Video tampering detection method, device, system, equipment, medium and product

Abstract

The present disclosure provides a video tampering detection method, apparatus, system, device, medium, and product. The video tampering detection method comprises: receiving a video to be detected; automatically detecting the video to be detected using an ORB algorithm that includes feature point extraction and feature point description; and, if the automatic detection result indicates that the video to be detected has been tampered with, sending the video to be detected and the detection result to a backend maintenance subsystem, where system maintenance personnel manually review the video to be detected in the backend maintenance subsystem. The disclosed solution can implement video tampering detection based on a combination of automatic detection using ORB feature point extraction and manual review, while also supporting dynamic adjustment of detection parameters.

Description

Technical Field

[0001] The present invention relates to the technical field of video detection and processing, and in particular to a video tampering detection method, device, system, equipment, medium and product. Background Art

[0002] With the widespread adoption of e-commerce applications and the rapid development of 5G technology, more and more e-commerce product information is being displayed via video. If the displayed videos are tampered with, they can mislead customers and even pose the risk of fraud. Current video tampering detection methods primarily rely on active detection techniques, which insert active forensic information such as digital watermarks and digital signatures during video generation to detect video tampering. However, passive detection, or post-video tampering detection, is relatively understudied in the prior art. Therefore, there is an urgent need to develop a passive method for detecting video tampering. Summary of the Invention

[0003] In view of the above technical problems, the present disclosure provides a video tampering detection method, apparatus, system, equipment, medium and product, which uses a combination of automatic detection based on feature point extraction of ORB (Oriented FAST and Rotated BRIEF) and manual review to achieve video tampering detection, and also supports dynamic adjustment of detection parameters.

[0004] According to a first aspect of the present disclosure, a video tampering detection method is provided, the video tampering detection method comprising:

[0005] Receive the video to be detected;

[0006] Use the ORB algorithm including feature point extraction and feature point description to automatically detect the video to be detected;

[0007] If the automatic detection result shows that the video to be detected has been tampered with, the video to be detected and the detection result are sent to the background maintenance subsystem, and the system maintenance personnel manually review the video to be detected in the background maintenance subsystem.

[0008] According to an embodiment of the present disclosure, the video tampering detection method also includes: when the automatic detection result is that the video to be detected has not been tampered with, the video to be detected and the detection result are sent to the background maintenance subsystem, and the background maintenance subsystem stores and registers the detection result that the video to be detected has not been tampered with.

[0009] According to an embodiment of the present disclosure, the video tampering detection method further includes: receiving detection parameter adjustment information for adjusting detection parameters of the ORB algorithm, and using the ORB algorithm after adjusting the detection parameters to detect the video to be detected.

[0010] According to an embodiment of the present disclosure, the automatic detection of a video to be detected using an ORB algorithm including feature point extraction and feature point description specifically includes:

[0011] Use oFast algorithm to extract feature points;

[0012] Use rBRIEF algorithm to describe feature points;

[0013] The similarity relationship between each frame in the image is determined by comparing the feature point description values, and whether the video to be detected has been tampered with is determined based on the similarity relationship.

[0014] According to an embodiment of the present disclosure, the feature point extraction using the oFast algorithm specifically includes:

[0015] Extract preliminary feature points;

[0016] Extracting optimal feature points from the preliminary feature points;

[0017] Use the non-maximum suppression algorithm (NMS) to eliminate feature points at adjacent positions;

[0018] Realize the scale invariance of feature points and obtain oFast feature points.

[0019] According to an embodiment of the present disclosure, extracting preliminary feature points specifically includes:

[0020] First, a large number of detection points are roughly selected from the video to be detected, and then the preliminary feature points are screened out from the large number of detection points.

[0021] According to an embodiment of the present disclosure, the step of selecting the preliminary feature points from a large number of detection points specifically includes:

[0022] A point p is selected from the large number of roughly selected detection points, and it is determined whether there are a certain number of adjacent pixel points within the range of 3 pixels around the point p whose grayscale values ​​are all greater than point p or all smaller than point p. If there are a certain number of adjacent pixel points whose grayscale values ​​are all greater than point p or all smaller than point p, then point p is a preliminary feature point.

[0023] According to an embodiment of the present disclosure, extracting the optimal feature point from the preliminary feature points specifically includes:

[0024] A decision tree is constructed using the ID3 greedy algorithm of machine learning, and all pixel points within a 3-pixel range around each of the preliminary feature points are input into the decision tree to select the optimal feature point.

