Data processing method, device and equipment

Abstract

The embodiment of the invention discloses a data processing method, device and equipment. The method comprises the following steps: acquiring a video data set; wherein the video data set comprises video data acquired through different video acquisition platforms, and the video data set relates to real video data and forged video data; setting an attribute value of each piece of video data in the video data set for a set video attribute; performing data annotation on the video data with the attribute values in the video data set to obtain training samples corresponding to the video data, and forming a training sample set; wherein the training samples corresponding to the video data with different types of attribute values are used for training different video forgery detection models.

Description

Technical Field

[0001] The present disclosure relates to the field of data processing technology, and more specifically, to a data processing method, device and equipment. Background Art

[0002] In the related art, deep learning image forgery algorithms can generate forged images that are close to real images, so deep learning image forgery algorithms can be applied to different fields. However, due to the advancement of deep learning image forgery algorithms, realistic images that are increasingly difficult to distinguish between true and false have appeared, which may lead to social suspicion of authenticity and credibility. Therefore, forgery detection of images can be performed based on forgery detection models.

[0003] Currently, when training forgery detection models, they often rely on limited data sets, which results in the trained forgery detection models having a single usage scenario and poor detection accuracy. Summary of the invention

[0004] The embodiments of the present disclosure provide a data processing method, apparatus and device.

[0005] According to a first aspect of the present disclosure, a method for detecting face forgery is provided, the method comprising:

[0006] Acquire a video data set; wherein the video data set includes video data collected through different video collection platforms, and the video data set involves real video data and forged video data;

[0007] Setting an attribute value of each of the video data in the video data set for a set video attribute;

[0008] Performing data labeling on the video data set having the attribute value, obtaining training samples corresponding to the video data, and forming a training sample set;

[0009] The training samples corresponding to the video data of different types of attribute values ​​are used for training different video forgery detection models.

[0010] Optionally, acquiring the video data set includes:

[0011] Get a list of search keywords;

[0012] In response to a set trigger event, based on the search keywords involved in the search keyword list, video data collection is performed on different video collection platforms to obtain the video data set.

[0013] Optionally, the different video collection platforms include at least two of a short video platform, a media information platform and a social platform.

[0014] Optionally, the set video attribute includes at least one of video quality, video content and video format.

[0015] Optionally, the step of performing data labeling on the video data set having the attribute value set therein to obtain training samples corresponding to the video data to form a training sample set includes:

[0016] Performing video frame extraction processing on the video data set having the attribute value set therein to obtain a first video frame sequence corresponding to the video data;

[0017] Performing image enhancement processing on a first video frame sequence corresponding to the video data to obtain a second video frame sequence corresponding to the video data;

[0018] Performing noise removal processing on a second video frame sequence corresponding to the video data to obtain a third video frame sequence corresponding to the video data;

[0019] Data annotation is performed on a third video frame sequence corresponding to the video data in the video data set to obtain corresponding training samples to form the training sample set.

[0020] Optionally, the video forgery detection model involves one of the network architectures of a convolutional neural network, a survivable adversarial network, a Transformer network, and a variational autoencoder network;

[0021] Furthermore, different video forgery detection models involve different network architectures and / or training methods.

[0022] Optionally, the method further comprises:

[0023] Provide input interface;

[0024] Obtaining a target attribute value, a target network architecture, and a target training method input through the input interface;

[0025] Determining a target training sample set from the training sample set according to the target attribute value;

[0026] Training is performed based on the target network architecture according to the target training sample set and the target training method to obtain the corresponding video forgery detection model.

[0027] Optionally, the method further comprises:

[0028] Storing the training sample set and the trained video forgery detection model in a set database;

[0029] Get new video data;

[0030] Setting the attribute value of the new video data for the set video attribute, and performing data annotation on the new video data with the attribute value set to obtain a corresponding new training sample, and storing the new training sample in the set database;

[0031] According to the new training sample, the corresponding video forgery detection model in the setting database is updated.

