Evaluation tool and method for detecting whether face recognition can be bypassed
By replacing the frame image captured by the face recognition system camera as a test image and inserting the action image for feature comparison, the problem of limited application scope of security testing of face recognition system in the prior art is solved, and the application scope of face recognition system in various scenarios is realized effective testing and efficient verification.
Patent Information
- Application Number
- CN202411022155.5
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-29
- Publication Date
- 2024-11-26
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
When conducting security testing on facial recognition systems, the existing technology has limited scope of application and poor flexibility, and cannot effectively detect the risk of attackers circumventing facial recognition systems through masks, photos and other means.
The frame image captured by the face recognition system camera is replaced as a test image, and the action image is inserted according to the prompt information to compare features to verify the security of the face recognition system. It is suitable for face recognition systems in various scenarios.
It realizes effective testing of the face recognition system, improves testing efficiency and applicability, and can flexibly detect whether the system is safe, and is suitable for a variety of recognition scenarios, improving the flexibility and accuracy of security verification.
Smart Images

Figure CN119028030A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of face recognition technology, and more specifically, to an evaluation tool and method for detecting whether face recognition can be bypassed. Background Art
[0002] In the digital age, face recognition technology has gradually become an indispensable part of daily life. However, with the popularization and development of face recognition technology, the security verification challenges it faces are becoming increasingly prominent. At present, attackers can use masks, photos, videos, deep fake technology and other means to deceive face recognition systems, thereby obtaining unauthorized access or potentially conducting illegal activities. These bypass attacks not only pose a threat to personal privacy, but may also have a serious impact on social security and information security.
[0003] Currently, when testing the face recognition system, a malicious application APP is used to manage the system camera. When the face recognition camera is turned on, the camera data stream passes through the APP. The APP uses the whitelist of the camera interface to detect the data stream. However, this method can only be used to identify the system camera name provided by major mobile phone manufacturers, and its scope of application is limited.
[0004] Therefore, how to provide a technical solution with wide applicability to detect whether a test method can bypass face recognition has become a technical problem that needs to be solved urgently. Summary of the invention
[0005] The purpose of some embodiments of the present application is to provide an evaluation tool and method for detecting whether face recognition can be bypassed. Through the technical solutions of the embodiments of the present application, effective testing of face recognition systems in different scenarios can be achieved, with high test efficiency and wide applicability.
[0006] In a first aspect, some embodiments of the present application provide an evaluation method for detecting whether facial recognition can be bypassed, comprising: replacing a frame image of a test person captured by a camera during the recognition process of a facial recognition system with a test image, wherein the test image is a frontal face image of the test person; in response to an operation instruction of the test person based on prompt information, inserting and displaying an action image corresponding to the operation instruction; determining a test result of the facial recognition system by verifying the action image and the test image, wherein the test result represents whether the facial recognition system is safe.
[0007] Some embodiments of the present application replace the frame image captured by the camera in the face recognition system with a test image, then select the corresponding operation instruction according to the prompt information, insert and display the corresponding action image, and finally verify the action image and the test image to obtain the test result. Some embodiments of the present application can realize effective testing of the face recognition system in different scenarios by image replacement, with high test efficiency, wide applicability and high flexibility.
[0008] In some embodiments, before using the test image to replace the frame image of the test person captured by the camera during the recognition process of the face recognition system, the method also includes: photographing the test person to obtain a video frame recognizable by the camera; producing the test image based on the video frame; based on the test image, generating the action image corresponding to the test person, wherein the action image is related to the facial movements of the test person, and the facial movements include: turning the head, nodding, raising the head, closing the eyes and opening the mouth.
[0009] Some embodiments of the present application obtain a test image by taking a photo of the test person, and then generate a corresponding action image to provide test data support for subsequent face recognition system testing.
[0010] In some embodiments, after inserting and displaying the action image corresponding to the operation instruction, the method further includes: restoring to the test image.
[0011] In some embodiments, the test result of the face recognition system is determined by verifying the action image and the test image, including: comparing features of the action image with the test image, and comparing features of the test image with a pre-stored image to obtain a verification result; and obtaining the test result based on the verification result.
[0012] Some embodiments of the present application can implement effective testing of the face recognition system by separately verifying the action image and the test image to obtain the verification result and then obtain the test result.
[0013] In some embodiments, the feature comparison between the action image and the test image, and the feature comparison between the test image and the pre-stored image to obtain a verification result includes: confirming that the features of the action image are consistent with those of the test image, and that the features of the test image are consistent with those of the pre-stored image, then the verification result is passed; and obtaining the test result based on the verification result includes: when confirming that the verification result is passed, the test result is unsafe.
