Method and arrangement for preventing image recordings of content to be protected
The method uses sensors and AI to detect and prevent unauthorized recording of protected content by identifying camera presence and orientation, enhancing detection accuracy and preventing unauthorized reproduction of sensitive information.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- SIEMENS AG
- Filing Date
- 2025-12-15
- Publication Date
- 2026-06-25
AI Technical Summary
Existing methods to prevent unauthorized recording of protected content, such as screenshots or external camera captures, are inadequate, particularly when users can circumvent privacy filters and flash-based countermeasures, allowing partial or complete capture of sensitive information.
A method utilizing sensors and AI modules to detect and prevent unauthorized recording by identifying the presence and orientation of cameras, employing artificial infrared light sources, flash devices, and AI models like LLMs to analyze sensor data for subtle patterns indicative of recording intent, and trigger suppression actions.
Effectively prevents unauthorized recording by accurately detecting camera presence and orientation, reducing unnoticed captures, and ensuring protected content is not reproduced when unauthorized recording is detected, while minimizing disruption to authorized users.
Smart Images
Figure EP2025087164_25062026_PF_FP_ABST
Abstract
Description
[0001] 202416212
[0002] 1
[0003] Description
[0004] Procedures and arrangements for protection against the recording of images containing protected content
[0005] The invention relates to a method for protection against image recordings of protectable content according to the preamble of claim 1 and an arrangement for protection against image recordings of protectable content according to the preamble of claim 17.
[0006] It is known that users of computer systems are generally able to record screen content, for example by taking a screenshot. This can infringe on protected content that affects the rights of third parties, such as data sovereignty policies, know-how protection, data privacy, and / or personality rights.
[0007] It is known that screenshots of the screen content can be prevented via software settings in the operating system. However, this does not prevent users from taking a picture with an external camera.
[0008] Known technical measures designed to prevent recording by external cameras include:
[0009] - Privacy filters that protect against side photographs
[0010] - Backflashes, which, upon detecting flashes triggered by a camera belonging to unauthorized persons, cause overexposure by means of a backflash.
[0011] A disadvantage of privacy filters is that they only offer protection against frontal shots to the extent that a user sitting in front of the screen obscures it. If the screen is at least partially visible from the front, data can be captured, at least partially.
[0012] The solution based on counter-flashes only protects insofar as flashes are detected, so that, depending on the given lighting conditions, at least partial recognition of content cannot be ruled out by taking pictures without a flash, which consequently cannot be detected.
[0013] The object of the invention is to offer a technical solution that overcomes the disadvantages of the prior art, in particular to provide a method and an arrangement that prevents unauthorized recording of copyrighted content. 202416212
[0014] 2
[0015] This problem is solved starting from the method according to the features of the preamble of claim 1 by its characterizing features, as well as by the arrangement according to the features of the preamble of claim 17 by its characterizing features.
[0016] Unless otherwise specified in the following description, the terms "perform," "implement," "transform," "transmit," "calculate," "computer-aided," "compute," "determine," "generate," "configure," "reconstruct," "ascertain," "capture," and the like—where mentioned—preferably refer to actions and / or processes and / or processing steps that modify and / or generate data and / or convert data into other data, wherein the data may be represented or exist as physical quantities, for example, as electrical impulses. In particular, the term "computer" should be interpreted as broadly as possible to encompass all electronic devices with data processing capabilities.Computers can therefore be, for example, personal computers, servers, programmable logic controllers (PLCs), handheld computer systems, pocket PC devices, mobile phones and other communication devices that can process data using a computer, processors and other electronic devices for data processing.
[0017] The term “computer-aided” or “computer-supported” – where mentioned – may, for example, refer to an implementation of the method in which, in particular, a processor performs at least one process step of the method.
[0018] In the context of the invention, a processor—where mentioned—can be understood to mean, for example, a machine or an electronic circuit. A processor can be, in particular, a central processing unit (CPU), a microprocessor, or a microcontroller. A processor can also be, for example, an integrated circuit (IC), especially an FPGA (field-programmable gate array) or an ASIC (application-specific integrated circuit), or a digital signal processor (DSP), or a graphics processing unit (GPU). A processor can also be understood to be a virtualized processor, a virtual machine, or a soft CPU. It can also be, for example, a programmable processor, such as the one described in 202416212.