[0025] According to an embodiment of the present disclosure, the use of a non-maximum suppression (NMS) algorithm to remove feature points at adjacent positions specifically includes:

[0026] Calculate the response value of each feature point, which is the sum of the absolute values ​​of the deviations of all pixels within the feature point and its surrounding 3 pixels. Then, among the adjacent feature points, the ones with larger response values ​​are retained and the ones with smaller response values ​​are deleted.

[0027] According to an embodiment of the present disclosure, achieving scale invariance of feature points and obtaining oFast feature points specifically includes:

[0028] By using the pyramid algorithm to reduce the original image according to the scale factor, multiple sets of feature point sets of graphics with different scales are obtained. This set is the oFast feature point of the image.

[0029] According to an embodiment of the present disclosure, the feature point description using the rBRIEF algorithm specifically includes:

[0030] Perform noise reduction on the image;

[0031] Take a large neighborhood window of size S*S with the feature point as the center, randomly select two sub-windows of size 5*5 in the large window, compare the pixels in the sub-windows, and assign binary values ​​to them;

[0032] N pairs of sub-windows are randomly selected from the large window, and binary values ​​are assigned to the sub-windows to form a series of binary codes, where the binary codes are the feature point description values ​​of the feature points.

[0033] According to an embodiment of the present disclosure, the performing noise reduction processing on the image specifically includes: performing noise reduction processing on the image using a Gaussian filtering algorithm.

[0034] According to an embodiment of the present disclosure, the binary assignment specifically includes:

[0035]

[0036] Among them, p(x) and p(y) are the pixel values ​​of the two selected sub-windows.

[0037] According to an embodiment of the present disclosure, the feature point description value specifically includes:

[0038]

[0039] Among them, 1≤n≤N.

[0040] According to an embodiment of the present disclosure, determining the similarity relationship between frames in the image by comparing the feature point description values, and determining whether the video to be detected has been tampered with based on the similarity relationship specifically includes:

[0041] After obtaining the feature point description values ​​of each frame image, if there are substantially identical feature point description values ​​between different frames, it can be determined that there is a duplication or copy smear relationship between the different frames, and then it can be determined that the video to be detected is a tampered video.

[0042] According to an embodiment of the present disclosure, receiving a video to be detected specifically includes:

[0043] Receive the video to be detected sent by the background maintenance subsystem, where the video to be detected is uploaded to the background maintenance subsystem by the customer.

[0044] According to an embodiment of the present disclosure, the method further includes:

[0045] When the result of manual review of the video to be detected is that the video to be detected has not been tampered with, the detection result that the video to be detected has not been tampered with is stored and registered in the background maintenance subsystem.

[0046] According to an embodiment of the present disclosure, the method further includes:

[0047] The backend maintenance subsystem provides a query portal, and the customer queries the detection result of the video through the query portal of the backend maintenance subsystem.

[0048] According to an embodiment of the present disclosure, the adjusted parameters include:

[0049] The rejection threshold of the non-maximum suppression algorithm in the oFast algorithm can be adjusted to adjust the result of feature point extraction; or

[0050] The number of sub-windows randomly selected from the large window in the rBRIEF algorithm can be adjusted to adjust the result of feature point description.

[0051] A second aspect of the present disclosure provides a video tampering detection device, the device comprising a video receiving module, an automatic detection module, and a first sending module, wherein:

[0052] The video receiving module is used to receive the video to be detected;

[0053] The automatic detection module is used to automatically detect the video to be detected using the ORB algorithm including feature point extraction and feature point description;

[0054] The first sending module is used to send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result shows that the video to be detected has been tampered with, and the system maintenance personnel manually review the video to be detected in the background maintenance subsystem.

[0055] According to an embodiment of the present disclosure, the device also includes a second sending module, which is used to send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result is that the video to be detected has not been tampered with, and the background maintenance subsystem stores and registers the detection result that the video to be detected has not been tampered with.

[0056] According to an embodiment of the present disclosure, the device further includes a parameter adjustment module, which is configured to receive detection parameter adjustment information for adjusting detection parameters of the ORB algorithm, and use the ORB algorithm after adjusting the detection parameters to detect the video to be detected.

[0057] According to an embodiment of the present disclosure, the automatic detection module includes a feature point extraction module, a feature point description module and a tampering determination module, wherein:

[0058] The feature point extraction module is used to extract feature points using the oFast algorithm;

[0059] The feature point description module is used to describe feature points using the rBRIEF algorithm;

[0060] The tampering determination module is used to determine the similarity relationship between each frame in the image by comparing the feature point description values, and determine whether the video to be detected has been tampered with based on the similarity relationship.