[0014] In some embodiments, the feature comparison between the action image and the test image, and the feature comparison between the test image and the pre-stored image to obtain a verification result includes: confirming that the features of the action image and the test image are inconsistent, and / or the features of the test image and the pre-stored image are inconsistent, then the verification result is failed; obtaining the test result based on the verification result includes: when confirming that the verification result is failed, the test result is safe.
[0015] Some embodiments of the present application obtain verification results and then test results through the relationship between the features of the action image and the test image and the corresponding image, which can achieve effective testing of the face recognition system and has wide applicability.
[0016] In a second aspect, some embodiments of the present application provide an evaluation tool for detecting whether facial recognition can be bypassed, comprising: a first replacement module, configured to replace a frame image of a test person captured by a camera during the recognition process of a facial recognition system with a test image, wherein the test image is a frontal face image of the test person; a second replacement module, configured to respond to an operation instruction of the test person based on prompt information, insert and display an action image corresponding to the operation instruction; a test module, configured to determine a test result of the facial recognition system by verifying the action image and the test image, wherein the test result represents whether the facial recognition system is safe.
[0017] In a third aspect, some embodiments of the present application provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, can implement the method described in any embodiment of the first aspect.
[0018] In a fourth aspect, some embodiments of the present application provide an electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor can implement a method as described in any embodiment of the first aspect when executing the program.
[0019] In a fifth aspect, some embodiments of the present application provide a computer program product, wherein the computer program product comprises a computer program, wherein the computer program, when executed by a processor, can implement the method described in any embodiment of the first aspect. BRIEF DESCRIPTION OF THE DRAWINGS
[0020] In order to more clearly illustrate the technical solutions of some embodiments of the present application, the drawings required for use in some embodiments of the present application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present application and therefore should not be regarded as limiting the scope. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.
[0021] Figure 1 One of the flow charts of the evaluation method for detecting whether face recognition can be bypassed provided in some embodiments of the present application;
[0022] Figure 2 A schematic diagram of a test interface provided for some embodiments of the present application;
[0023] Figure 3 A flow chart of function calls in testing provided for some embodiments of the present application;
[0024] Figure 4 A second flow chart of a method for evaluating whether face recognition can be bypassed provided in some embodiments of the present application;
[0025] Figure 5 A block diagram of an evaluation tool for detecting whether face recognition can be bypassed provided in some embodiments of the present application;
[0026] Figure 6 A schematic diagram of an electronic device is provided for some embodiments of the present application. DETAILED DESCRIPTION
[0027] The technical solutions in some embodiments of the present application will be described below in conjunction with the drawings in some embodiments of the present application.
[0028] It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in subsequent drawings. At the same time, in the description of this application, the terms "first", "second", etc. are only used to distinguish the description and cannot be understood as indicating or implying relative importance.
[0029] In the related art, the principle of face recognition technology is to analyze and identify geometric features and biometric features in face images through computer vision and pattern recognition technology. Although this technology brings many advantages in terms of improving security and convenience, it is accompanied by a series of challenges of bypassing attacks. In order to cope with the bypass challenges faced by face recognition technology, methods such as liveness detection, multimodal recognition, anti-attack algorithms, regular model updates, and enhanced system security have been proposed and applied to actual systems to improve the robustness and security of face recognition systems. There are two main ways in the existing technology to bypass face tools at the software level. One is to use a malicious APP to take over the camera of the face recognition system, so that the camera opened by the face recognition system is the malicious APP, and all camera data streams pass through the malicious APP. This solution has a large disadvantage. The protection on the APP side is relatively easy. The APP can use the whitelist of the camera interface to make a judgment. The whitelist is the name of the system camera provided by major mobile phone manufacturers, but it is not applicable to other recognition scenarios. The other is to hijack the return interface of the face recognition system: bypass face recognition by modifying the return result of the server. This solution has poor applicability. Generally, after recognition, a token will be returned for the next step. Without the correct token, the purpose of bypassing face recognition cannot be achieved.
[0030] It can be seen from the above-mentioned related technologies that the methods for performing security testing on face recognition systems in the prior art have limited scope of application and poor flexibility.
