 is not part of the authentication system. Figure 10shows how the authentication system can be used by a system 250 for personalizing documents, e.g., identity documents, valuable documents, and / or security documents. For example, the system 250 can be a system for personalizing identity cards, which are designed as chip cards based on a plastic or paper carrier.
[0268] Person 224 first registers in the secure premises of a government agency to obtain an identity card for the first time. This involves capturing a facial image of this person for the first time and saving it as an acquaintance image. Because the image is captured in secure premises by specially trained personnel and / or with a particularly secure camera, it is prevented that the facial image captured during registration depicts a person other than the person who appeared in person for registration. Years later, the person then successfully authenticates remotely using a smartphone and a linked KLHA sensor, as described for embodiments.
[0269] Following successful authentication, a further step is taken in an unsecured environment (e.g., no specially trained personnel are present, no camera with a special signature function is used), capturing a current facial image of the person, e.g., using a standard smartphone camera. The facial image is displayed to the person again on the smartphone's display 216. While the person views the current facial image 232 on the display, the person's brain activity signals are captured and compared with the first and second brain activity signals measured during the authentication process. Only if the captured brain activity signals are similar to the first brain activity signals and dissimilar to the second brain activity signals is the current facial image used to print a new identity card 222 with it, thus personalizing it.
[0270] Figure 11shows a plot 950 with a first brain activity signal 952 in the form of a pupil diameter profile, which was measured by an optical sensor during the viewing of a familiar image (here: an image of the viewer's face), and with a second brain activity signal 904 in the form of a second pupil diameter profile, which was measured by the optical sensor during the viewing of an external image. The two pupil diameter profiles, or rather the respective diameters, were determined by automatic image analysis of the captured images. In contrast to the viewing of an external image, viewing a familiar image leads to a significant increase in the pupil diameter within the first 1000 ms in the person viewed here.
[0271] The authentication method according to the invention is not limited to the described embodiments. It can be combined with other authentication methods to enable secure and convenient authentication. A combination of several biometric methods is advantageous to ensure multimodal authentication. List of reference symbols
[0272] 102-114 Steps 150-160 Steps 200 Authentication system 201 Portable telecommunications device 202 Server computer system 203 (Secure) camera 204 Authentication module 205 Server application 206 Image database 207 Server-side data storage 208 Facial image (third-party image) 209 Client-side temporary data storage 210 Facial image (third-party image) 212 Facial image (third-party image) 214 Facial image (third-party image) 216 Display 218 KLHA sensor(s) 219 GUI 222 Document 224 Person 230 Known image 232 Known image 233 Authentication request 234 Specification of an image sequence 235 Server-side data storage 236 Client-side data storage 240 Client application 242 Camera 262 Secure room 504 Facial image (known image) 602 Sequence of facial images 702 Sequence of brain activity signals 704 Second brain activity signals 706 Second brain activity signals 708 First brain activity signals 710 Second brain activity signals 712 First brain activity signals 800 Neural network-based authentication module 802 PreprocessingModule 804 Recurrent neural network 806 Evaluation module 900 Signal plot recorded by electrical sensor 902 First brain activity signal (when displaying a familiar image) 904 Second brain activity signal (when displaying an unrelated image) 950 Signal plot recorded by optical sensor 952 First brain activity signal (when displaying a familiar image) 954 Second brain activity signal (when displaying an unrelated image)
Claims
1. A method comprising the steps of: - generating (106) a specification (234) of a sequence of known images and foreign images (208-214), wherein a known image is an image containing an image motif that is known to a person and a foreign image is an image containing a motif that is not known to a person; - displaying (108) a plurality of images on a display (216), wherein the displayed images include a sequence of one or more known images and one or more foreign images according to the specification; - whilst the plurality of images are being displayed to the person, detecting (110), by a sensor (218), first brain activity signals (708, 712, 902, 952) and second brain activity signals (704, 706, 710, 904, 954) that are triggered in the person as a result of looking at the images, wherein first brain activity signals are detected whilst a known image (230) is displayed to the person, and wherein second brain activity signals are detected whilst a foreign image is displayed to a person; characterized in that the method is a method for authenticating a person (224), in that the display (216) is associated with a portable telecommunications device (201) which is assigned to a person, in that the generation (106) of the specification (234) of the sequence of known images and foreign images (208-214) is performed by an authentication module (204), in that the authentication module is part of a server application or a client application, and in that the display (108) of the plurality of images on the display (216) of the portable telecommunications device (201) is performed by the client application outside a secure premises, wherein the method further comprises the steps of: - registering (102) the person with the server application (205), wherein the server application is instantiated on a server computer system (202), wherein the registration comprises capturing at least one known image (230) of the person within secure premises (262) when the person is present in the secure premises; - receiving (104) the at least one known image from the server application via a network connection by the client application (240), wherein the client application is instantiated on the telecommunications device (201), wherein the telecommunications device is connected reversibly or irreversibly to the sensor (218), wherein the sensor (218) is a contactless sensor (218) for detecting brain activity and is referred to hereinafter as a KLHA sensor (218); - authenticating (112) the person as a function of the similarity of the detected first and second brain activity signals under consideration of the specification of the image sequence by the authentication module, wherein the method for authenticating a person treats the person as having been successfully authenticated if the first brain activity signals are similar to one another, if the second brain activity signals are similar to one another, if the first brain activity signals are dissimilar to the second brain activity signals, and if the sequence of similar first brain activity signals and similar second brain activity signals corresponds to the sequence of displayed known images and foreign images; and - outputting (114) an authentication result in relation to the person.