[0019] 3
[0020] is equipped with configuration steps for carrying out the said method according to the invention, or is configured with configuration steps such that the programmable processor realizes the features of the method, the component, the modules, or other aspects and / or partial aspects of the invention according to the invention.
[0021] In connection with the invention, a “storage unit”, “storage device” or “storage module” and the like – if mentioned – can be understood to mean, for example, volatile memory in the form of random-access memory (RAM) or permanent memory such as a hard drive or data carrier, as well as a combination of one or more of these elements to form a so-called cloud.
[0022] In the context of the invention, a "module" can be understood to mean, for example, at least one processor and / or at least one memory unit for storing program instructions, which are physically connected at one location, for example, a part of a printed circuit board, and functionally interact with one another. For example, the processor is specifically configured to execute the program instructions in such a way that the processor performs functions to implement or realize the method according to the invention or a step thereof.
[0023] In the inventive method for protection against the recording of images of protected content, the following steps are performed: a) at least parts of an area allowing a view of the protected content are detected by at least one sensor; b) at least parts of the detected sensor data are forwarded to a cls module; c) the cls module evaluates the sensor data to determine data that correlates with at least the presence and / or orientation of a first camera, for example, a digital camera, particularly one integrated into a mobile phone; d) a suppression action is carried out depending on the data evaluation such that the recording of the protected content is prevented if the data evaluation at least indicates an unauthorized recording process.
[0024] The steps of the method according to the invention describe a solution that overcomes the disadvantages of the prior art. This is achieved, among other things, through the use of the sensor. This sensor provides a means of monitoring the environment in which the protected content is reproduced, specifically that part of the environment in which it is located.
[0025] 4. It is possible to use a camera to record the content without authorization. This further reduces the possibilities for unnoticed recording described at the beginning. Furthermore, the combination of sensor-acquired data and the AI module allows the monitoring to contribute to accurate results, meaning that, in principle, sensors with any characteristics can be used, since an AI module specifically tailored to the sensor, and especially trained for it, can improve the evaluation of the sensor data, particularly the reliability of the alert.
[0026] The arrangement according to the invention for protection against image recordings of a protectable content is characterized by means for carrying out the method and / or one of its further developments, thereby contributing to the implementation and mutatis mutandis to the realization of the advantages mentioned in connection with the method for protection against image recordings of a protectable content.
[0027] Further advantageous embodiments and developments of the invention are specified in the dependent claims.
[0028] According to a first embodiment of the inventive method, the sensor is operated in such a way that an artificial infrared light source can be detected. This allows the detection of cameras that emit a minimal amount of light to ensure sufficient brightness. This emission is generally detectable as a point-like infrared source. In combination with the AI module, the mere presence of such a pattern and / or a change in the position of this point-like source detected over time can then serve as the basis for indicating unauthorized recording.
[0029] Alternatively or additionally, the method according to the invention can be further developed in such a way that a flash device is operated in such a way that it is triggered at least temporarily, in particular at recurring times, wherein the sensor is operated synchronously with the triggering times in such a way that reflections of the flash are detected.
[0030] This further development utilizes the sensor-Kl combination according to the invention to recognize patterns in the reflection that indicate a lens ready to record and / or the alignment of a lens with the content requiring protection, and ensures, through the flash device, that a known light source with flashes triggered at known times is known as a fixed quantity when determining the reflection during an evaluation by the Kl module, so that the Kl module can detect the reflections of random light reflections.
[0031] 5 can be distinguished. In addition to other approaches, such as the detection of infrared light by the sensor, it can, for example, allow the area to be narrowed down by indicating a point-like infrared source for recording.
[0032] Preferably, the method according to the invention is further developed in such a way that the detection is carried out in such a way that at least one lens of a second camera is operated as a sensor such that the camera lens at least temporarily in a first period of time, in particular at discrete times, takes at least one picture of the area, wherein the picture is forwarded to the AI module as sensor data, wherein the first period of time corresponds at least partially to a given reproduction of the content.
[0033] Firstly, this advanced training allows the use of common or already existing peripheral devices such as cameras, thus reducing the costs of using the method. Secondly, it utilizes the ability of these cameras to make the aforementioned camera light sources visible as bright points in an image. This works particularly well when the image is captured in a dark room, but thanks to the use of the AI module, it also works in brighter environments, since an AI module can detect changes in pixel brightness that are imperceptible to the human eye and is therefore able to generate an indication of point-like infrared sources.
