Video detection processing method and device

Abstract

The embodiment of the invention provides a video detection processing method and device, and the method comprises the steps: carrying out the object detection and image extraction of a video frame of a to-be-detected video in a detection process of the to-be-detected video, and obtaining an object image, and on the basis of performing differential processing on the object image, performing feature extraction on the obtained differential image to obtain a differential feature sequence, splicing the video feature sequence of each video frame with the differential feature sequence, and inputting the spliced feature sequence into a video detection module to perform detection of a machine-generated video to obtain a detection result. Therefore, whether the to-be-detected video is the machine-generated video or not can be detected.

Description

Technical Field

[0001] This document relates to the field of video processing technology, and in particular to a video detection processing method and device. Background Art

[0002] With the continuous development and promotion of the Internet and artificial intelligence technologies, the application scope of various models obtained through machine learning is becoming wider and wider, and has gradually covered many fields. In this case, not only various forms of data processing models have emerged, but also various large language models, such as text generation models, code generation models, image generation models, video generation models and other generative models, which can efficiently generate content. However, generative models may bring some potential risks in the process of content generation. For example, images generated by image generation models may have a high degree of similarity with existing images, which may lead to copyright conflicts. Summary of the invention

[0003] One or more embodiments of the present specification provide a video detection processing method, comprising: performing object detection and image extraction on each video frame of a video to be detected to obtain an object image. Performing differential processing on the object image, and performing feature extraction on the obtained differential image to obtain a differential feature sequence. Splicing the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence. Inputting the spliced ​​feature sequence into a video detection module to perform machine-generated video detection to obtain a detection result.

[0004] One or more embodiments of the present specification provide a video detection and processing device, including: a detection and extraction module, configured to perform object detection and image extraction on each video frame of a video to be detected, and obtain an object image. A differential processing module, configured to perform differential processing on the object image, and perform feature extraction on the obtained differential image to obtain a differential feature sequence. A splicing module, configured to splice the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence. A video detection module, configured to input the spliced ​​feature sequence into a video detection module to perform detection of a machine-generated video and obtain a detection result.

[0005] One or more embodiments of the present specification provide a video detection processing device, including: a processor; and a memory configured to store computer executable instructions, wherein when the computer executable instructions are executed, the processor: performs object detection and image extraction on each video frame of the video to be detected to obtain an object image. Perform differential processing on the object image, and perform feature extraction on the obtained differential image to obtain a differential feature sequence. Splice the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence. Input the spliced ​​feature sequence into a video detection module to detect machine-generated video and obtain a detection result.

[0006] One or more embodiments of the present specification provide a computer-readable storage medium for storing computer-executable instructions, which implement the following process when executed: performing object detection and image extraction on each video frame of the video to be detected to obtain an object image. Performing differential processing on the object image, and performing feature extraction on the obtained differential image to obtain a differential feature sequence. Splicing the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence. Inputting the spliced ​​feature sequence into a video detection module to perform machine-generated video detection to obtain a detection result. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] In order to more clearly illustrate one or more embodiments of the present specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present specification. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative labor. Figure 1 A schematic diagram of an implementation environment of a video detection processing method provided by one or more embodiments of this specification; Figure 2 A processing flow chart of a video detection processing method provided by one or more embodiments of this specification; Figure 3 A schematic diagram of a machine-generated video detection process provided for one or more embodiments of this specification; Figure 4 A processing flow chart of a video detection processing method applied to a video detection scenario provided by one or more embodiments of this specification; Figure 5 A schematic diagram of an embodiment of a video detection processing device provided by one or more embodiments of this specification; Figure 6 A schematic diagram of the structure of a video detection and processing device provided in one or more embodiments of this specification. DETAILED DESCRIPTION

[0008] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.