[0031] In view of this, some embodiments of the present application provide an evaluation method for detecting whether face recognition can be bypassed. During the face verification process of the face recognition system, the method obtains the test image of the tester to replace the frame image taken by the camera of the face recognition system, and then performs relevant liveness detection based on the prompt information. During the liveness detection process, relevant action images are inserted and displayed for the corresponding actions. Finally, the test image and the action image are verified to determine the test result of the face recognition system. Some embodiments of the present application can realize effective testing of whether the face recognition system can bypass the face, and can be applicable to face recognition systems in various scenarios, with high flexibility.
[0032] The following is combined with Figure 1 The implementation process of the face recognition system test performed by a terminal device with a face recognition system provided by some embodiments of the present application is exemplified.
[0033] Please see attached Figure 1 , Figure 1 A method for detecting whether face recognition can be bypassed is provided for some embodiments of the present application. The method for detecting whether face recognition can be bypassed may include:
[0034] S110, using a test image to replace a frame image of a test person captured by a camera during a recognition process of a face recognition system, wherein the test image is a frontal face image of the test person.
[0035] For example, in some embodiments of the present application, during face recognition, the frame image captured by the camera is replaced by a test image obtained by pre-capturing the face of the test person through Hook interception.
[0036] In one embodiment, any application can operate the camera through the Camera.java class. Among them, the handleMessage function can be used to preview the image taken by the camera, and the frame image taken by the camera can also be obtained through this function. This function processes the frame image message transmitted from the bottom layer, and is called in a loop to continuously process the frame image. During face recognition, in order to recognize the face, it is necessary to obtain a suitable frontal face picture from all preview frame images. Therefore, you can start with the handleMessage function and use the Hook interception method to replace each frame image received by handleMessage from the bottom layer with a test image. After replacing these frames, regardless of whether the camera captures a normal face, the frame images that the APP can receive are all replaced normal face photos (as a specific example of a test image).
[0037] It should be noted that in the entire process from the camera capturing the image to the data transmission verification image, any function carrying frame data can be used as an interception point to replace the corresponding video frame. The detection tool of the embodiment provided in this application only uses the top-level function for interception and replacement. In practical applications, the interception point can be flexibly found. Specifically, in the process of testing the face recognition system, the data transmission process mainly involves the following: Figure 3The calling function flow shown in the figure. First, when starting the face recognition system, you need to adjust the camera as follows: call the Open function to open the camera, call the native_setup function to set the state of the camera, call the initialize function to initialize the camera, and set the callback setCallbacks function so that after shooting and obtaining the frame data, the camera acquisition frame (i.e., the test image) is obtained to replace the frame data currently captured by the camera. After completing the camera initialization, the face recognition system calls the takePicture function to drive the camera to shoot and obtain the frame data. The camera acquisition frame can be obtained by calling the setCallbacks function, and then the data submission postData function is called to submit the acquired frame, the data copy copyAndPost function is called, the event postEventFromNative function is called to pass the event to the Java layer, and the camera message queue handleMessage function is called to obtain the replaced preview frame data to display in the interface of the face recognition system. As long as the above-mentioned functions carry frame data, they can be replaced as interception points to achieve effective testing of whether the face recognition system can bypass the face.
[0038] Specifically, the Hook injection plug-in is used to replace the video frame captured by the camera (as a specific example of a frame image). During face recognition verification, the video frame of the camera can be replaced in real time by clicking on the relevant operation items on the terminal device to achieve the effect of disguising a normal video. That is, when the face recognition verification just begins and the camera is turned on, the Hook tool is used to intercept the video frame captured by the camera and replace it with the frontal photo of the current tester (that is, the frontal image). When the APP recognizes the currently replaced video frame, it will successfully recognize the frontal photo of the tester, and then proceed to the next step to start liveness verification. Since it is necessary to keep the person's face looking at the camera, the video frame needs to be replaced by a pre-generated frontal photo by default.
[0039] In some embodiments of the present application, before executing S110, the evaluation method for detecting whether face recognition can be bypassed may also include: photographing the test person to obtain a video frame recognizable by the camera; producing the test image based on the video frame; based on the test image, generating the action image corresponding to the test person, wherein the action image is related to the facial movements of the test person, and the facial movements include: turning the head, nodding, raising the head, closing the eyes and opening the mouth.