[0034] Preferably, the method according to the invention can be further developed such that at least one second camera, operating the sensor, is at least functionally connected to a device reproducing the protected content and is operated for at least temporary purposes. Due to the functional connection, the control, in particular by a processor of the device, can be carried out jointly, and thus at least the acquisition step, if not even the step of calling the AI module and / or the evaluation, can be carried out by the control unit of the reproducing device.
[0035] In a further embodiment of the method according to the invention, the second camera is positioned and operated in such a way that, relative to the content to be protected, it can capture at least parts of that content. That is, in this embodiment, the second camera, or at least one of the cameras, is not necessarily part of the display device. This embodiment is advantageous, for example, when the camera is an external peripheral device and / or is used in an environment where, for example, surveillance cameras are installed that can contribute to detection. 202416212
[0036] 6
[0037] Alternatively or additionally, the method according to the invention can be further developed in such a way that the second camera is placed in relation to the device and, in particular, is physically connected to it and operated in such a way that the at least one lens opening is arranged in the same plane as the reproduction of the content worthy of protection.
[0038] Preferably, the method according to the invention is further developed in such a way that the determination of data correlating at least with the presence and / or orientation of an image recording device is carried out in such a way that the AI module operates a trained, in particular multimodal, LLM in such a way that it is called with a prompt formed at least from the sensor data in such a way that a prediction about the presence and / or orientation is output.
[0039] Training can detect even very subtle patterns, imperceptible to humans, that indicate a recording process, thus improving prediction accuracy. This training can be done beforehand, but can also be continued during operation through fine-tuning, such as providing feedback on incorrect or correct indications.
[0040] The AI module can alternatively or additionally operate other generative AI models, either as an alternative or in combination with an LLM. For example, the use of generative AI models that generate videos, images, and / or music is also conceivable. In particular, so-called discriminative AI models are also suitable as an alternative or additional tool for recording recognition, since they are dedicated to classification and recognition and are therefore very well trained, at least for the recording analysis purpose according to the invention. LLMs, especially multimodal ones, offer diverse application possibilities, not least due to their popularity, and can therefore be used for the purpose of the invention cost-effectively and with little effort.
[0041] According to a further development of the inventive method, the LLM can also be carried out in such a way that it is trained to predict a camera lens of the first camera that at least partially captures the area of the protected content. This ensures that content blocking is only triggered when there is a recording capability that captures at least parts of the protected content. Thus, recording devices not directed at the content would be excluded, and unjustified interruptions of the display would be avoided. 202416212
[0042] 7
[0043] Alternatively or additionally, the method according to the invention can be further developed such that the LLM is trained in such a way that the prediction of the area is carried out at least by means of a reflection of light, in particular emitted by a stationary source, for example the flash unit, detected from the image acquisition. With this further development, it is possible to use reflections caused, for example, solely by the given lighting conditions to detect camera lenses even without a fixed source, since, as mentioned, a light source can detect subtle changes at the pixel level. With a fixed source such as the aforementioned flash unit, this property can be used advantageously in conjunction with the flash unit in such a way that the flash is not set to a particularly high brightness.It can be very weak and therefore more pleasantly designed, yet still lead to accurate recognition, and this further training can also be provided as an alternative or supplement.
[0044] Due to the ability to detect the most subtle changes at the pixel level of an image or patterns in the sensor data, a further development of the inventive method is advantageous in which the LLM is trained to predict the operating state of the camera lens of the first camera. In other words, it is trained to infer that the camera is currently filming simply by detecting the lens.
[0045] Alternatively or additionally, the method according to the invention can be further developed in such a way that the LLM is trained in such a way that the prediction of the operating state is carried out at least on the basis of infrared points detectable in the image recording.
[0046] A further alternative or supplementary development is given if the inventive method is further developed in such a way that a further camera, disjoint from the second camera and connected with a device for control, its control and / or image evaluation, is operated in such a way that it calls the LLM with recordings in such a way that the LLM makes a prediction about the presence and / or orientation of the first camera.
