Video forgery detection method, device, equipment and readable storage medium

Abstract

This application relates to a video forgery detection method, apparatus, device, and readable storage medium. The method comprises: obtaining a first silhouette image of a target object based on a video frame captured by a monitoring device; and obtaining a second silhouette image of the target object based on a Wi-Fi signal captured by a Wi-Fi device; and determining whether the video frame is forged based on the similarity between the first and second silhouette images. This method addresses the complex, computationally expensive, and delayed issues of video forgery detection in related technologies, achieving real-time, fine-grained, and computationally inexpensive video forgery detection.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence security technology, and in particular to a video forgery detection method, apparatus, device and readable storage medium. Background Art

[0002] Video / camera-based surveillance systems are widely used in stores, banks, logistics centers, and other areas to prevent crimes like theft and preserve video evidence. With the development of the Internet of Things (IoT), these surveillance systems are becoming more distributed, cost-effective, and autonomous, but this also increases security risks. Attackers can hijack cameras or the Ethernet cables connecting them and manipulate the live video stream, replacing real frames with fake ones. Furthermore, attackers can use deepfake technology to modify video frames, making them indistinguishable from fakes. These forgery attacks pose a serious threat to the authenticity and credibility of existing video surveillance systems.

[0003] To address this issue, methods have been proposed to detect forged frames in suspicious videos. However, these approaches all have their drawbacks. For example, watermark-based methods require additional dedicated modules within the camera, but such modules are difficult to implement in mainstream cameras. Video forensics methods typically extract spatiotemporal features from consecutive frames to achieve fine-grained forgery detection, but this feature extraction results in high computational overhead and latency, hindering their use in real-time video. Summary of the Invention

[0004] Based on this, it is necessary to provide a video forgery detection method, device, equipment and readable storage medium to address the above technical problems.

[0005] In a first aspect, an embodiment of the present application provides a method for detecting video forgery, the method comprising:

[0006] Acquire a first silhouette image of the target object based on the video frames captured by the monitoring device; and

[0007] Acquire a second silhouette image of the target object based on a WIFI signal collected by the WIFI device;

[0008] Based on the similarity between the first silhouette image and the second silhouette image, it is determined whether the video frame is forged.

[0009] In one embodiment, acquiring a first silhouette image of the target object based on a video frame captured by a monitoring device includes:

[0010] The video frames captured by the monitoring device are input into a pre-trained instance segmentation model for classification, regression, and segmentation processing to obtain the first silhouette image of the target object.

[0011] In one embodiment, acquiring a second silhouette image of the target object based on the WIFI signal includes:

[0012] Acquiring a wavefront of the target object based on the WIFI signal;

[0013] The wavefront of the target object is input into a pre-trained semantic segmentation network model to extract a second silhouette image of the target object.

[0014] In one embodiment, the WIFI device includes a WIFI transmitter and a WIFI receiver, and acquiring the wavefront of the target object based on the WIFI signal includes:

[0015] A spatial rectangular coordinate system is constructed with the WIFI receiver as the coordinate origin O, the direction of the line connecting the WIFI receiver and the WIFI transmitter as the X-axis direction, the direction perpendicular to the ground as the Z-axis direction, and the Y-axis direction being perpendicular to the X-axis and the Z-axis respectively;

[0016] Based on formula (1), the wavefront s of each reflection point on the target object is calculated using two-dimensional inverse Fourier transform p,q ;

[0017]

[0018] Wherein, S(t,f) is the WIFI signal of the target object received by the WIFI receiver at the subcarrier frequency f and time t, α is the amplitude attenuation constant, p and q are the coordinate values of the reflection point in the X-axis and Z-axis in the spatial rectangular coordinate system, respectively, the reflection plane where the reflection point is located is parallel to the coordinate plane XOZ, P and Q are the length and width of the reflection plane, j is a complex unit, and f c is the center frequency of the subcarrier, c is the propagation speed of the subcarrier, λ c is the center frequency f c The corresponding wavelength, is the initial displacement of the target object from the coordinate origin O at time t0, v y is the relative motion speed of the target object along the Y axis, γ is the azimuth angle of the reflection point relative to the coordinate origin, and β is the elevation angle of the reflection point relative to the coordinate origin;

[0019] The wavefront of the target object is constructed based on the wavefronts of each reflection point of the target object.