[0032] According to a second aspect of the present disclosure, a data processing device is provided, the device comprising:

[0033] An acquisition module, used to acquire a video data set; wherein the video data set includes video data collected through different video acquisition platforms, and the video data set involves real video data and forged video data;

[0034] A setting module, used for setting an attribute value of each video data in the video data set for a set video attribute;

[0035] A labeling module, used for labeling the video data with the attribute value in the video data set, obtaining training samples corresponding to the video data, and forming a training sample set;

[0036] The training samples corresponding to the video data of different types of attribute values ​​are used for training different video forgery detection models.

[0037] According to a third aspect of the present disclosure, an electronic device is provided, comprising a memory and a processor, wherein the memory is used to store an executable computer program; and the computer program is used to control the processor to execute the method according to the first aspect of the present disclosure.

[0038] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the method described in the first aspect of the present disclosure is implemented.

[0039] According to the data processing method of the embodiment of the present disclosure, video data can be collected through different video collection platforms to obtain a video data set, and the collected video data set involves real video data and forged video data, and the attribute value of each video data in the video data set for the set video attribute is set, and then the video data with the attribute value set in the video data set is data labeled to obtain training samples corresponding to the video data to form a training sample set, and the training samples corresponding to video data with different types of attribute values ​​can be used for the training of different video forgery detection models. Through the embodiment of the present disclosure, it collects video data through multiple channels, and the video data set involves real video data and forged video data, so as to construct a diversified video data set, and the training samples corresponding to video data with different types of attribute values ​​can be used for the training of different video forgery detection models, so as to ensure the robustness of the video forgery detection model in different scenarios.

[0040] Other features and advantages of the present disclosure will become apparent from the following detailed description of exemplary embodiments of the present disclosure with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] The accompanying drawings, which are incorporated in and constitute a part of the specification, illustrate embodiments of the present disclosure and, together with the description, serve to explain the principles of the present disclosure.

[0042] Figure 1 is a schematic diagram of the hardware configuration of an electronic device according to an embodiment of the present disclosure;

[0043] Figure 2 is a flowchart of a data processing method according to an embodiment of the present disclosure;

[0044] Figure 3 is a principle block diagram of a data processing device according to an embodiment of the present disclosure;

[0045] Figure 4 is a schematic diagram of the hardware configuration of an electronic device according to another embodiment of the present disclosure. DETAILED DESCRIPTION

[0046] Various exemplary embodiments of the present disclosure will now be described in detail with reference to the accompanying drawings. It should be noted that the relative arrangement of components and steps, numerical expressions and numerical values ​​set forth in these embodiments do not limit the scope of the present disclosure unless otherwise specifically stated.

[0047] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present disclosure, its application, or uses.

[0048] Technologies, methods, and equipment known to ordinary technicians in the relevant art may not be discussed in detail, but where appropriate, the technologies, methods, and equipment should be considered as part of the specification.

[0049] In all examples shown and discussed herein, any specific values ​​should be interpreted as merely exemplary and not limiting. Therefore, other examples of the exemplary embodiments may have different values.

[0050] It should be noted that like reference numerals and letters refer to similar items in the following figures, and therefore, once an item is defined in one figure, it need not be further discussed in subsequent figures.

[0051] <Hardware Configuration>

[0052] Figure 1 is a block diagram of a hardware configuration of an electronic device 1000 according to an embodiment of the present disclosure.

[0053] The electronic device 1000 may be a terminal device, a portable computer, a desktop computer, a server, a server cluster, etc. Figure 1 As shown, the electronic device 1000 may include a processor 1100, a memory 1200, an interface device 1300, a communication device 1400, a display device 1500, an input device 1600, a speaker 1700, a microphone 1800, and the like. Among them, the processor 1100 may be a processor CPU, a microprocessor MCU, and the like. The memory 1200 includes, for example, a ROM (read-only memory), a RAM (random access memory), a non-volatile memory such as a hard disk, and the like. The interface device 1300 includes, for example, a USB interface, a headphone interface, and the like. The communication device 1400 is, for example, capable of wired or wireless communication, and may specifically include Wifi communication, Bluetooth communication, 2G / 3G / 4G / 5G communication, and the like. The display device 1500 is, for example, a liquid crystal display screen, a touch display screen, and the like. The input device 1600 may include, for example, a touch screen, a keyboard, a somatosensory input, and the like. The user may input / output voice information through the speaker 1700 and the microphone 1800.