[0047] This is particularly advantageous when using surveillance cameras already present in the room, such as those found in industrial environments. These cameras can complement the second camera and, for example, even if the view of the first camera from the second camera might be obstructed (e.g., by an operator), they can still trigger the blocking of the display of the content to be protected before the 202416212
[0048] 8
[0049] When the operator moves aside, the content is revealed, and only now can the second camera record the first camera.
[0050] Preferably, the method according to the invention is further developed in such a way that the suppression action is carried out in such a way that the device reproducing the content worthy of protection is manipulated in such a way that, as long as there is an indication of an unauthorized recording, the reproduction of the content to be protected does not occur.
[0051] This provides a simple and quick method to prevent the content from being recorded.
[0052] In a further development, the method according to the invention can be further developed such that the suppression action is carried out with the following steps: a) Identifying a communication link to the first camera, b) Forming a message that correlates with at least one control signal, c) Sending the correlated message, d) Receiving the correlated message by a device designed to influence the function of the first camera, e) Manipulating the function by the influencing device in such a way that the formation of a recording reproducing the content is prevented.
[0053] This training is particularly beneficial where suppression functions are planned. For example, this training could be suitable if smartphones are designed in such a way that, to protect privacy, a manipulation function is included, allowing the camera and / or the general functionality of the smartphone to be blocked by external entities.
[0054] If the inventive method is further developed such that, upon detection of unauthorized recording, at least one warning signal is issued visually, in particular on the playback device, and / or audibly via a loudspeaker device, authorized personnel can intervene as appropriate and, for example, clarify whether an unauthorized recording is taking place, and then take appropriate measures. 202416212
[0055] 9
[0056] Further advantages and details of the invention, as well as further developments of the invention, are explained in more detail below with reference to an exemplary embodiment shown in the single figure. It shows the
[0057] Figure (FIG) schematically illustrates an exemplary process of the computer-implemented procedure for testing a user experience of at least a part, in particular the user interface, of a software.
[0058] The embodiment described below in the figure (FIG) is a preferred embodiment, the advantages of which, as well as further embodiments or developments of the invention, are explained in more detail.
[0059] In particular, the following explanations merely show exemplary implementation possibilities of how such implementations of the teaching according to the invention could look, since it is impossible and also not helpful or necessary for understanding the invention to name all these implementation possibilities.
[0060] Furthermore, a person skilled in the art, with knowledge of the independent claims, will of course be aware of all the possibilities for realizing the invention that are customary in the prior art, so that in particular there is no need for a separate disclosure in the description.
[0061] In the exemplary embodiment(s), the described components of the embodiments each represent individual features of the invention that can be considered independently of one another, which further develop the invention independently of one another and can therefore be regarded as part of the invention individually or in a combination other than that shown.
[0062] Furthermore, the described embodiments can also be supplemented by further features of the invention already described.
[0063] Figure FIG schematically shows both exemplary embodiments of the arrangement according to the invention and a flowchart as an exemplary embodiment of the method according to the invention. 202416212
[0064] 10
[0065] On the left side of the figure, a schematic and simplified embodiment of the arrangement according to the invention is shown.
[0066] The image shows a laptop (L) on which protected content can be displayed. It also shows an operator (BP) who, for example, would access such content during a demonstration for a group of visitors (BG).
[0067] It is further indicated that a person from the visitor group BG points a smartphone with a first camera K1 at the laptop L, so that the laptop L is recorded.
[0068] Since the visitor group BG is located in an area B, B*, from which the content on the laptop L could be captured using a camera K1, the operation of a second camera K2, K* is provided for monitoring the area B, B* according to the embodiment of the arrangement, but also the embodiment of the method.
[0069] In this process, the second camera K2, connected to the laptop L according to a first embodiment, can be used. This will typically be a camera integrated into the laptop. However, it can also be an external, alternatively or additionally connected peripheral camera. This camera is then arranged such that its lens(es) is in the same plane as the content to be displayed, i.e., the screen surface of the laptop L, or parallel to it. From this plane, a first area B can be captured directly opposite the content. This makes it possible to directly capture the lens(es) of the first camera K1 of the smartphone. This is also advantageous if smartphones are not used, but only a microcamera, of which only the microlens would be visible.