[0020] In one embodiment, determining whether the video frame is forged based on the similarity between the first silhouette image and the second silhouette image includes:

[0021] Inputting the first silhouette image and the second silhouette image into two branches of the twin neural network respectively, and outputting a first semantic feature vector and a second semantic feature vector;

[0022] Based on the similarity between the first semantic feature vector and the second semantic feature vector, it is determined whether the video frame is forged.

[0023] In one embodiment, determining whether the video frame is forged based on the similarity between the first semantic feature vector and the second semantic feature vector includes:

[0024] Calculating the Euclidean distance between the first semantic feature vector and the second semantic feature vector; determining the similarity; wherein the similarity is the reciprocal of the Euclidean distance;

[0025] Determining whether the similarity is less than a preset threshold;

[0026] If so, it is determined that the video frame is forged;

[0027] If not, it is determined that the video frame has not been forged.

[0028] In one embodiment, before acquiring the second silhouette image of the target object based on the WIFI signal, the method includes:

[0029] The WIFI signal is resampled so that the WIFI signal is temporally aligned with the video frame, and the WIFI signal is filtered using a median filter, a mean filter, and a low-pass Butterworth filter in sequence.

[0030] In a second aspect, an embodiment of the present application further provides a video forgery detection device, the device comprising:

[0031] an acquisition module, configured to acquire a first silhouette image of a target object based on a video frame acquired by a monitoring device; and acquire a second silhouette image of the target object based on a WIFI signal acquired by a WIFI device;

[0032] A determination module is configured to determine whether the video frame is forged based on a similarity between the first silhouette image and the second silhouette image.

[0033] In a third aspect, an embodiment of the present application further provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the method described in the first aspect above when executing the computer program.

[0034] In a fourth aspect, an embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored. When the computer program is executed by a processor, the method described in the first aspect above is implemented.

[0035] The above-described video forgery detection method, apparatus, device, and readable storage medium obtain a first silhouette image of a target object based on video frames captured by a monitoring device; and obtain a second silhouette image of the target object based on Wi-Fi signals captured by a Wi-Fi device. Based on the similarity between the first and second silhouette images, the method determines whether the video frame has been forged. This method addresses the complex, computationally expensive, and delayed nature of video forgery detection in related technologies, achieving real-time, fine-grained, and computationally efficient video forgery detection.

[0036] The details of one or more embodiments of the present application are set forth in the following drawings and description to make other features, objects, and advantages of the present application more readily apparent. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0038] Figure 1 is a schematic diagram of an application environment of a video forgery detection method in an embodiment;

[0039] Figure 2 is a flow chart of a method for detecting video forgery in an embodiment;

[0040] Figure 3 is a flowchart of specific steps for obtaining a second silhouette image in an embodiment;

[0041] Figure 4 is a flowchart of specific steps for obtaining the wavefront of a target object in one embodiment;

[0042] Figure 4-1 is a schematic diagram of establishing a spatial rectangular coordinate system in an embodiment;

[0043] Figure 5 is a schematic diagram of a flow chart for executing specific steps of S202 in an embodiment;

[0044] Figure 6 is a flowchart of executing the specific steps of S502 in an embodiment;

[0045] Figure 7 is a structural block diagram of a video forgery detection device in an embodiment;

[0046] Figure 8 It is a schematic diagram of the structure of a computer device in an embodiment. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solutions and advantages of this application more clearly understood, the present application is described and illustrated below in conjunction with the accompanying drawings and examples. It should be understood that the specific embodiments described herein are merely used to explain this application and are not intended to limit this application. Based on the embodiments provided in this application, all other embodiments obtained by those of ordinary skill in the art without making any creative efforts are within the scope of protection of this application.

[0048] Obviously, the drawings described below are merely examples or embodiments of the present application. Those skilled in the art can, without inventive effort, apply the present application to other similar scenarios based on these drawings. Furthermore, it is also understood that, although the effort involved in such a development process may be complex and lengthy, for those skilled in the art related to the content disclosed in this application, changes in design, manufacturing, or production based on the technical content disclosed in this application are merely conventional technical means and should not be construed as an insufficiency of the content disclosed in this application.