[0054] Figure 1 The electronic device shown is merely illustrative and does not in any way imply any limitation on the present disclosure, its application or use. In the embodiments of the present disclosure, the memory 1200 of the electronic device 1000 is used to store instructions, and the instructions are used to control the processor 1100 to operate to perform any data processing method provided in the embodiments of the present disclosure. It should be understood by those skilled in the art that although Figure 1In the electronic device 1000, multiple devices are shown, but the present disclosure may only involve some of the devices, for example, the electronic device 1000 only involves the processor 1100 and the storage device 1200. A technician can design instructions according to the scheme disclosed in the present disclosure. How instructions control the processor to operate is well known in the art, so it will not be described in detail here.

[0055] <Method Example>

[0056] In this embodiment, a data processing method is provided. The data processing method can be implemented by an electronic device. The electronic device can be as follows: Figure 1 The electronic device 1000 shown may be a terminal device, a portable computer, a desktop computer, a server, a server cluster, etc.

[0057] according to Figure 2 As shown, the face forgery detection method of the embodiment of the present disclosure may include the following steps S2100 to S2300.

[0058] Step S2100, obtaining a video data set.

[0059] The video data set includes video data collected through different video collection platforms. The different video collection platforms include at least two of short video platforms, media information platforms and social platforms. The short video platform may include a domestic short video platform or a foreign short video platform, the social platform may include a domestic social platform or a foreign social platform, and the media information platform may include a domestic media information platform or a foreign media information platform.

[0060] Generally, the video data involved in a video dataset should include both real video data and forged video data to ensure the diversity and representativeness of the video dataset.

[0061] In an optional embodiment, the step S2100 of acquiring the video data set may further include the following steps S2110 to S2120:

[0062] Step S2110, obtaining a search keyword list.

[0063] Among them, the search keyword list includes at least one search keyword. The search keywords in the search keyword list can be pre-set according to actual needs and actual scenarios. This embodiment does not limit the method for obtaining the search keywords in the search keyword list.

[0064] Step S2120, in response to a set trigger event, based on the search keywords involved in the search keyword list, video data collection is performed on the different video collection platforms to obtain the video data set.

[0065] The set trigger event may be at least one of receiving a user input and reaching a set time.

[0066] The above-set time may be, for example, 10:00 am every day. Then, at 10:00 am every day, the electronic device may collect video data on different video collection platforms based on the search keywords involved in the search keyword list.

[0067] The above-mentioned user input can be a user's touch input, such as a click input. For example, the display interface of the electronic device displays a video acquisition control. If the user clicks the video acquisition control, the electronic device can perform video data acquisition on different video acquisition platforms based on the search keywords involved in the search keyword list.

[0068] In this step S2120, a trigger event is pre-set in the electronic device. When the trigger event occurs, video data can be collected by crawling on different video collection platforms based on the search keywords involved in the search keyword list to obtain a video data set.

[0069] According to the above steps S2110 to S2120, video data from different channels can be obtained, and the video data includes real video data and forged video data to ensure the diversity and representativeness of the video data set.

[0070] After executing the above step S2100 to obtain the video data set, proceed to:

[0071] Step S2200: setting an attribute value of each video data in the video data set for a set video attribute.

[0072] The setting of the video attributes includes at least two of the video quality, the video content and the video format.

[0073] In this embodiment, after acquiring the video data set, the electronic device can set the attribute value of each video data in the video data set for the above-mentioned set video attribute. This step can also be understood as video data classification, that is, video data with different attribute values ​​can be classified into one category of video data.