[0061] According to an embodiment of the present disclosure, the feature point extraction using the oFast algorithm specifically includes:

[0062] Extract preliminary feature points;

[0063] Extracting optimal feature points from the preliminary feature points;

[0064] Use the non-maximum suppression algorithm (NMS) to eliminate feature points at adjacent positions;

[0065] Realize the scale invariance of feature points and obtain oFast feature points.

[0066] According to an embodiment of the present disclosure, extracting preliminary feature points specifically includes:

[0067] First, a large number of detection points are roughly selected from the video to be detected, and then the preliminary feature points are screened out from the large number of detection points.

[0068] According to an embodiment of the present disclosure, the step of selecting the preliminary feature points from a large number of detection points specifically includes:

[0069] A point p is selected from the large number of roughly selected detection points, and it is determined whether there are a certain number of adjacent pixel points within the range of 3 pixels around the point p whose grayscale values ​​are all greater than point p or all smaller than point p. If there are a certain number of adjacent pixel points whose grayscale values ​​are all greater than point p or all smaller than point p, then point p is a preliminary feature point.

[0070] According to an embodiment of the present disclosure, extracting the optimal feature point from the preliminary feature points specifically includes:

[0071] A decision tree is constructed using the ID3 greedy algorithm of machine learning, and all pixel points within a 3-pixel range around each of the preliminary feature points are input into the decision tree to select the optimal feature point.

[0072] According to an embodiment of the present disclosure, the use of a non-maximum suppression (NMS) algorithm to remove feature points at adjacent positions specifically includes:

[0073] Calculate the response value of each feature point, which is the sum of the absolute values ​​of the deviations of all pixels within the feature point and its surrounding 3 pixels. Then, among the adjacent feature points, the ones with larger response values ​​are retained and the ones with smaller response values ​​are deleted.

[0074] According to an embodiment of the present disclosure, achieving scale invariance of feature points and obtaining oFast feature points specifically includes:

[0075] By using the pyramid algorithm to reduce the original image according to the scale factor, multiple sets of feature point sets of graphics with different scales are obtained. This set is the oFast feature point of the image.

[0076] According to an embodiment of the present disclosure, the feature point description using the rBRIEF algorithm specifically includes:

[0077] Perform noise reduction on the image;

[0078] Take a large neighborhood window of size S*S with the feature point as the center, randomly select two sub-windows of size 5*5 in the large window, compare the pixels in the sub-windows, and assign binary values ​​to them;

[0079] N pairs of sub-windows are randomly selected from the large window, and binary values ​​are assigned to the sub-windows to form a series of binary codes, where the binary codes are the feature point description values ​​of the feature points.

[0080] According to an embodiment of the present disclosure, the performing noise reduction processing on the image specifically includes: performing noise reduction processing on the image using a Gaussian filtering algorithm.

[0081] According to an embodiment of the present disclosure, the binary assignment specifically includes:

[0082]

[0083] Among them, p(x) and p(y) are the pixel values ​​of the two selected sub-windows.

[0084] According to an embodiment of the present disclosure, the feature point description value specifically includes:

[0085]

[0086] Among them, 1≤n≤N.

[0087] According to an embodiment of the present disclosure, determining the similarity relationship between frames in the image by comparing the feature point description values, and determining whether the video to be detected has been tampered with based on the similarity relationship specifically includes:

[0088] After obtaining the feature point description values ​​of each frame image, if there are substantially identical feature point description values ​​between different frames, it can be determined that there is a duplication or copy smear relationship between the different frames, and then it can be determined that the video to be detected is a tampered video.

[0089] According to an embodiment of the present disclosure, receiving a video to be detected specifically includes:

[0090] Receive the video to be detected sent by the background maintenance subsystem, where the video to be detected is uploaded to the background maintenance subsystem by the customer.

[0091] According to an embodiment of the present disclosure, the adjusted parameters include:

[0092] The rejection threshold of the non-maximum suppression algorithm in the oFast algorithm can be adjusted to adjust the result of feature point extraction; or

[0093] The number of sub-windows randomly selected from the large window in the rBRIEF algorithm can be adjusted to adjust the result of feature point description.

[0094] A third aspect of the present disclosure provides a video tampering detection system, the system comprising a background maintenance subsystem and an automatic detection system, wherein:

[0095] The background maintenance subsystem is used to send the video to be detected to the automatic detection system for detection, receive the video to be detected and the detection result returned by the automatic detection system, and manually review the video to be detected that is detected to be tampered with.