[0009] The video detection processing method provided in one or more embodiments of this specification can be applied to the implementation environment of video detection. Figure 1 , the implementation environment at least includes: A video encoding module 101 for encoding video frames of a video to be detected, a differential encoding module 102 for encoding differential images of a video to be detected, and a video detection module 103 for performing machine-generated video detection; The video encoding module 101 may adopt a network architecture such as a deep neural network or ViT (Vision Transformer), and the differential encoding module 102 may also adopt a network architecture such as a deep neural network or ViT (Vision Transformer); the video detection module 103 may adopt a Transformer network architecture; In this implementation environment, in the process of detecting whether the video to be detected is machine-generated video, on the one hand, feature extraction is performed on the video frame of the video to be detected by the video encoding module 101 to obtain a video feature sequence composed of video features of the video to be detected, and on the other hand, object detection and image extraction are performed on each video frame of the video to be detected to obtain an object image contained in the video frame, and differential processing is performed on the object image by the differential encoding module 102, and differential feature extraction is performed on the obtained differential image to obtain a differential feature sequence, on this basis, the video feature sequence and the differential feature sequence are spliced, and the spliced ​​feature sequence is input into the video detection module 103 for machine-generated video detection to obtain a detection result of whether the video to be detected is a machine-generated video, thereby realizing machine-generated video detection of the video to be detected.

[0010] One or more embodiments of a video detection processing method provided in this specification are as follows: Reference Figure 2 The video detection processing method provided in this embodiment specifically includes steps S202 to S208.

[0011] Step S202: performing object detection and image extraction on each video frame of the video to be detected to obtain an object image.

[0012] In this embodiment, the video to be detected refers to a video for machine-generated video detection, and the video to be detected contains multiple video frames. In the process of machine-generated video detection, machine-generated video detection is performed on the video frames contained in the video to be detected, thereby realizing machine-generated video detection of the video to be detected.

[0013] Specifically, in the process of performing machine-generated video detection on the video frames of the video to be detected, machine-generated video detection can be performed on all video frames contained in the video to be detected; in addition, considering that the contents of similar video frames in the video to be detected are relatively similar, in order to improve the detection efficiency of the video to be detected, the video frames contained in the video to be detected can also be sampled, and machine-generated video detection can be performed on the sampled video frames obtained by sampling.

[0014] Optionally, each video frame of the video to be detected is obtained in the following manner: uniformly sampling a video frame sequence of the video to be detected, and using each obtained sampled video frame as the video frame.

[0015] In the specific implementation, in the process of performing machine-generated video detection on each video frame of the video to be detected, machine-generated video detection is performed on each video frame of the video to be detected based on two dimensions: the video features of each video frame and the difference in object images contained in each video frame. Here, based on the dimension of the difference in object images contained in each video frame of the video to be detected, object detection and image extraction are performed on each video frame of the video to be detected to obtain the object image.

[0016] Among them, object detection performed on each video frame refers to detecting whether each video frame contains a corresponding object. The object can be a biological object (such as a person, animal, plant, etc.), a physical object, a building object or a landscape object, etc. It should be noted that biological objects often change over time due to their own biological nature, such as various types of changes such as movement, deformation or displacement; therefore, in addition to biological objects, in order to ensure the effectiveness of machine-generated video detection of the video to be detected, in the process of performing object detection of other object types other than biological objects on each video frame of the video to be detected, the detected object must undergo a certain degree of change in each video frame of the video to be detected, such as changes in at least two video frames of the video to be detected, such as various types of changes such as movement, deformation or displacement. Based on this, physical objects, building objects or landscape objects other than biological objects can be referred to as non-biological objects.

[0017] In the process of performing object detection and image extraction of biological objects on each video frame of the video to be detected, an optional implementation manner provided by this embodiment is to perform object detection and image extraction on each video frame of the video to be detected to obtain an object image, including: performing biological object detection on each video frame to obtain a biological object and an object video frame containing the biological object; and extracting an image block of the biological object from the object video frame as the object image.

[0018] For example, Figure 3 The schematic diagram of the machine-generated video detection process shown in FIG. 1 inputs a sampled video frame of a video stream as a video to be detected into an object detection and extraction module to perform biological object detection and image extraction to obtain a biological object image extracted from the sampled video frame.