[0074] Exemplarily, the video data set involves video data 1, video data 2, video data 3 and video 4, and the set video attributes include video quality, video content and video format. For example, the electronic device can set the attribute value of video data 1 for the set video attributes to be "video quality: high, video content: quarrel, video format: RMVB"; for another example, the electronic device can set the attribute value of video data 2 for the set video attributes to be "video quality: high, video content: raining, video format: AVI"; for another example, the electronic device can set the attribute value of video data 3 for the set video attributes to be "video quality: low, video content: quarreling, video format: RMVB"; for another example, the electronic device can set the attribute value of video data 4 for the set video attributes to be "video quality: low, video content: raining, video format: AVI".

[0075] It can be understood that if the video data set is classified by video quality, video data 1 and video data 2 belong to one category, and video data 3 and video data 4 belong to one category. If the video data set is classified by whether the video content is a quarrel, video data 1 and video data 3 belong to one category, and video data 2 and video data 4 belong to one category. If the video data set is classified by whether the video format is RMVB, video data 1 and video data 3 belong to one category, and video data 2 and video data 4 belong to one category.

[0076] After executing the above step S2200 to set the attribute value of each video data in the video data set for setting the video attribute, proceed to:

[0077] Step S2300, performing data labeling on the video data set having the attribute value, obtaining training samples corresponding to the video data, and forming a training sample set;

[0078] In an optional embodiment, this step S2300 performs data annotation on the video data set having the attribute value, to obtain training samples corresponding to the video data, and forming the training sample set may further include the following steps S2310 to S2340:

[0079] Step S2310: Perform video frame extraction processing on the video data set having the attribute value set therein to obtain a first video frame sequence corresponding to the video data.

[0080] The video frame extraction process is used to decompose the video data into independent video frames.

[0081] In this step S2310, after the electronic device sets the attribute value corresponding to the set video attribute for each video data in the video data set, it can perform video frame extraction processing on any video data to obtain a first video frame sequence corresponding to the video data, so as to decompose the video data into independent video frames.

[0082] Step S2320: Perform image enhancement processing on the first video frame sequence corresponding to the video data to obtain a second video frame sequence corresponding to the video data.

[0083] The image enhancement processing is used to improve the quality of the video frames, for example, to improve the brightness adjustment and contrast enhancement of the video frames.

[0084] In this step S2320, the electronic device performs video frame extraction processing on any video data to obtain a first video frame sequence corresponding to the video data, and then performs image enhancement processing on each video frame in the first video frame sequence to obtain a second video frame sequence corresponding to the video data.

[0085] Step S2330: performing noise removal processing on the second video frame sequence corresponding to the video data to obtain a third video frame sequence corresponding to the video data.

[0086] The noise removal process is used to remove background noise in the video frame so as to enhance the clarity of the image.

[0087] In this step S2330, the electronic device performs image enhancement processing on the first video frame sequence corresponding to any video data to obtain a second video frame sequence corresponding to the video data, and then performs noise removal processing on each video frame in the second video frame sequence to obtain a third video frame sequence corresponding to the video data.

[0088] Step S2340: perform data annotation on a third video frame sequence corresponding to the video data in the video data set to obtain corresponding training samples to form the training sample set.

[0089] In this step S2340, the electronic device can annotate the third video frame corresponding to any video data by manual or semi-automatic tools to distinguish real video data from forged video data, and these annotated data will be used for training the video forgery detection model. Generally, annotating the third video frame sequence corresponding to any video data based on semi-automatic tools can improve the annotation efficiency.

[0090] The training samples corresponding to the video data of different types of attribute values ​​are used for training different video forgery detection models.

[0091] The video forgery detection model involves a network architecture selected from the group consisting of a convolutional neural network, a survivable adversarial network, a transformer network, and a variational autoencoder network. In addition, different video forgery detection models involve different network architectures and / or training methods.

[0092] Here, the data processing method of the embodiment of the present disclosure further includes the following steps S3100 to S3400:

[0093] Step S3100, providing an input interface.

[0094] The input interface may be a text input box, a voice input box, etc., which is not limited in this embodiment.

[0095] Step S3200, obtaining the target attribute value, target network architecture and target training method input through the input interface.