[0049] References to "embodiments" in this application mean that a particular feature, structure, or characteristic described in connection with the embodiment may be included in at least one embodiment of the application. The appearance of this phrase in various places in the specification does not necessarily refer to the same embodiment, nor does it refer to independent or alternative embodiments that are mutually exclusive of other embodiments. It is understood, both explicitly and implicitly, by those skilled in the art that the embodiments described in this application may be combined with other embodiments unless there is a conflict.

[0050] Unless otherwise defined, the technical or scientific terms used in this application should have the ordinary meaning understood by a person of ordinary skill in the technical field to which this application belongs. The words "one", "a", "the" and the like used in this application do not indicate a limit on quantity and may indicate the singular or plural. The terms "include", "comprise", "have" and any variations thereof used in this application are intended to cover non-exclusive inclusions; for example, a process, method, system, product or device that includes a series of steps or modules (units) is not limited to the listed steps or units, but may also include steps or units that are not listed, or may also include other steps or units that are inherent to these processes, methods, products or devices. The words "connect", "connected", "coupled" and the like used in this application are not limited to physical or mechanical connections, but may include electrical connections, whether direct or indirect. The word "multiple" used in this application refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships can exist. For example, "A and / or B" can mean: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects before and after are in an "or" relationship. The terms "first", "second", "third", etc. involved in this application are only used to distinguish similar objects and do not represent a specific order for the objects.

[0051] The video forgery detection method provided in this application can be applied to Figure 1 In the application environment shown, the terminal 106 communicates with the WIFI device 102 and the monitoring device 104 via a network. The WIFI device 102 is used to obtain a collected WIFI signal, and the monitoring device 104 is used to collect video frames. The terminal 106 is used to obtain a first silhouette image of a target object based on the video frame collected by the monitoring device; and obtain a second silhouette image of the target object based on the WIFI signal collected by the WIFI device; and determine whether the video frame is forged based on the similarity between the first silhouette image and the second silhouette image.

[0052] The terminal 102 may include one or more processors (e.g., a single-chip processor or a multi-chip processor). By way of example only, the terminal 102 may include a central processing unit (CPU), an application-specific integrated circuit (ASIC), an application-specific instruction set processor (ASIP), a graphics processing unit (GPU), a physical processing unit (PPU), a digital signal processor (DSP), a field programmable gate array (FPGA), a programmable logic device (PLD), a controller, a microcontroller unit, a reduced instruction set computer (RISC), a microprocessor, or any combination thereof.

[0053] The present application provides a method for detecting video forgery. Figure 2 As shown, the method includes the following steps:

[0054] Step S201: acquiring a first silhouette image of a target object based on a video frame captured by a monitoring device; and acquiring a second silhouette image of the target object based on a WIFI signal captured by a WIFI device;

[0055] Specifically, a camera is used to record the target object in the monitoring area to obtain a video stream, which is then sliced and segmented to obtain a video frame of the target object. A first silhouette image of the target object is obtained from the video frame. A WiFi device is also installed in the monitoring area. The transmitter signal propagates to the surface of the object. After a series of reflection, transmission, and diffraction processes, the signal reaches the receiver, which receives the WiFi signal of the target object, namely the Channel State Information (CSI). CSI describes the attenuation factors of the signal on each transmission path in the frequency domain and time domain, such as signal scattering, environmental attenuation, distance attenuation, etc., in the form of complex amplitude. CSI is used to measure the wireless signal to capture the target object information, thereby obtaining a second silhouette image of the target object.

[0056] Step S202: Determine whether the video frame is forged based on the similarity between the first silhouette image and the second silhouette image.

[0057] Based on the similarity between the first silhouette image and the second silhouette image of the same target object, it is determined whether the video frame is forged.

[0058] Through steps S201 and S202, a first silhouette image of the target object is obtained based on the video frame captured by the monitoring device; a second silhouette image of the target object is obtained based on the Wi-Fi signal captured by the Wi-Fi device; and based on the similarity between the first and second silhouette images, a determination is made as to whether the video frame has been forged. This solves the problems of complex, high computational overhead, and latency in video forgery detection in related technologies, achieving real-time, fine-grained, and low-computational-cost video forgery detection.