[0019] In actual applications, in the process of performing biological object detection on each video frame, there may be a situation where the object organisms contained in different video frames are the same object organism, and there may also be a situation where the object organisms contained in different video frames are different object organisms. Here, for the situation where the object organisms contained in different video frames are the same object organism, an optional implementation provided in this embodiment is to perform biological object detection on each video frame to obtain biological objects and object video frames containing biological objects, including: identifying the biological objects contained in each video frame; if the biological objects contained in each video frame are the same biological object, screening the object video frames containing the biological objects in each video frame.

[0020] Based on this, for the case where the object organisms contained in different video frames are different object organisms, this embodiment provides an optional implementation method, if the biological objects contained in the video frames are different biological objects, the significant index of each biological object is calculated, and the target biological object is determined based on the significant index, and the object video frame containing the target biological object is selected in each video. In the process of determining the target biological object based on the significant index, the biological objects can be sorted in descending order according to the significant index of each biological object, and the biological object ranked first is used as the target biological object.

[0021] Among them, the saliency index of the biological object refers to an index that characterizes the degree of saliency of the biological object in each video frame. The saliency index can be specifically determined according to the number of video frames containing the biological object, or the number of times the biological object appears in each video frame. For example, the number of video frames containing the biological object or the number of times the biological object appears in each video frame is used as the saliency index; in addition, the saliency index can also be determined according to other parameters, such as according to the ratio of the image area of ​​the biological object in the video frame to the frame image area of ​​the video frame.

[0022] In addition, in the process of performing object detection and image extraction on the video to be detected and the like, in an optional implementation manner provided by this embodiment, performing object detection and image extraction on each video frame of the video to be detected and obtaining an object image includes: performing non-biological object detection on each video frame and obtaining a non-biological object and an object video frame containing the non-biological object; extracting an image block of the non-biological object from the object video frame as the object image; Among them, in order to improve the efficiency and detection accuracy of object detection and image extraction of non-biological objects, an object detection model for non-biological objects can also be pre-trained, and non-biological object detection can be performed by inputting each video frame into the object detection model. During the process of non-biological object detection, if the object detection model detects that each video frame contains only one non-biological object, then the non-biological object can be output; if it is detected that each video frame contains multiple non-biological objects, the non-biological objects can be sorted in descending order according to their significant indicators in each video frame, and the non-biological object ranked first can be output.

[0023] It should be pointed out that in order to improve the accuracy and comprehensiveness of machine-generated video detection on each video frame of the video to be detected, object detection of biological objects and non-biological objects can also be performed on each video frame of the video to be detected. For example, biological object detection and non-biological object detection can be performed on each video frame to obtain object video frames of biological objects, non-biological objects, and object objects containing biological objects and non-biological objects, and image blocks of biological objects and non-biological objects can be extracted from the object video frames as object images.

[0024] Step S204: performing differential processing on the object image, and performing feature extraction on the obtained differential image to obtain a differential feature sequence.

[0025] In specific implementation, based on the object images extracted from each video frame of the video to be detected, differential processing is performed on the object images to obtain changes in biological objects or non-biological objects in the video frames between different video frames. Specifically, in the process of differential processing of the object images, differential processing is performed on adjacent object images based on determining the image order of the object images to obtain corresponding differential images.

[0026] In an optional implementation manner provided by this embodiment, performing difference processing on the object image includes: Sort the object images based on the video frames to which the object images belong to obtain an object image sequence; A difference process is performed on adjacent object images in the object image sequence.

[0027] During the specific execution process, in the process of sorting the object images based on the video frames to which the object images belong to obtain an object image sequence, since each video frame in the video to be detected has a playback sequence, the sorting order of the video frames to which the object images belong can be determined based on the playback time or frame order of each video frame in the video to be detected, and the sorting order of the video frames to which the object images belong can be further used as the sorting order of the object images, so that the object images are sorted on the basis of the sorting order of the object images to obtain an object image sequence composed of object images.