[0096] In this embodiment, the user can input the type of training samples, network architecture and training method of the video forgery detection model to be trained through the input interface according to his own wishes. The electronic device can use the type of training samples selected by the user as the target attribute value, the network architecture selected by the user as the target network architecture, and the training method selected by the user as the target training method.

[0097] For example, the user can manually input the target attribute value "video content: noisy", the target network architecture "variational autoencoder", and the target training method "deep learning".

[0098] Step S3300: determining a target training sample set from the training sample set according to the target attribute value.

[0099] Continuing with the above example, if the target attribute value is "video content: noisy", the electronic device can combine the training data 1 corresponding to the above video data 1 and the training samples corresponding to the video data 3 to form a target training sample set.

[0100] Step S3400: Perform training based on the target network architecture according to the target training sample set and the target training method to obtain the corresponding video forgery detection model.

[0101] Continuing with the above example, the electronic device can perform deep learning on the variational autoencoder based on the target training sample set to obtain a corresponding video forgery detection model.

[0102] According to the method of the embodiment of the present disclosure, video data can be collected through different video acquisition platforms to obtain a video data set, and the collected video data set involves real video data and forged video data, and the attribute value of each video data in the video data set for the set video attribute is set, and then the video data with the attribute value set in the video data set is data labeled to obtain training samples corresponding to the video data to form a training sample set, and the training samples corresponding to video data with different types of attribute values ​​can be used for the training of different video forgery detection models. Through the embodiment of the present disclosure, it collects video data through multiple channels, and the video data set involves real video data and forged video data, so as to construct a diversified video data set, and the training samples corresponding to video data with different types of attribute values ​​can be used for the training of different video forgery detection models, so as to ensure the robustness of the video forgery detection model in different scenarios.

[0103] In one embodiment, the data processing method of the embodiment of the present disclosure further includes the following steps S4100 to S4400:

[0104] Step S4100: storing the training sample set and the trained video forgery detection model in a setting database.

[0105] The setting database may be at least one of a relational database and a distributed file system, so as to facilitate efficient storage and rapid retrieval of data.

[0106] In this embodiment, the electronic device may store the training sample set and the trained video forgery detection model in a setting database.

[0107] Step S4200, obtaining new video data.

[0108] In this embodiment, the electronic device can obtain new video data when the set second time arrives. It should be noted that the set second time is usually different from the set first time.

[0109] Step S4300, setting the attribute value of the new video data for the set video attribute, and performing data annotation on the new video data with the attribute value to obtain corresponding new training samples, and storing the new training samples in the set database.

[0110] In this embodiment, the electronic device can set the attribute value of the set video attribute for the new video data, and first perform frame extraction processing on the new video data with the attribute value set to obtain a first video frame sequence corresponding to the new video data, then perform image enhancement processing on the first video frame sequence corresponding to the new video data to obtain a second video frame sequence corresponding to the new video data, and remove noise from the second video frame sequence corresponding to the new video data to obtain a third video frame sequence corresponding to the new video data, and finally perform data annotation on the third video frame sequence corresponding to the new video data to obtain the corresponding new training sample.

[0111] Step S4400: updating the corresponding video forgery detection model in the setting database according to the new training sample.

[0112] In this embodiment, the electronic device can update the corresponding video forgery detection model in the set database according to the new training sample.

[0113] Through this embodiment, the electronic device can regularly collect new video data and update the corresponding video forgery detection model based on the new video data, which can ensure the timeliness and reliability of the set database, not only provide historical data support, but also maintain efficient forgery detection capabilities in practical applications.

[0114] In one embodiment, the data processing method of the embodiment of the present disclosure further includes: providing a viewing interface for a user to search for required video data or a trained video forgery model in a set database.

[0115] <Example>

[0116] Next, an example of a data processing method is shown. In this example, the data processing method may further include:

[0117] Step 1: obtain a search keyword list, and in response to a set trigger event, based on the search keywords involved in the search keyword list, perform video data collection on different video collection platforms to obtain a video data set.