[0059] In one embodiment, acquiring a first silhouette image of the target object based on a video frame captured by a monitoring device includes:

[0060] The video frames captured by the monitoring device are input into a pre-trained instance segmentation model for classification, regression, and segmentation processing to obtain the first silhouette image of the target object.

[0061] Specifically, the instance segmentation model in this embodiment uses a Mask Region-based Convolutional Neural Network (Mask R-CNN), a machine learning model that performs object detection and segmentation, trained on a large number of video frame samples. Video frames captured by the monitoring device are used as input to the instance segmentation model. The pre-trained Mask R-CNN network performs classification, regression, and segmentation processing to output a first silhouette image of the target object.

[0062] In one embodiment, Figure 3 As shown, the obtaining of the second silhouette image of the target object based on the WIFI signal includes the following steps:

[0063] Step S301, acquiring the wavefront of the target object based on the WIFI signal;

[0064] Specifically, a two-dimensional inverse Fourier transform is applied to the WiFi signal to construct the wavefront of the target object.

[0065] Step S302: input the wavefront of the target object into a pre-trained semantic segmentation network model to extract a second silhouette image of the target object.

[0066] Specifically, the semantic segmentation network model in this embodiment is a pre-trained U-Net-based network model. The wavefront of the target object is input into the trained U-Net network model to extract the second silhouette image of the target object.

[0067] In one embodiment, Figure 4 As shown, the WIFI device includes a WIFI transmitter and a WIFI receiver. The WIFI transmitter is used to transmit WIFI signals, and the WIFI receiver is used to receive WIFI signals that reach the WIFI receiver after a series of reflection, transmission, and diffraction processes.

[0068] The step of acquiring the wavefront of the target object based on the WIFI signal comprises the following steps:

[0069] Step S401, as Figure 4-1 As shown, the WIFI receiver is taken as the coordinate origin O, the direction of the line connecting the WIFI receiver and the WIFI transmitter is taken as the X-axis direction, the direction perpendicular to the ground is taken as the Z-axis direction, and the Y-axis direction is perpendicular to the X-axis and the Z-axis respectively, to construct a spatial rectangular coordinate system;

[0070] Step S402: Based on equation (1), a two-dimensional inverse Fourier transform is used to calculate the wavefront s of each reflection point on the target object. p,q ;

[0071]

[0072] Wherein, S(t,f) is the WIFI signal of the target object received by the WIFI receiver at the subcarrier frequency f and time t, α is the amplitude attenuation constant, p and q are the coordinate values of the reflection point in the X-axis and Z-axis in the spatial rectangular coordinate system, respectively, the reflection plane where the reflection point is located is parallel to the coordinate plane XOZ, P and Q are the length and width of the reflection plane, j is a complex unit, and f c is the center frequency of the subcarrier, c is the propagation speed of the subcarrier, λ c is the center frequency f c The corresponding wavelength, is the initial displacement of the target object from the coordinate origin O at time t0, v y is the relative motion speed of the target object along the Y axis, γ is the azimuth angle of the reflection point relative to the coordinate origin, and β is the elevation angle of the reflection point relative to the coordinate origin;

[0073] Specifically, each point on the target object can reflect the WIFI signal. Take any reflection point Wp,q as the research object. The reflection plane where the reflection point Wp,q is located is parallel to the coordinate plane XOZ and has a size of P×Q. Let W p,q The coordinates of (x p，q ,y p，q , z p，q )=(p,dp ，d， q) then the coordinate vector is Azimuth γ = arctan(p / d p，q ), elevation angle Therefore, the coordinate vector of Wp,q can be expressed as:

[0074]

[0075] A virtual antenna is set at the reflection point Wp,q, which is equivalent to a WiFi signal transmitter with the signal strength and phase information at the reflection point. The signal strength and phase information of the reflection point are the wavefront s at the reflection point. p,q According to the theoretical model of signal propagation principle, the WIFI signal S(t,f) received by the WIFI receiver at the subcarrier frequency f and time t from the target object can be expressed as:

[0076]

[0077] Among them, α is the amplitude attenuation constant, and λ is the wavelength of the WIFI signal.