2. The method according to claim 1, wherein the at least one known image is received by the client application during the course of the registration of the person, wherein the method preferably further comprises the steps of: - saving (150) the at least one received known image in a protected way in a non-volatile memory (236) of the telecommunications device; - sending (152) a receipt confirmation in relation to the known image from the client application to the server application following successful receipt and successful saving of the known image; and - deleting (154) all other existing copies of the known image by the server application in response to the receipt of the receipt confirmation.
3. The method according to claim 1, wherein the at least one known image is received by the client application during the course of each authentication process for the person, wherein each authentication process comprises the steps of: - receiving an authentication request for the person from the client application by the server application; - in response to the receipt of the authentication request, identifying, by the server application, the at least one known image that was detected during the registration of the person; - sending the identified at least one known image to the client application by the server application via the network connection.
4. The method according to claim 1, 2 or 3, further comprising the steps of: - saving (160) the at least one known image detected during the registration of the person in a non-voltage memory of the server computer system by the server application in a protected way; - saving the at least one received known image on a memory of the telecommunications device exclusively temporarily for the duration of the authentication process.
5. The method according to any one of the preceding claims, wherein the at least one known image is an image selected from a group comprising: - a facial image of the person; - an image, captured by means of a camera or video camera (203), of an image motif demonstrably known to the person; - a multiplicity of images, captured by means of a camera or video camera (203), of an image motif demonstrably known to the person; - a multiplicity of copies of an image, captured by means of a camera or video camera (203), of an image motif demonstrably known to the person.
6. The method according to any one of the preceding claims, wherein the registration of the person further comprises the steps of: - providing at least one biometric sensor within the secure premises, wherein the sensor is configured in particular as a sensor for distinguishing people wearing a face mask from people not wearing a face mask; - capturing a biometric feature of the person by the biometric sensor; - evaluating the captured biometric feature; - if the evaluation of the captured biometric feature reveals that the person is not wearing a mask, performing the registration and providing the at least one known image from the server application to the client application; - if the evaluation of the captured biometric feature reveals that the person is wearing a mask, aborting the registration without capturing the at least one known image and / or without providing the at least one known image to the client application.
7. The method according to any one of the preceding claims, wherein at least one known image, which is captured in the secure premises, is captured by a security camera (203), wherein a security camera is a camera selected from a group comprising: - a camera with access protection, wherein the access protection causes the camera to be able to be used by an operator only following a prior authentication step to the security camera; and / or - a camera with a signature key which is configured to sign the known images captured by said camera, wherein the server application has a signature verification key, which is configured to verify the signature of images and, during the course of the registration process, to accept as known images only those images that have been verified and signed by the security camera.
8. The method according to any one of the preceding claims, - wherein the KLHA sensor includes a signature that was generated by a KLHA signature key; - wherein the client application includes a KLHA signature verification key corresponding to the KLHA signature key; and - wherein the client application is configured to prompt the detection of the first and second brain activity signals and / or use thereof for the authentication only if a verification of the signature of the KLHA sensor with the KLHA signature verification key reveals that the KLHA sensor signature is valid.
9. The method according to any one of the preceding claims, wherein the sequence of known images and foreign images serves as a challenge in the sense of a challenge-response method, and wherein the authentication, as a function of the similarity of the detected first and second brain activity signals, comprises a check as to whether the chronological order of the first and second brain activity signals reflects the order of known images and foreign images coded in the challenge.