[0118] Step 2, setting the attribute value of the set video attribute for each video data in the video data set, and performing video frame extraction processing on the video data with the attribute value set in the video data set to obtain a first video frame sequence corresponding to the video data, and performing image enhancement processing on the first video frame sequence corresponding to the video data to obtain a second video frame sequence corresponding to the video data, and performing noise removal processing on the second video frame sequence corresponding to the video data to obtain a third video frame sequence corresponding to the video data, and finally performing data annotation on the third video frame sequence corresponding to the video data in the video data set to obtain corresponding training samples to form a training sample set.

[0119] Step 3: Provide an input interface, and obtain the target attribute value, target network architecture, and target training method input through the input interface, and determine the target training sample set from the training sample set based on the target attribute value, and then train the target network architecture based on the target training sample set and the target training method to obtain the corresponding video forgery detection model.

[0120] Step 4: Store the training sample set and the trained video forgery detection model in a set database.

[0121] Step 5, obtain new video data and store it in the set database, set the attribute value of the new video data for the set video attribute, and perform data annotation on the new video data with the set attribute value to obtain the corresponding new training sample, and store the new training sample in the set database, and update the corresponding video forgery detection model in the set database according to the new training sample.

[0122] <Device Example>

[0123] In this embodiment, a data processing device 300 is provided. Figure 3 As shown, the data processing device 300 may include an acquisition module 310 , a setting module 320 and a labeling module 330 .

[0124] An acquisition module 310 is used to acquire a video data set; wherein the video data set includes video data collected through different video acquisition platforms, and the video data set involves real video data and forged video data;

[0125] A setting module 320, configured to set an attribute value of each video data in the video data set for a set video attribute;

[0126] A labeling module 330 is used to label the video data with the attribute value in the video data set, obtain training samples corresponding to the video data, and form a training sample set;

[0127] The training samples corresponding to the video data of different types of attribute values ​​are used for training different video forgery detection models.

[0128] In one embodiment, the acquisition module 310 is specifically used to obtain a search keyword list; in response to a set trigger event, based on the search keywords involved in the search keyword list, video data acquisition is performed on different video acquisition platforms to obtain the video data set.

[0129] In one embodiment, the different video collection platforms include at least two of a short video platform, a media information platform, and a social platform.

[0130] In one embodiment, the setting of the video attribute includes at least one of video quality, video content and video format.

[0131] In one embodiment, the labeling module 330 is specifically used to perform video frame extraction processing on the video data set with the attribute value to obtain a first video frame sequence corresponding to the video data; perform image enhancement processing on the first video frame sequence corresponding to the video data to obtain a second video frame sequence corresponding to the video data; perform noise removal processing on the second video frame sequence corresponding to the video data to obtain a third video frame sequence corresponding to the video data; perform data labeling on the third video frame sequence corresponding to the video data in the video data set to obtain corresponding training samples to form the training sample set.

[0132] In one embodiment, the video forgery detection model involves a network architecture selected from the group consisting of a convolutional neural network, a survivable adversarial network, a transformer network, and a variational autoencoder network;

[0133] Furthermore, different video forgery detection models involve different network architectures and / or training methods.

[0134] In one embodiment, the apparatus 300 further includes a providing module, a determining module, and an updating module (all not shown in the figure).

[0135] Providing a module for providing an input interface;

[0136] The acquisition module 310 is further used to acquire the target attribute value, target network architecture and target training method input through the input interface;

[0137] A determination module, configured to determine a target training sample set from the training sample set according to the target attribute value;

[0138] A training module is used to perform training based on the target network architecture according to the target training sample set and the target training method to obtain the corresponding video forgery detection model.

[0139] According to this embodiment, video data can be collected through different video collection platforms to obtain a video data set, and the collected video data set involves real video data and forged video data, and the attribute value of each video data in the video data set for the set video attribute is set, and then the video data with the attribute value set in the video data set is data labeled to obtain training samples corresponding to the video data to form a training sample set, and the training samples corresponding to video data with different types of attribute values ​​can be used for the training of different video forgery detection models. Through the disclosed embodiment, it collects video data through multiple channels, and the video data set involves real video data and forged video data, so as to construct a diversified video data set, and the training samples corresponding to video data with different types of attribute values ​​can be used for the training of different video forgery detection models, which can ensure the robustness of the video forgery detection model in different scenarios.