[0028] In the process of performing differential processing on adjacent object images in the object image sequence, differential processing may be performed on every two adjacent object images in the object image sequence. The differential processing may be inputting every two adjacent object images into a differential algorithm for differential calculation to obtain a differential image.

[0029] After performing the differential processing on the object image to obtain the object image, feature extraction is performed on the obtained differential image to obtain a differential feature sequence. Specifically, feature extraction can be performed on the obtained differential image to obtain the differential features of each differential image, and a differential feature sequence is obtained based on the differential features of each differential image.

[0030] As described above, object detection is performed on each video frame of the video to be detected, which may be biological object detection on the video frames. In this regard, in an optional implementation provided by the present embodiment, feature extraction is performed on the obtained differential images to obtain a differential feature sequence, including: biological posture feature extraction is performed on each differential image to obtain biological posture features, and the differential feature sequence is constructed based on the obtained biological posture features.

[0031] Among them, in the process of extracting the biological posture features of each differential image to obtain the biological posture features, the biological posture features of each differential image can be obtained by inputting the differential image into a pre-trained biological posture extraction model for biological posture feature extraction; the biological posture extraction model can be a model that adopts a deep neural network or a network architecture such as ViT (Vision Transformer) and is obtained after model training; on the basis of obtaining the biological posture features of each differential image, the biological posture features of each differential image can be feature assembled to obtain a differential feature sequence, and the biological posture features of each differential image can be feature spliced ​​to obtain a differential feature sequence.

[0032] For example, Figure 3The schematic diagram of the machine-generated video detection process shown in the figure, after obtaining the biological object image, the biological object image is input into the differential module of the image differential algorithm for differential processing to obtain the differential image; further, the differential image is input into the biological posture extraction model of the deep neural network for biological posture feature extraction to obtain the biological posture feature, and the biological posture features are assembled to obtain the biological posture feature sequence.

[0033] In addition, object detection can be performed on each video frame of the video to be detected, and non-biological object detection can also be performed on the video frames. In this regard, in an optional implementation provided by this embodiment, feature extraction is performed on the obtained differential images to obtain a differential feature sequence, including: dynamic feature extraction of non-biological objects is performed on each differential image to obtain dynamic features, and the differential feature sequence is constructed based on the obtained dynamic features.

[0034] Similarly, in the process of extracting dynamic features of non-biological objects from each differential image to obtain dynamic features, the dynamic features of each differential image can be obtained by inputting the differential image into a pre-trained dynamic feature extraction model of non-biological objects for dynamic feature extraction; the dynamic feature extraction model can be a model that adopts a deep neural network or a network architecture such as ViT (Vision Transformer) and is obtained after model training; on the basis of obtaining the dynamic features of each differential image, the dynamic features of each differential image can be feature assembled to obtain a differential feature sequence, and the dynamic features of each differential image can be feature spliced ​​to obtain a differential feature sequence.

[0035] Step S206: splicing the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence.

[0036] In this embodiment, in the process of performing machine-generated video detection on each video frame of the video to be detected, starting from the two dimensions of video features of each video frame and the difference of object images contained in each video frame, machine-generated video detection is performed on each video frame of the video to be detected. In the above steps S202 to S204, object detection and image extraction are performed on each video frame of the video to be detected in the dimension of the difference of object images contained in each video frame of the video to be detected to obtain the object image; Here, before the video feature sequence of each video frame is spliced ​​with the differential feature sequence, video feature extraction is performed on each video frame of the video to be detected in the dimension of video features to obtain a video feature sequence. In an optional implementation provided by this embodiment, the video feature sequence is obtained in the following manner: video feature extraction is performed on each video frame to obtain video features, and the video feature sequence is constructed based on the obtained video features.

[0037] During the specific implementation process, each video frame of the video to be detected can be input into a pre-trained video feature extraction model to perform video feature extraction to obtain video features of each video frame; the video feature extraction model can be a model that adopts a deep neural network or a network architecture such as ViT (Vision Transformer) and is obtained after model training; based on the video features of each video frame, the video features of each video frame can be assembled to obtain a video feature sequence, and the video features of each video frame can be spliced ​​to obtain a video feature sequence.