[0070] Alternatively or additionally, and therefore shown with dashed lines, a (further) second camera K2* can be offset to record a second area B*. For example, this second camera K2* could be a surveillance camera in an industrial environment IU, as illustrated in the exemplary embodiment by the communication with the factory pictogram IU.
[0071] In an industrial environment (IU), the use of smartphones or cameras (K1) is generally prohibited. This applies at least to visitor groups (BG). In such an environment, it is therefore sufficient to merely detect the presence of a smartphone with a camera (K1). Therefore, monitoring from an offset point, i.e., not from a directly opposite point, would be conceivable, either alone or as a supplement, according to the invention. 202416212
[0072] 11
[0073] During surveillance, recordings A, A* of the first area B and / or second area B* are made. If only the second camera K2 of the laptop L is implemented, this will be a first recording A; if another second camera K2* is used alternatively or additionally, a second recording A* will be created.
[0074] The recordings can be made as a video recording, i.e., a continuous sequence of recordings A, A*, and / or generated at discrete time intervals as individual image recordings A, A*.
[0075] After generation, the recordings A, A* are passed to a Kl module. This Kl module operates a Kl model, for example, a multimodal LLM. A pre-trained standard LLM, such as Chat-GPT, can be used and trained to examine the recordings for indications that a camera and / or camera lens is pointed at the laptop L.
[0076] Camera K2*, assigned to an industrial environment IU, would typically send the recording to a control unit also assigned to the industrial environment IU. Thus, in the example shown, the recordings A* from this control unit would be sent to the Kl module via the industrial environment IU. It is also conceivable that the Kl module Kl is integrated into or assigned to a control unit within the industrial environment IU, and / or integrated into the laptop L, or implemented through its control system, for example, by providing a software client that enables access to an LLM.
[0077] Since conventional LLMs would capture all pixels, they are capable of detecting subtle changes that indicate a camera (lens) is ready to record or is recording. Here, training with feedback using labels to indicate whether a positive signal, i.e., a camera indication, is correct would be necessary.
[0078] However, it is also conceivable to use individually trained LLMs. According to the invention, the training will be carried out on at least one of the following indicators or a combination thereof:
[0079] A change in brightness, especially in images taken in infrared mode, in a very small, particularly point-like area; reflections; objects that point to a smartphone, a camera, a camera lens; unusual 202416212
[0080] 12
[0081] Movements / postures of people who are opposite the laptop L, etc., whereby for the investigation of reflections it is also conceivable that the arrangement is supplemented by a flash device that generates flashes synchronized with the second cameras K2, K2*, so that one is not dependent on random light sources for reflections and there are fewer degrees of freedom in the evaluation, so that the use of reflections for evaluation can deliver more accurate results.
[0082] Furthermore, due to the use of the device according to the invention, the brightness of the flash can be reduced to such an extent that very subtle flashes are produced, because the device can also detect reflections that are barely perceptible to the human eye, so that this does not disturb the persons present BP, BG.
[0083] If the AI trained in this way detects such a clue, it generates a control signal SSG or a message generating a control signal in the laptop, according to the example, which causes an alarm to be triggered and the display on the screen to be suppressed, i.e. the screen of the laptop L to be switched off.
[0084] However, the invention is not limited to the use of LLM as AI; rather, the AI module according to the invention mentioned in this disclosure can also comprise one or more alternative implementations that realize an artificial intelligence capable of performing the functions mentioned.
[0085] An embodiment of the method according to the invention could therefore be schematically summarized as follows:
[0086] Starting from a first state Z1, in which the laptop L is running or, for example, the demonstration is taking place, in a first step S1 a recording A of a first area B or a further recording A* of a second area B* from other angles is generated.
[0087] In a second step, S2, the recordings A, A* are then forwarded to a Kl module. This can be an image sequence, especially video recordings, individual images, and / or continuous recording, so that the Kl module forwards this image sequence to an LLM, at least as part of a prompt, and generates a prediction as to whether unauthorized recording / camera use is occurring. 202416212
[0088] 13
[0089] In a third step, S3, it is then checked whether such a prediction exists. If not, the previous steps S1-S2 are repeated continuously or at discrete time intervals. If it does, an alarm is triggered in a fourth step. This can be audible via a loudspeaker and / or visual with an alarm light, or simply due to the laptop screen L switching off / dimming in response to the warning, leading to a second state, Z2, in which the screen remains dimmed until an all-clear is given, for example, by the operator.