[0140] <Equipment Embodiment>

[0141] Figure 4 FIG. 1 is a schematic diagram of the hardware structure of an electronic device according to an embodiment. Figure 4 As shown, the electronic device 400 includes a processor 410 and a memory 420 .

[0142] The memory 420 may be used to store executable computer instructions.

[0143] The processor 410 can be used to execute the data processing method described in the embodiment of the method of the present disclosure under the control of the executable computer instructions.

[0144] The electronic device 400 may be Figure 1 The electronic device 1000 shown may also be a device with other hardware structures, which is not limited here.

[0145] In another embodiment, the electronic device 400 may include the above data processing device 300 .

[0146] In one embodiment, each module of the above data processing device 300 can be implemented by the processor 410 running computer instructions stored in the memory 420 .

[0147] <Computer Readable Storage Medium>

[0148] The embodiment of the present disclosure further provides a computer-readable storage medium on which computer instructions are stored. When the computer instructions are executed by a processor, the data processing method provided by the embodiment of the present disclosure is executed.

[0149] The present disclosure may be a system, a method and / or a computer program product. The computer program product may include a computer-readable storage medium carrying computer-readable program instructions for causing a processor to implement various aspects of the present disclosure.

[0150] A computer-readable storage medium may be a tangible device that can hold and store instructions used by an instruction execution device. A computer-readable storage medium may be, for example, but not limited to, an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples of computer-readable storage media (a non-exhaustive list) include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disk read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanical encoding device, such as a punch card or a raised structure in a groove on which instructions are stored, and any suitable combination of the foregoing. As used herein, a computer-readable storage medium is not to be interpreted as a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagating through a waveguide or other transmission medium (e.g., a light pulse through a fiber optic cable), or an electrical signal transmitted through a wire.

[0151] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to each computing / processing device, or downloaded to an external computer or external storage device via a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network can include copper transmission cables, optical fiber transmissions, wireless transmissions, routers, firewalls, switches, gateway computers, and / or edge servers. The network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in the computer-readable storage medium in each computing / processing device.

[0152] The computer program instructions for performing the operation of the present disclosure may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine-related instructions, microcode, firmware instructions, state setting data, or source code or object code written in any combination of one or more programming languages, including object-oriented programming languages, such as Smalltalk, C++, etc., and conventional procedural programming languages, such as "C" language or similar programming languages. Computer-readable program instructions may be executed completely on a user's computer, partially on a user's computer, as an independent software package, partially on a user's computer, partially on a remote computer, or completely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer via any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., using an Internet service provider to connect via the Internet). In some embodiments, an electronic circuit, such as a programmable logic circuit, a field programmable gate array (FPGA), or a programmable logic array (PLA), may be customized by utilizing the state information of the computer-readable program instructions, and the electronic circuit may execute the computer-readable program instructions, thereby realizing various aspects of the present disclosure.

[0153] Various aspects of the present disclosure are described herein with reference to the flowcharts and / or block diagrams of the methods, devices (systems) and computer program products according to the embodiments of the present disclosure. It should be understood that each box in the flowchart and / or block diagram and the combination of each box in the flowchart and / or block diagram can be implemented by computer-readable program instructions.

[0154] These computer-readable program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, thereby producing a machine, so that when these instructions are executed by the processor of the computer or other programmable data processing device, a device that implements the functions / actions specified in one or more boxes in the flowchart and / or block diagram is generated. These computer-readable program instructions can also be stored in a computer-readable storage medium, and these instructions cause the computer, programmable data processing device, and / or other equipment to work in a specific manner, so that the computer-readable medium storing the instructions includes a manufactured product, which includes instructions for implementing various aspects of the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0155] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device so that a series of operating steps are performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process, thereby causing the instructions executed on the computer, other programmable data processing apparatus, or other device to implement the functions / actions specified in one or more boxes in the flowchart and / or block diagram.