[0038] In a specific implementation, on the basis of obtaining the video feature sequence and the differential feature sequence of each video frame of the video to be detected, the video feature sequence and the differential feature sequence are spliced ​​to obtain a spliced ​​feature sequence. Specifically, in the splicing process, the video feature sequence and the differential feature sequence can be spliced ​​first in a preset splicing order to obtain a spliced ​​feature sequence. For example, if the preset splicing order is that the video feature sequence comes first and the differential feature sequence comes second, then the differential feature sequence is spliced ​​after the video feature sequence to obtain a spliced ​​feature sequence. For another example, if the preset splicing order is that the video feature sequence comes second and the differential feature sequence comes first, then the video feature sequence is spliced ​​after the differential feature sequence to obtain a spliced ​​feature sequence.

[0039] For example, Figure 3 The schematic diagram of the machine-generated video detection process shown in FIG. 1 is to input the sampled video frame of the video stream to be detected as the video to be detected, and the video feature extraction model of the ViT (Vision Transformer) network is used to extract video features to obtain video features, and the video features are assembled to obtain a video feature sequence. Thereafter, the video feature sequence is spliced ​​with the biological posture feature sequence.

[0040] In addition, in order to improve the processing efficiency of subsequently inputting the spliced ​​feature sequence into the video detection module for machine-generated video detection, and to facilitate the video detection module to distinguish the video feature sequence and the differential feature sequence contained in the spliced ​​feature sequence, in the process of splicing the video feature sequence and the differential feature sequence, a splicing identifier or an interval identifier may be added before the video feature sequence and the differential feature sequence, so that the video detection module can quickly distinguish the video feature sequence and the differential feature sequence contained in the spliced ​​feature sequence based on the splicing identifier or the interval identifier.

[0041] Step S208: input the splicing feature sequence into a video detection module to detect the machine-generated video and obtain a detection result.

[0042] As described above, this embodiment starts from the two dimensions of the video features of each video frame and the differences in the object images contained in each video frame, and performs machine-generated video detection on each video frame of the video to be detected. Starting from the spliced ​​feature sequence of the feature sequences containing these two dimensions obtained above, the spliced ​​feature sequence is input into the video detection module to perform machine-generated video detection to obtain the detection result.

[0043] Machine-generated video refers to video generated by a model. The model here can be a model for generating videos obtained by machine learning methods or a model that includes the function of generating video. Machine-generated video can be a video generated based on text (text-generated video), such as a video obtained by inputting text into a model for video generation. It can also be a video generated based on various forms of content such as images, voice, and video. Or, it can also be a video generated based on at least two of various forms of content such as text, images, voice, and video, such as a video generated based on text and images.

[0044] During the specific execution process, the video detection module may be a detection model using a Transformer network, wherein the network parameters of the Transformer network are determined after model training. Specifically, during the detection process of machine-generated videos, the spliced ​​feature sequence is input into the detection model, and the detection model outputs a detection result after detecting the machine-generated video, wherein the detection result may be a label of whether the video to be detected is a machine-generated video, for example, if the video to be detected is a machine-generated video, the label is "yes" or "machine-generated video"; if the video to be detected is not a machine-generated video, the label is "no" or "non-machine-generated video".

[0045] Optionally, performing machine-generated video detection includes: detecting whether the video to be detected is a video generated by a generative model based on the spliced ​​feature sequence. The generative model can be a model for generating videos obtained by machine learning methods, or a large language model including a function of generating video obtained by machine learning methods.

[0046] Specifically, the generative model can be a generative model that generates video based on text, or a generative model that generates video based on various forms of content such as images, voice, video, etc.; or, it can be a generative model that generates video based on at least two of various forms of content such as text, images, voice, video, etc., such as a large language model that generates video based on text and images.