[0078] In order to express the wavefront phase change at each reflection point caused by signal propagation, a function related to each reflection point is defined, let the function The above formula is:

[0079]

[0080] The relative motion between the WIFI device and the target object is studied. Assume that the initial position of the target object at time t0 is: The direction at time t is: Its relative motion speed along the Y axis is v y , then the displacement of the reflection point Wp,q relative to the Y axis can be expressed as: For different reflection points, d in p,q is a variable, so the function K p,q The phase can be expressed as a function of time:

[0081]

[0082] In formula (5), the center frequency f c Wavelength λ c Replace λ, that is, λ=c / f c , c is the propagation speed of electromagnetic waves, then formula (5) can be expressed as:

[0083]

[0084] In formula (6), g(p,q) is used instead of s p,q , replace t with a and f with b c , let G(a,b)=S(t,f). Then formula (6) can be expressed as:

[0085]

[0086] Equation (7) is a standard two-dimensional inverse Fourier transform. Therefore, based on the known S(t,f), the two-dimensional inverse Fourier transform can be used to obtain g(p,q), that is, the wavefront s at the target reflection point p,q .

[0087] Step S403 : constructing a wavefront of the target object based on the wavefronts of each reflection point of the target object.

[0088] In one embodiment, Figure 5 As shown, the determining whether the video frame is forged based on the similarity between the first silhouette image and the second silhouette image includes the following steps:

[0089] Step S501: input the first silhouette image and the second silhouette image into two branches of the twin neural network respectively, and output a first semantic feature vector and a second semantic feature vector;

[0090] Specifically, the Siamese network has two branches, which have the same structure and parameters. The first silhouette image and the second silhouette image are used as inputs of the two branches respectively. In each branch, the feature extractor of the network uses three two-dimensional convolutional networks to mine the similar semantic information of the first silhouette image and the second silhouette image. Each convolutional network is followed by a batch normalization (BN) function and a linear activation function (ReLU). BN prevents the offset of data distribution, and ReLU reduces the dependency of neurons and improves the generalization ability of the feature extractor. At the end of the branch are two fully connected layers, which project the high-dimensional features into a low-dimensional semantic feature vector and output the first semantic feature vector and the second semantic feature vector. Among them, the first semantic feature vector is The second semantic feature vector

[0091] Step S502: Determine whether the video frame is forged based on the similarity between the first semantic feature vector and the second semantic feature vector.

[0092] Whether the video frame is forged is determined based on the similarity between the first semantic feature vector and the second semantic feature vector.

[0093] In one embodiment, Figure 6 As shown, the determining whether the video frame is forged based on the similarity between the first semantic feature vector and the second semantic feature vector includes the following steps:

[0094] Step S601: Calculate the Euclidean distance between the first semantic feature vector and the second semantic feature vector to determine the similarity; wherein the similarity is the reciprocal of the Euclidean distance;

[0095] Specifically, the Euclidean distance is used to quantify the similarity between the first semantic feature vector and the second semantic feature vector. The Euclidean distance D between the first semantic feature vector and the second semantic feature vector is: W (X1,X2) is calculated as follows:

[0096]

[0097] The similarity between the first semantic feature vector X1 and the second semantic feature vector X2 is the Euclidean distance D W The reciprocal of (X1,X2),

[0098] Step S602, determining whether the similarity is less than a preset threshold;

[0099] Step S603: If yes, it is determined that the video frame is forged;

[0100] Step S604: If not, determine that the video frame has not been forged.

[0101] In one embodiment, before acquiring the second silhouette image of the target object based on the WIFI signal, the method includes:

[0102] The WIFI signal is resampled so that the WIFI signal is aligned with the video frame in time, and the WIFI signal is filtered using a median filter, a mean filter, and a low-pass Butterworth filter in sequence to remove noise from the WIFI signal.