[0040] For example, in some embodiments of the present application, before the test is conducted, a frontal photo of the tester's face is obtained. First, a video frame that can be recognized by the camera and collected by the tester is made. The pixel size and the photo deflection method can be based on the model of the current terminal device to save and view the camera video frame. For example, the model of the terminal device tested in this test is Google Pixel. After saving the video frame, it is found that the image format is 640x480 pixels, rotated 90 degrees clockwise, and a frontal photo is made. After that, artificial intelligence technology is used to generate action images such as turning the head left, turning the head right, nodding, raising the head, closing the eyes and opening the mouth of the tester for backup. Finally, the frontal picture and action image related to the tester can be pre-stored in the terminal device to be verified, or stored in other databases or servers, and then obtained in real time through the network. Specifically, it can be selected according to the actual application scenario, and the embodiments of the present application are not specifically limited here.
[0041] S120, in response to the tester's operation instruction based on the prompt information, inserting and displaying an action image corresponding to the operation instruction.
[0042] In some embodiments of the present application, after the action image is displayed, the test image is restored.
[0043] For example, in some embodiments of the present application, when performing liveness detection, Figure 2 As shown, according to the prompt information of the APP of the terminal device, the tester can click the corresponding button on the left side of the Hook tool, such as "turn left", "close eyes", "open mouth", etc. When clicking the relevant button, the Hook tool will automatically find the pre-generated corresponding action image according to the photo name and path. For example, for the eye-closing action, the static photo of the closed eyes is named eye.jpg. The tester clicks the eye-closing button to replace the front face photo displayed on the interface with the eye-closing photo (as a specific example of an action image). Clicking once only inserts one frame (such as closing eyes). The next frame is controlled by the Hook tool and automatically replaced with the default front face photo. In this way, multiple clicks can insert multiple action frames, which are connected together to form a dynamic eye-closing and eye-opening action.
[0044] S130, determining a test result of the face recognition system by verifying the action image and the test image, wherein the test result indicates whether the face recognition system is safe.
[0045] For example, in some embodiments of the present application, in liveness detection, the action image and the test image are verified to obtain the test result.
[0046] Specifically, in some embodiments of the present application, S130 may include: performing feature comparison between the action image and the test image, and performing feature comparison between the test image and a pre-stored image to obtain a verification result; and obtaining the test result based on the verification result.
[0047] For example, in some embodiments of the present application, the front end of the terminal device compares the facial features of the frontal face photo with the facial features of the action image to confirm whether they are consistent, and the back end compares the facial features of the frontal face picture with the facial features of the pre-stored images in the terminal device to obtain a verification result and then a test result.
[0048] In some embodiments of the present application, S130 may include: confirming that the features of the action image are consistent with those of the test image, and that the features of the test image are consistent with those of the pre-stored image, then the verification result is passed; when confirming that the verification result is passed, the test result is unsafe.
[0049] For example, in some embodiments of the present application, when the facial features of the frontal photo and the action image are consistent, the front end passes the verification; when the facial features of the frontal photo and the predicted image are consistent, the back end also passes the verification; under the premise that both the front end and the back end pass the verification, the test result is unsafe, which indicates that the security of the face recognition system is weak and can be easily bypassed by attackers, and further security protection is required. Among them, whether the facial features in the image are consistent can be compared and identified by using a face recognition algorithm or an image recognition algorithm, which is not specifically limited in the embodiments of the present application.
[0050] In other embodiments of the present application, S130 may include: confirming that the features of the action image are inconsistent with those of the test image, and / or that the features of the test image are inconsistent with those of the pre-stored image, then the verification result is failed; when confirming that the verification result is failed, the test result is safe.
[0051] For example, in some embodiments of the present application, when the facial features of the frontal photo and the action image are inconsistent, the front end fails the verification; when the facial features of the frontal photo and the predicted image are inconsistent, the back end also fails the verification; if at least one of the verifications fails on the front end and the back end, the test result is safe, which indicates that the face recognition system is highly secure and not easily bypassed by attackers.
[0052] The following is combined with Figure 4 The specific process of whether the detection provided by some embodiments of the present application can bypass face recognition is exemplified.
[0053] Please see attached Figure 4 , Figure 4A flow chart of an evaluation method for detecting whether face recognition can be bypassed is provided for some embodiments of the present application.
[0054] The above process is explained below as an example.
[0055] S410, filming the test person to obtain video frames recognizable by the camera to produce a test image.
[0056] S420: Generate an action image corresponding to the test person based on the test image.
[0057] S430, using the Hook tool, using the test image to replace the frame image of the test person captured by the camera during the recognition process of the face recognition system.
[0058] S440, according to the prompt information, in response to the tester's operation instruction on the Hook tool operation item, insert and display the action image corresponding to the operation instruction, and then immediately replace it with the test image.