[0047] For example, Figure 3The schematic diagram of the machine-generated video detection process shown in the figure, after splicing the video feature sequence and the biological posture feature sequence, the spliced ​​feature sequence is input into the detection model using the Transformer network, based on the video feature sequence and the biological posture feature sequence contained in the spliced ​​feature sequence, it is detected whether the video to be detected is a video generated by the generative model, and the detection result is output.

[0048] To sum up, the video detection processing method provided by the present embodiment, during the detection process of the video to be detected, on the one hand, performs object detection and image extraction on each video frame of the video to be detected to obtain the object image contained in the video frame, and after differential processing of the object image, performs differential feature extraction on the obtained differential image to obtain a differential feature sequence, and on the other hand, performs feature extraction on the video frame of the video to be detected to obtain a video feature sequence of the video to be detected, on this basis, splices the video feature sequence and the differential feature sequence, and inputs the spliced ​​feature sequence into the video detection module to detect machine-generated videos, and obtains a detection result of whether the video to be detected is a machine-generated video, thereby combining the features of the two dimensions of video features of the video frame and differential features of the objects contained in the video frame to realize machine-generated video detection of the video to be detected, thereby improving the detection accuracy and comprehensiveness.

[0049] The following takes the application of a video detection processing method provided by this embodiment in a video detection scenario as an example, combined with Figure 4 , the video detection processing method provided in this embodiment is further described, see Figure 4 , a video detection processing method applied to a video detection scenario specifically includes the following steps.

[0050] Step S402 : uniformly sample the video frame sequence of the video to be detected to obtain sampled video frames.

[0051] Step S404: perform biological object detection on each sampled video frame to obtain biological objects and object video frames containing biological objects.

[0052] Step S406: extracting an object image of the biological object from the object video frame.

[0053] Step S408 , sorting the object images based on the sampled video frames to which the object images belong to obtain an object image sequence.

[0054] Step S410: performing differential processing on adjacent object images in the object image sequence to obtain a differential image.

[0055] Step S412: extracting biological posture features from each differential image to obtain biological posture features, and constructing a differential feature sequence according to the obtained biological posture features.

[0056] Step S414, splicing the video feature sequence of each sampled video frame with the differential feature sequence to obtain a spliced ​​feature sequence.

[0057] Step S416, inputting the spliced ​​feature sequence into the detection model to detect the machine-generated video and obtain the detection result.

[0058] It should be noted that any one of steps S402 to S416 or any combination of multiple steps can be combined with any one of steps S202 to S208 to form a new implementation method according to the needs of implementation deployment; in addition, according to the needs of actual deployment, any one or multiple technical features can be selected from steps S402 to S416 and combined with any one or multiple technical features provided by steps S202 to S208 to form a new implementation method; or, any one or multiple technical features in steps S402 to S416 can be replaced with any one or multiple technical features provided by steps S202 to S208 to form a new implementation method according to the needs of actual deployment, which will not be repeated here.

[0059] An embodiment of a video detection processing device provided in this specification is as follows: In the above-mentioned embodiment, a video detection processing method is provided, and correspondingly, a video detection processing device is also provided, which will be described below with reference to the accompanying drawings.

[0060] Reference Figure 5 , which shows a schematic diagram of an embodiment of a video detection processing device provided in this embodiment.

[0061] Since the device embodiment corresponds to the method embodiment, the description is relatively simple, and the relevant parts can refer to the corresponding description of the method embodiment provided above. The device embodiment described below is only illustrative.

[0062] This embodiment provides a video detection processing device, the device comprising: The detection and extraction module 502 is configured to perform object detection and image extraction on each video frame of the video to be detected to obtain an object image; The differential processing module 504 is configured to perform differential processing on the object image and perform feature extraction on the obtained differential image to obtain a differential feature sequence; The splicing module 506 is configured to splice the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence; The video detection module 508 is configured to input the splicing feature sequence into the video detection module to perform machine-generated video detection to obtain a detection result.

[0063] An embodiment of a video detection and processing device provided in this specification is as follows: Corresponding to the video detection processing method described above, based on the same technical concept, one or more embodiments of this specification also provide a video detection processing device, which is used to execute the video detection processing method provided above. Figure 6 A schematic diagram of the structure of a video detection and processing device provided in one or more embodiments of this specification.