[0090] Regardless of the grammatical gender of a particular term, persons with male, female or other gender identities are included.
Claims
202416212 14 Patent claims 1. A method for protection against the recording of images of protected content, characterized by the following steps: a) capturing at least parts of an area allowing a view of the protected content by at least one sensor, b) forwarding at least parts of the captured sensor data to a control module, c) evaluating the sensor data by the control module to determine data that correlates at least with the presence and / or orientation of a first image recording device, for example, a digital camera, in particular integrated into a mobile phone, d) performing a suppression action depending on the data evaluation such that the recording of images of the protected content is prevented if the data evaluation at least indicates an unauthorized recording process.
2. Method according to one of the preceding claims, characterized in that the sensor is operated in such a way that an artificial infrared light source can be detected.
3. Method according to one of the preceding claims, characterized in that a flash device is operated in such a way that it is triggered at least temporarily, in particular at recurring times, wherein the sensor is operated synchronously with the triggering times in such a way that reflections of the flash are detected.
4. Method according to the preceding claim, characterized in that the detection is carried out in such a way that at least one lens of a second camera is operated as a sensor such that the camera lens at least temporarily in a first period of time, in particular at discrete times, takes at least one picture of the area, wherein the picture is forwarded to the AI module as sensor data, wherein the first period of time corresponds at least partially to a given reproduction of the content.
5. Method according to the preceding claim, characterized in that the second camera operating the sensor is at least one that is at least functionally equipped with a 202416212 15. A camera connected to a device that reproduces content worthy of protection is operated for recording purposes, at least temporarily.
6. Method according to the preceding claim, characterized in that the second camera is positioned and operated in such a way that, in relation to the content to be protected, it is able to capture at least parts of the content to be protected.
7. Method according to one of the two preceding claims, characterized in that the second camera is positioned in relation to the device and, in particular, is physically connected to it and operated in such a way that the at least one lens opening is arranged in the same plane as the reproduction of the content worthy of protection.
8. Method according to one of the preceding claims, characterized in that the determination of at least one data correlating with the presence and / or orientation of an image recording device is carried out in such a way that the AI module, a trained, in particular multimodal, LLM, is operated in such a way that it is called with a prompt formed at least from the sensor data in such a way that a prediction about the presence and / or orientation is output.
9. Method according to the preceding claim, characterized in that the LLM is trained in such a way that it predicts a camera lens of the first camera which at least partially captures the area of the content worthy of protection.
10. Method according to one of the two preceding claims, characterized in that the LLM is trained such that the prediction of the area is carried out at least by means of a reflection of light, in particular emitted by a stationary source, for example the flash device, detected on the basis of the image recording.
11. Method according to one of the three preceding claims, characterized in that the LLM is trained in such a way that it predicts the operating state of the camera lens of the first camera. 202416212 16 12. Method according to the preceding claim, characterized in that the LLM is trained such that the prediction of the operating state is carried out at least on the basis of infrared points detectable in the image acquisition.
13. Method according to claim 8, characterized in that at least one further camera, disjoint from the second camera and connected to a device for control, its control and / or image evaluation, is operated in such a way that it calls the LLM with recordings in such a way that the LLM makes a prediction about the presence and / or orientation of the first camera.
14. Method according to one of the preceding claims, characterized in that the suppression action is carried out in such a way that the device reproducing the content worthy of protection is manipulated in such a way that, as long as there is an indication of an unauthorized recording, the reproduction of the content to be protected is prevented.
15. Method according to one of the preceding claims, characterized in that the suppression action is carried out in the following steps: a) identifying a communication link to the first camera, b) generating a message correlated with at least one control signal, c) sending the correlated message, d) receiving the correlated message by a device designed to influence the function of the first camera, e) manipulating the function by the influencing device in such a way that the generation of a recording reproducing the content is prevented.
16. Method according to one of the preceding claims, characterized in that, in the presence of an indication of unauthorized recording, at least one warning signal is issued, visually, in particular on the playback device, and / or acoustically via a loudspeaker device.
17. Arrangement for protection against image recordings of protected content characterized by means for carrying out the method according to one of the preceding claims.