[0059] Among them, the operation items are as follows Figure 2 The “Face up, turn left, turn right, close eyes, open mouth” buttons on the interface shown. The type of operation item can be set as needed, and the embodiment of the present application is not limited thereto.
[0060] S450, determine whether both the test image and the action image pass the verification, if so, execute S451, otherwise execute S452.
[0061] S451, determining that the test result of the face recognition system is unsafe.
[0062] S452, determining that the test result of the face recognition system is safe.
[0063] It can be understood that the specific implementation process of S410 to S452 can refer to the method embodiment provided above, and in order to avoid repetition, the detailed description is appropriately omitted here.
[0064] It should be noted that the Hook tool provided in some embodiments of the present application is based on the system layer, and since all Android phone manufacturers use Android source code for customized development, the tool can be used by all Android-based systems, and the mobile phone system has high adaptability. By intercepting the preview frame of the system layer and replacing it, it is possible to deceive the face recognition of any APP, and by freely clicking on the action, any action required for liveness detection can be performed, which is highly flexible. Since the operation of this tool occurs at the system level, the APP cannot perform detection, nor can it recognize that the camera data has been replaced, so it is highly secure.
[0065] Please refer to Figure 5 , Figure 5The block diagram of the composition of the evaluation tool for detecting whether face recognition can be bypassed provided by some embodiments of the present application is shown. It should be understood that the evaluation tool for detecting whether face recognition can be bypassed corresponds to the above method embodiment and can execute each step involved in the above method embodiment. The specific functions of the evaluation tool for detecting whether face recognition can be bypassed can refer to the description above. To avoid repetition, the detailed description is appropriately omitted here.
[0066] Figure 5 The evaluation tool for detecting whether face recognition can be bypassed includes at least one software function module that can be stored in a memory in the form of software or firmware or solidified in the evaluation tool for detecting whether face recognition can be bypassed. The evaluation tool for detecting whether face recognition can be bypassed includes: a first replacement module 510, configured to replace the frame image of the test person captured by the camera during the recognition process of the face recognition system with a test image, wherein the test image is a frontal face image of the test person; a second replacement module 520, configured to respond to the test person's operation instruction based on prompt information, insert and display an action image corresponding to the operation instruction; a test module 530, configured to determine the test result of the face recognition system by verifying the action image and the test image, wherein the test result represents whether the face recognition system is safe.
[0067] In some embodiments of the present application, before the first replacement module 510, the evaluation tool for detecting whether face recognition can be bypassed also includes: an acquisition module (not shown in the figure) is configured to shoot the test person and obtain a video frame recognizable by the camera; produce the test image based on the video frame; based on the test image, generate the action image corresponding to the test person, wherein the action image is related to the facial movements of the test person, and the facial movements include: turning the head, nodding, raising the head, closing the eyes and opening the mouth.
[0068] In some embodiments of the present application, the second replacement module 520 is configured to restore to the test image after inserting and displaying the action image corresponding to the operation instruction.
[0069] In some embodiments of the present application, the test module 530 is configured to perform feature comparison between the action image and the test image, and to perform feature comparison between the test image and a pre-stored image to obtain a verification result; and obtain the test result based on the verification result.
[0070] In some embodiments of the present application, the test module 530 is configured to confirm that the features of the action image are consistent with those of the test image, and that the features of the test image are consistent with those of the pre-stored image, and the verification result is passed; when the verification result is confirmed to be passed, the test result is unsafe.
[0071] In some embodiments of the present application, the test module 530 is configured to confirm that the features of the action image are inconsistent with those of the test image, and / or that the features of the test image are inconsistent with those of the pre-stored image, and the verification result is failed; when it is confirmed that the verification result is failed, the test result is safe.
[0072] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the device described above can refer to the corresponding process in the aforementioned method, and will not be described in detail here.
[0073] Some embodiments of the present application further provide a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, can implement the operations of the method corresponding to any of the above methods provided in the above embodiments.
[0074] Some embodiments of the present application further provide a computer program product, which includes a computer program, wherein when the computer program is executed by a processor, it can implement the operations corresponding to any of the above methods provided in the above embodiments.
[0075] like Figure 6 As shown, some embodiments of the present application provide an electronic device 600, which includes: a memory 610, a processor 620, and a computer program stored in the memory 610 and executable on the processor 620, wherein the processor 620 can implement a method as described in any of the above embodiments when reading the program from the memory 610 through a bus 630 and executing the program.