[0064] This embodiment provides a video detection processing device, including: like Figure 6 As shown, the video detection processing device may have relatively large differences due to different configurations or performances, and may include one or more processors 601 and memory 602, and the memory 602 may store one or more storage applications or data. Among them, the memory 602 may be a short-term storage or a persistent storage. The application stored in the memory 602 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the video detection processing device. Furthermore, the processor 601 may be configured to communicate with the memory 602, and execute a series of computer executable instructions in the memory 602 on the video detection processing device. The video detection processing device may also include one or more power supplies 603, one or more wired or wireless network interfaces 604, one or more input / output interfaces 605, one or more keyboards 606, etc.

[0065] In a specific embodiment, the video detection processing device includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions for the video detection processing device, and the one or more programs are configured to be executed by one or more processors, including computer executable instructions for performing the following: Perform object detection and image extraction on each video frame of the video to be detected to obtain an object image; Performing differential processing on the object image, and performing feature extraction on the obtained differential image to obtain a differential feature sequence; Splicing the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence; The spliced ​​feature sequence is input into a video detection module to detect the machine-generated video to obtain a detection result.

[0066] An embodiment of a computer-readable storage medium provided in this specification is as follows: Corresponding to the video detection processing method described above, based on the same technical concept, one or more embodiments of this specification also provide a computer-readable storage medium.

[0067] The computer-readable storage medium provided in this embodiment is used to store computer-executable instructions, and the computer-executable instructions implement the following process when executed: Perform object detection and image extraction on each video frame of the video to be detected to obtain an object image; Performing differential processing on the object image, and performing feature extraction on the obtained differential image to obtain a differential feature sequence; Splicing the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence; The spliced ​​feature sequence is input into a video detection module to detect the machine-generated video to obtain a detection result.

[0068] It should be noted that an embodiment of a computer-readable storage medium in this specification and an embodiment of a video detection and processing method in this specification are based on the same inventive concept, so the specific implementation of this embodiment can refer to the implementation of the aforementioned corresponding method, and the repeated parts will not be repeated.

[0069] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referenced to each other. Each embodiment focuses on the differences from other embodiments. For example, the device embodiment, equipment embodiment and computer-readable storage medium embodiment are similar to the method embodiment, so the description is relatively simple. To read the relevant contents in the device embodiment, equipment embodiment and computer-readable storage medium embodiment, please refer to the partial description of the method embodiment.

[0070] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0071] In the 1930s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to circuit structures such as diodes, transistors, switches, etc.) or software improvements (improvements to method flows). However, with the development of technology, many of today's method flow improvements can be regarded as direct improvements to hardware circuit structures. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using hardware entity modules. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask chip manufacturers to design and produce dedicated integrated circuit chips. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one kind of HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed ​​Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages ​​and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.

[0072] The controller may be implemented in any suitable manner, for example, the controller may take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (e.g., software or firmware) executable by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller, examples of which include but are not limited to the following microcontrollers: ARC 625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320, and the memory controller may also be implemented as part of the control logic of the memory. It is also known to those skilled in the art that, in addition to implementing the controller in a purely computer-readable program code manner, the controller may be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, such a controller may be considered as a hardware component, and the devices for implementing various functions included therein may also be considered as structures within the hardware component. Or even, the devices for implementing various functions may be considered as both software modules for implementing the method and structures within the hardware component.

[0073] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.

[0074] For the convenience of description, the above devices are described in terms of functions and are divided into various units. Of course, when implementing the embodiments of this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.

[0075] It should be understood by those skilled in the art that one or more embodiments of the present specification may be provided as a method, system or computer program product. Therefore, one or more embodiments of the present specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present specification may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0076] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0077] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0078] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0079] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.

[0080] Memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. Memory is an example of a computer-readable medium.