10. The method according to any one of the preceding claims, wherein the authentication module treats the person as having been successfully authenticated only if, besides a significant difference in the first and second brain activity signals, the following further criteria apply: - all first brain activity signals detected during the display of any one of the known images are identical or sufficiently similar to one another; - all second brain activity signals detected during the display of any one of the foreign images are identical or sufficiently similar to one another.
11. The method according to any one of the preceding claims, - wherein the KLHA sensor is an electrical sensor; and • wherein the electrical sensor is configured to detect the first and / or second brain activity signals by measuring charge changes and / or by detecting charge redistributions; and / or • wherein the electrical sensor is configured to detect the amplitude of the first and / or second brain activity signals and / or the change in amplitude during a time interval; or - wherein the KLHA sensor is an optical sensor which is configured to detect images or image sequences of the face or of parts of the face of the person that contain the first and / or second brain activity signals, wherein the authentication, as a function of the similarity of the detected first and second brain activity signals, comprises an automatic image analysis of the images or image sequences by the authentication module, wherein the image analysis comprises: - computationally determining the pupil diameter of the person; and / or - computationally determining a pupil diameter change profile of the person.
12. The method according to any one of the preceding claims, wherein the telecommunications device includes a camera (242), wherein the at least one known image that was detected during the registration is a facial image of the at least one person, said method further comprising: - if the authentication result includes the fact that the person has successfully authenticated themself: • displaying a GUI (129) of the client application on a display of the telecommunications device, wherein the GUI includes a request to hold the face of the person in the detection range of the camera of the telecommunications device; • capturing a current facial image (232) of the person; sending the current facial image via the network connection from the client application to the server application; said method preferably further comprising: • saving and using the current facial image (232) of the person as the at least one known image of the person instead of the at least one known image (230) that was captured during the registration of the person; and / or • providing the current facial image (232) of the person from the server application to a software (250) for personalising identity or security documents (222) in order to personalise an identity or security document for the person.
13. A method for producing an identity or security document for a person, said method comprising the steps of: - authenticating the person by a method according to claim 12; - producing an identity or security document (222) which is uniquely assigned to the person, wherein the current facial image of the person that was transmitted to the server application is used as security element (270) of the identity or security document.
14. A system (200), characterised in that the system (200) is a system (200) for authenticating a person (224), wherein the system (200) comprises: - a server computer system (202) with a server application (205), wherein the server application includes a register containing the profiles of a plurality of people registered with the server application, wherein each profile is stored linked with at least one known image (230, 232), wherein the at least one known image was captured within a secure premises (262) when this person was present in the secure premises, wherein a known image of a person is an image containing an image motif known to this person; - a portable telecommunications device (201), which is reversibly or irreversibly connected to a sensor (218), wherein the sensor is a contactless sensor (218) for detecting brain activity and is referred to hereinafter as a KLHA sensor (218), and on which a client application is instantiated which is interoperable with the server application, wherein the telecommunications device is assigned to the person; - an authentication module (204), which is part of the server application or the client application; wherein the client application is configured to receive the at least one known image from the server application via a network connection; wherein the authentication module is configured to generate a specification (234) of a sequence of known images and foreign images (208-214), wherein a foreign image is an image containing a motif that is not known to the person; wherein the client application is configured to display (108) a plurality of images on a display (216) of the telecommunications device outside the secure premises, wherein the displayed images include a sequence of one or more known images and one or more foreign images according to the specification; wherein the KLHA sensor is configured, whilst the plurality of images are being displayed to the person, to detect first brain activity signals (708, 712, 902, 952) and second brain activity signals (704, 706, 710, 904, 954) that are triggered in the person as a result of looking at the images, wherein first brain activity signals are detected whilst a known image (230) is displayed to the person, and wherein second brain activity signals are detected whilst a foreign image is displayed to a person; - wherein the authentication module is configured to authenticate the person as a function of the similarity of the detected first and second brain activity signals under consideration of the specification of the image sequence; and to output an authentication result in relation to the person, wherein the authentication module treats the person as having been successfully authenticated if the first brain activity signals are similar to one another, if the second brain activity signals are similar to one another, if the first brain activity signals are dissimilar to the second brain activity signals, and if the sequence of similar first brain activity signals and similar second brain activity signals corresponds to the sequence of displayed known images and foreign images.