[0076] Processor 620 can process digital signals and can include various computing structures, such as complex instruction set computer structure, reduced instruction set computer structure, or a structure that implements a combination of multiple instruction sets. In some examples, processor 620 can be a microprocessor.
[0077] The memory 610 may be used to store instructions executed by the processor 620 or data related to the execution of instructions. These instructions and / or data may include codes for implementing some or all functions of one or more modules described in the embodiments of the present application. The processor 620 of the disclosed embodiment may be used to execute instructions in the memory 610 to implement the method shown above. The memory 610 includes a dynamic random access memory, a static random access memory, a flash memory, an optical memory, or other memory known to those skilled in the art.
[0078] The above description is only an embodiment of the present application and is not intended to limit the scope of protection of the present application. For those skilled in the art, the present application may have various changes and variations. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the scope of protection of the present application. It should be noted that similar reference numerals and letters represent similar items in the following drawings, so once an item is defined in one drawing, it does not need to be further defined and explained in the subsequent drawings.
[0079] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
[0080] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device. In the absence of further restrictions, the elements defined by the sentence "comprise a ..." do not exclude the existence of other identical elements in the process, method, article or device including the elements.
Claims
1. A method for detecting whether face recognition can be bypassed, characterized in that: include: Using the test image to replace the frame image of the test person captured by the camera during the recognition process of the face recognition system, wherein the test image is a frontal face image of the test person; In response to the tester's operation instruction based on the prompt information, inserting and displaying an action image corresponding to the operation instruction; By verifying the action image and the test image, a test result of the face recognition system is determined, wherein the test result represents whether the face recognition system is safe.
2. The method according to claim 1, characterized in that Before using the test image to replace the frame image of the test person captured by the camera during the recognition process of the face recognition system, the method further includes: Filming the tester to obtain video frames recognizable by the camera; Creating the test image based on the video frame; Based on the test image, the action image corresponding to the test person is generated, wherein the action image is related to the facial movements of the test person, and the facial movements include: turning the head, nodding, raising the head, closing the eyes and opening the mouth.
3. The method according to claim 1 or 2, characterized in that After inserting and displaying the action image corresponding to the operation instruction, the method further includes: restoring to the test image.
4. The method according to claim 1 or 2, characterized in that: The step of determining the test result of the face recognition system by verifying the action image and the test image includes: Comparing the features of the action image with the test image, and comparing the features of the test image with the pre-stored image, to obtain a verification result; Based on the verification result, the test result is obtained.
5. The method according to claim 4, characterized in that The step of comparing the features of the action image with the test image, and comparing the features of the test image with a pre-stored image to obtain a verification result includes: Confirming that the features of the action image are consistent with those of the test image, and that the features of the test image are consistent with those of the pre-stored image, the verification result is passed; The obtaining the test result based on the verification result includes: When it is confirmed that the verification result is passed, the test result is unsafe.
6. The method according to claim 4, characterized in that The step of comparing the features of the action image with the test image, and comparing the features of the test image with a pre-stored image to obtain a verification result includes: If it is confirmed that the features of the action image and the test image are inconsistent, and / or the features of the test image and the pre-stored image are inconsistent, the verification result is failure; The obtaining the test result based on the verification result includes: When it is confirmed that the verification result is not passed, the test result is safe.
7. A tool for evaluating whether face recognition can be bypassed, characterized in that: include: A first replacement module is configured to replace a frame image of a test person captured by a camera during recognition by a face recognition system with a test image, wherein the test image is a frontal face image of the test person; A second replacement module is configured to, in response to an operation instruction of the tester based on the prompt information, insert and display an action image corresponding to the operation instruction; The testing module is configured to determine a test result of the face recognition system by verifying the action image and the test image, wherein the test result represents whether the face recognition system is safe.
8. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein the computer program executes the method according to any one of claims 1 to 6 when executed by a processor.
9. An electronic device, characterized in that: The method comprises a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the computer program executes the method according to any one of claims 1 to 6 when being run by the processor.
10. A computer program product, characterized in that The computer program product comprises a computer program, wherein the computer program executes the method according to any one of claims 1 to 6 when executed by a processor.
Citation Information
Patent Citations
Face image verification method and device, electronic equipment and storage medium
CN114091503A
Detection system for face recognition presentation attack
CN116434304A
Face recognition attack test method and system
CN116994342A
Interactive attack test method, optimization method and device
CN117093986A
Method for preventing AI face changing and attack in face biopsy process
CN117728955A