[0156] The flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the systems, methods and computer program products according to multiple embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of an instruction, and the module, a program segment or a part of an instruction contains one or more executable instructions for realizing the specified logical function. In some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or the flowchart, and the combination of the boxes in the block diagram and / or the flowchart can be implemented by a dedicated hardware-based system that performs the specified function or action, or can be implemented by a combination of dedicated hardware and computer instructions. It is well known to those skilled in the art that it is equivalent to implement it by hardware, implement it by software, and implement it by combining software and hardware.

[0157] The embodiments of the present disclosure have been described above, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The choice of terms used herein is intended to best explain the principles of the embodiments, practical applications, or technical improvements in the market, or to enable other ordinary skill in the art to understand the embodiments disclosed herein. The scope of the present disclosure is defined by the appended claims. .

Claims

1. A data processing method, characterized in that: The method comprises: Acquire a video data set; wherein the video data set includes video data collected through different video collection platforms, and the video data set involves real video data and forged video data; Setting an attribute value of each of the video data in the video data set for a set video attribute; Performing data labeling on the video data set having the attribute value, obtaining training samples corresponding to the video data, and forming a training sample set; The training samples corresponding to the video data of different types of attribute values ​​are used for training different video forgery detection models.

2. The method according to claim 1, characterized in that The step of obtaining a video data set includes: Get a list of search keywords; In response to a set trigger event, based on the search keywords involved in the search keyword list, video data collection is performed on different video collection platforms to obtain the video data set.

3. The method according to claim 1 or 2, characterized in that: The different video collection platforms include at least two of a short video platform, a media information platform and a social platform.

4. The method according to claim 1 or 2, characterized in that: The set video attribute includes at least one of video quality, video content and video format.

5. The method according to claim 1, characterized in that in, The step of labeling the video data in the video data set with the attribute value to obtain training samples corresponding to the video data to form a training sample set includes: Performing video frame extraction processing on the video data set having the attribute value set therein to obtain a first video frame sequence corresponding to the video data; Performing image enhancement processing on a first video frame sequence corresponding to the video data to obtain a second video frame sequence corresponding to the video data; Performing noise removal processing on a second video frame sequence corresponding to the video data to obtain a third video frame sequence corresponding to the video data; Data annotation is performed on a third video frame sequence corresponding to the video data in the video data set to obtain corresponding training samples to form the training sample set.

6. The method according to claim 1, characterized in that The video forgery detection model involves one of the network architectures of a convolutional neural network, a survivable adversarial network, a Transformer network, and a variational autoencoder network; Furthermore, different video forgery detection models involve different network architectures and / or training methods.

7. The method according to claim 1, characterized in that The method further comprises: Provide input interface; Obtaining a target attribute value, a target network architecture, and a target training method input through the input interface; Determining a target training sample set from the training sample set according to the target attribute value; Training is performed based on the target network architecture according to the target training sample set and the target training method to obtain the corresponding video forgery detection model.

8. The method according to claim 1, characterized in that The method further comprises: Storing the training sample set and the trained video forgery detection model in a set database; Get new video data; Setting the attribute value of the new video data for the set video attribute, and performing data annotation on the new video data with the attribute value set to obtain a corresponding new training sample, and storing the new training sample in the set database; According to the new training sample, the corresponding video forgery detection model in the setting database is updated.

9. A data processing device, characterized in that: The device comprises: An acquisition module, used to acquire a video data set; wherein the video data set includes video data collected through different video acquisition platforms, and the video data set involves real video data and forged video data; A setting module, used for setting an attribute value of each video data in the video data set for a set video attribute; A labeling module, used for labeling the video data with the attribute value in the video data set, obtaining training samples corresponding to the video data, and forming a training sample set; The training samples corresponding to the video data of different types of attribute values ​​are used for training different video forgery detection models.

10. An electronic device, characterized in that: The method comprises a memory and a processor, wherein the memory is used to store an executable computer program; and the computer program is used to control the processor to execute the method according to any one of claims 1 to 8.

Images