[0081] Computer-readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer-readable instructions, data structures, program modules or other data. Examples of computer-readable storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include transitory media such as modulated data signals and carrier waves.

[0082] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of further restrictions, the elements defined by the sentence "includes at least one ..." do not exclude the presence of other identical elements in the process, method, commodity or device including the elements.

[0083] One or more embodiments of the present specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of the present specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.

[0084] The above description is only an embodiment of this document and is not intended to limit this document. For those skilled in the art, this document may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this document should be included in the scope of the claims of this document.

Claims

1. A video detection processing method, comprising: Perform object detection and image extraction on each video frame of the video to be detected to obtain an object image; Performing differential processing on the object image, and performing feature extraction on the obtained differential image to obtain a differential feature sequence; Splicing the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence; The spliced ​​feature sequence is input into a video detection module to detect the machine-generated video to obtain a detection result.

2. The video detection processing method according to claim 1, wherein the step of performing object detection and image extraction on each video frame of the video to be detected to obtain the object image comprises: Performing biological object detection on each of the video frames to obtain a biological object and an object video frame containing the biological object; An image block of the biological object is extracted from the object video frame as the object image.

3. The video detection processing method according to claim 1, wherein the performing differential processing on the object image comprises: Sort the object images based on the video frames to which the object images belong to obtain an object image sequence; A difference process is performed on adjacent object images in the object image sequence.

4. The video detection processing method according to claim 1, wherein the step of extracting features from the obtained differential image to obtain a differential feature sequence comprises: Extracting biological posture features from each differential image to obtain biological posture features, and constructing the differential feature sequence according to the obtained biological posture features; or, Dynamic features of non-biological objects are extracted from each differential image to obtain dynamic features, and the differential feature sequence is constructed according to the obtained dynamic features.

5. The video detection processing method according to claim 1, wherein each video frame of the video to be detected is obtained in the following manner: The video frame sequence of the video to be detected is uniformly sampled, and the obtained sampled video frames are used as the video frames.

6. The video detection processing method according to claim 5, before the step of splicing the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence is performed, it also includes: Video features are extracted from each video frame to obtain video features, and the video feature sequence is constructed according to the obtained video features.

7. The video detection processing method according to claim 2, wherein the biological object detection is performed on each video frame to obtain the biological object and the object video frame containing the biological object, comprising: Identifying biological objects contained in each of the video frames; If the biological objects included in the video frames are the same biological object, object video frames including the biological object are screened from the video frames.

8. The video detection processing method according to claim 7, wherein the biological object detection is performed on each of the video frames to obtain the biological object and the object video frame containing the biological object, further comprising: If the biological objects contained in the video frames are different biological objects, a salient index of each biological object is calculated, and a target biological object is determined based on the salient index, and an object video frame containing the target biological object is selected from the videos.

9. The video detection processing method according to claim 1, wherein the detecting of machine-generated video comprises: Based on the spliced ​​feature sequence, it is detected whether the video to be detected is a video generated by a generative model.

10. A video detection processing device, comprising: The detection and extraction module is configured to perform object detection and image extraction on each video frame of the video to be detected to obtain an object image; A differential processing module is configured to perform differential processing on the object image and perform feature extraction on the obtained differential image to obtain a differential feature sequence; A splicing module is configured to splice the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence; The video detection module is configured to input the splicing feature sequence into the video detection module to detect the machine-generated video and obtain a detection result.

11. A video detection and processing device, comprising: processor; and a memory configured to store computer executable instructions that, when executed, cause the processor to: Perform object detection and image extraction on each video frame of the video to be detected to obtain an object image; Performing differential processing on the object image, and performing feature extraction on the obtained differential image to obtain a differential feature sequence; Splicing the video feature sequence of each video frame with the differential feature sequence to obtain a spliced ​​feature sequence; The spliced ​​feature sequence is input into a video detection module to detect the machine-generated video to obtain a detection result.

12. A computer-readable storage medium for storing computer-executable instructions, wherein the computer-executable instructions implement the steps of the method of claim 1 when executed.

Images