[0103] The present application also provides a video forgery detection device. Figure 7 As shown, the device includes:

[0104] An acquisition module 710 is configured to acquire a first silhouette image of a target object based on a video frame acquired by a monitoring device; and acquire a second silhouette image of the target object based on a Wi-Fi signal acquired by a Wi-Fi device;

[0105] The determination module 720 is configured to determine whether the video frame is forged based on the similarity between the first silhouette image and the second silhouette image.

[0106] The device provided in this embodiment obtains a first silhouette image of a target object based on video frames captured by a monitoring device and a second silhouette image of the target object based on Wi-Fi signals collected by a Wi-Fi device. Based on the similarity between the first and second silhouette images, the device determines whether the video frame has been forged. This solves the problems of complex, computationally expensive, and delayed video forgery detection in related technologies, achieving real-time, fine-grained, and computationally efficient video forgery detection.

[0107] The acquisition module 710 is further configured to input the video frames captured by the monitoring device into a pre-trained instance segmentation model to perform classification, regression, and segmentation processing to obtain a first silhouette image of the target object.

[0108] The acquisition module 710 is further configured to acquire a wavefront of the target object based on the WIFI signal; input the wavefront of the target object into a pre-trained semantic segmentation network model to extract a second silhouette image of the target object.

[0109] The WIFI device includes a WIFI transmitter and a WIFI receiver. The acquisition module 710 is further configured to construct a spatial rectangular coordinate system with the WIFI receiver as a coordinate origin O, a direction of a line connecting the WIFI receiver and the WIFI transmitter as an X-axis direction, a direction perpendicular to the ground as a Z-axis direction, and a Y-axis direction perpendicular to the X-axis and the Z-axis respectively.

[0110] Based on formula (1), the wavefront s of each reflection point on the target object is calculated using two-dimensional inverse Fourier transform p,q ;

[0111]

[0112] Wherein, S(t,f) is the WIFI signal of the target object received by the WIFI receiver at the subcarrier frequency f and time t, α is the amplitude attenuation constant, p and q are the coordinate values of the reflection point in the X-axis and Z-axis in the spatial rectangular coordinate system, respectively, the reflection plane where the reflection point is located is parallel to the coordinate plane XOZ, P and Q are the length and width of the reflection plane, j is a complex unit, and f c is the center frequency of the subcarrier, c is the propagation speed of the subcarrier, λ c is the center frequency f c The corresponding wavelength, is the initial displacement of the target object from the coordinate origin O at time t0, v y is the relative motion speed of the target object along the Y axis, γ is the azimuth angle of the reflection point relative to the coordinate origin, and β is the elevation angle of the reflection point relative to the coordinate origin;

[0113] The wavefront of the target object is constructed based on the wavefronts of each reflection point of the target object.

[0114] The determining module 720 is further configured to input the first silhouette image and the second silhouette image into two branches of the twin neural network respectively, and output a first semantic feature vector and a second semantic feature vector;

[0115] Based on the similarity between the first semantic feature vector and the second semantic feature vector, it is determined whether the video frame is forged.

[0116] The determining module 720 is further configured to calculate the Euclidean distance between the first semantic feature vector and the second semantic feature vector; and determine the similarity; wherein the similarity is the inverse of the Euclidean distance;

[0117] Determining whether the similarity is less than a preset threshold;

[0118] If so, it is determined that the video frame is forged;

[0119] If not, it is determined that the video frame has not been forged.

[0120] The acquisition module 710 is further configured to resample the WIFI signal so that the WIFI signal is temporally aligned with the video frame, and sequentially filter the WIFI signal using a median filter, a mean filter, and a low-pass Butterworth filter.

[0121] It should be noted that the above modules can be functional modules or program modules, and can be implemented through software or hardware. For modules implemented through hardware, the above modules can be located in the same processor; or the above modules can be located in different processors in any combination.

[0122] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a video forgery detection method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0123] Those skilled in the art will understand that Figure 8 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0124] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps of any of the above-mentioned message pushing method or message forwarding method embodiments are implemented.

[0125] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory or optical memory, etc. Volatile memory may include random access memory (RAM) or external cache memory. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM).

[0126] The technical features of the above-mentioned embodiments can be combined arbitrarily. In order to make the description concise, not all possible combinations of the technical features in the above-mentioned embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0127] The above-described embodiments merely represent several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present patent application shall be determined by the appended claims.

Claims

1. A video forgery detection method, characterized in that: The method comprises: Acquire a first silhouette image of the target object based on the video frames captured by the monitoring device; and Acquiring a second silhouette image of the target object based on a WIFI signal collected by a WIFI device; comprising: acquiring a wavefront of the target object based on the WIFI signal; inputting the wavefront of the target object into a pre-trained semantic segmentation network model to extract the second silhouette image of the target object; Based on the similarity between the first silhouette image and the second silhouette image, it is determined whether the video frame is forged.

2. The video forgery detection method according to claim 1, characterized in that: The acquiring of a first silhouette image of the target object based on the video frame captured by the monitoring device includes: The video frames captured by the monitoring device are input into a pre-trained instance segmentation model for classification, regression, and segmentation processing to obtain the first silhouette image of the target object.

3. The video forgery detection method according to claim 2, characterized in that: The WIFI device includes a WIFI transmitter and a WIFI receiver, and acquiring the wavefront of the target object based on the WIFI signal includes: A spatial rectangular coordinate system is constructed with the WIFI receiver as the coordinate origin O, the direction of the line connecting the WIFI receiver and the WIFI transmitter as the X-axis direction, the direction perpendicular to the ground as the Z-axis direction, and the Y-axis direction being perpendicular to the X-axis and the Z-axis respectively; Based on formula (1), the wavefront s of each reflection point on the target object is calculated using two-dimensional inverse Fourier transform p,q ; Wherein, S(t,f) is the WIFI signal of the target object received by the WIFI receiver at the subcarrier frequency f and time t, α is the amplitude attenuation constant, p and q are the coordinate values of the reflection point in the X-axis and Z-axis in the spatial rectangular coordinate system, respectively, the reflection plane where the reflection point is located is parallel to the coordinate plane XOZ, P and Q are the length and width of the reflection plane, j is a complex unit, and f c is the center frequency of the subcarrier, c is the propagation speed of the subcarrier, λ c is the center frequency f c The corresponding wavelength, is the initial displacement of the target object from the coordinate origin O at time t0, v y is the relative motion speed of the target object along the Y axis, γ is the azimuth angle of the reflection point relative to the coordinate origin, and β is the elevation angle of the reflection point relative to the coordinate origin; The wavefront of the target object is constructed based on the wavefronts of each reflection point of the target object.

4. The video forgery detection method according to claim 1, wherein: The determining whether the video frame is forged based on the similarity between the first silhouette image and the second silhouette image includes: Inputting the first silhouette image and the second silhouette image into two branches of the twin neural network respectively, and outputting a first semantic feature vector and a second semantic feature vector; Based on the similarity between the first semantic feature vector and the second semantic feature vector, it is determined whether the video frame is forged.

5. The video forgery detection method according to claim 4, characterized in that: The determining whether the video frame is forged based on the similarity between the first semantic feature vector and the second semantic feature vector includes: Calculating the Euclidean distance between the first semantic feature vector and the second semantic feature vector; determining the similarity; wherein the similarity is the reciprocal of the Euclidean distance; Determining whether the similarity is less than a preset threshold; If so, it is determined that the video frame is forged; If not, it is determined that the video frame has not been forged.

6. The video forgery detection method according to claim 1, characterized in that: Before acquiring the second silhouette image of the target object based on the WIFI signal, the method includes: The WIFI signal is resampled so that the WIFI signal is temporally aligned with the video frame, and the WIFI signal is filtered using a median filter, a mean filter, and a low-pass Butterworth filter in sequence.

7. A video forgery detection device, characterized in that: The device comprises: An acquisition module is configured to acquire a first silhouette image of a target object based on a video frame captured by a monitoring device; and acquire a second silhouette image of the target object based on a Wi-Fi signal captured by a Wi-Fi device; the acquisition module includes: acquiring a wavefront of the target object based on the Wi-Fi signal; and inputting the wavefront of the target object into a pre-trained semantic segmentation network model to extract the second silhouette image of the target object. A determination module is configured to determine whether the video frame is forged based on a similarity between the first silhouette image and the second silhouette image.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the method according to any one of claims 1 to 6 is implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.

Images