[0096] The automatic detection system is used to receive the video to be detected sent by the background maintenance subsystem, and automatically detect the video to be detected using the ORB algorithm including feature point extraction and feature point description. If the automatic detection result shows that the video to be detected has been tampered with, the video to be detected and the detection result are sent to the background maintenance subsystem.

[0097] According to an embodiment of the present disclosure, the automatic detection system is further configured to send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result shows that the video to be detected has not been tampered with.

[0098] According to an embodiment of the present disclosure, the background maintenance subsystem is further configured to store and register the detection result that the video to be detected has not been tampered with.

[0099] According to an embodiment of the present disclosure, the background maintenance subsystem is further configured to receive videos to be tested uploaded by customers.

[0100] According to an embodiment of the present disclosure, the background maintenance subsystem is further used to store and register the detection results of the video to be detected that has been manually reviewed as not tampered with.

[0101] According to an embodiment of the present disclosure, the background maintenance subsystem is further configured to provide a query portal to receive customer inquiries regarding video detection results.

[0102] According to an embodiment of the present disclosure, the automatic detection system is further configured to receive detection parameter adjustment information for adjusting detection parameters of the ORB algorithm, and use the ORB algorithm after adjusting the detection parameters to detect the video to be detected.

[0103] According to an embodiment of the present disclosure, the background maintenance subsystem is further configured to receive detection parameter adjustment information from a system maintenance person for adjusting detection parameters of the ORB algorithm, and send the detection parameter adjustment information to the automatic detection system.

[0104] The fourth aspect of the present disclosure provides an electronic device, comprising: one or more processors; a memory for storing one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors execute the above-mentioned video tampering detection method.

[0105] The fifth aspect of the present disclosure further provides a computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to execute the above-mentioned video tampering detection method.

[0106] The sixth aspect of the present disclosure further provides a computer program product, including a computer program, which implements the above-mentioned video tampering detection method when executed by a processor.

[0107] Compared to existing technologies, the video tampering detection method, apparatus, system, device, medium, and product provided in this disclosure can achieve video detection based on a combination of automatic detection based on video frame feature point comparison and manual review, ensuring the authenticity of the video. It uses an algorithm that combines oFast and rBRIEF, which is faster than traditional SIFT and SURF algorithms. The rotational invariance of the ORB algorithm makes the algorithm more robust. It also supports dynamically adjusted video detection and allows customers to query detection results at any time, enabling continuous system optimization and providing a good user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0108] The above contents and other objects, features and advantages of the present disclosure will become more apparent through the following description of the embodiments of the present disclosure with reference to the accompanying drawings, in which:

[0109] Figure 1 The following schematically shows a flow chart of a video tampering detection method according to an embodiment of the present disclosure;

[0110] Figure 2 Schematically shows a schematic diagram of displaying feature point description values ​​in a graphical manner according to an embodiment of the present disclosure;

[0111] Figure 3 The following schematically shows a structural block diagram of a video tampering detection device according to an embodiment of the present disclosure;

[0112] Figure 4 The following schematically shows a structural block diagram of a video tampering detection system according to an embodiment of the present disclosure;

[0113] Figure 5 A block diagram of an electronic device suitable for implementing a video tampering detection method according to an embodiment of the present disclosure is schematically shown. DETAILED DESCRIPTION

[0114] Hereinafter, embodiments of the present disclosure will be described with reference to the accompanying drawings. However, it should be understood that these descriptions are merely exemplary and are not intended to limit the scope of the present disclosure. In the detailed description below, for ease of explanation, many specific details are set forth to provide a comprehensive understanding of the embodiments of the present disclosure. However, it is apparent that one or more embodiments may also be implemented without these specific details. In addition, in the following description, descriptions of well-known structures and technologies are omitted to avoid unnecessary confusion of the concepts of the present disclosure.

[0115] The terms used herein are only for describing specific embodiments and are not intended to limit the present disclosure. The terms "comprise," "include," etc. used herein indicate the presence of the features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0116] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art unless otherwise defined. It should be noted that the terms used herein should be interpreted as having a meaning consistent with the context of this specification and should not be interpreted in an idealized or overly rigid manner.

[0117] When expressions such as "at least one of A, B and C, etc." are used, they should generally be interpreted in accordance with the meaning of the expression commonly understood by those skilled in the art (for example, "a system having at least one of A, B and C" should include but is not limited to a system having A alone, B alone, C alone, A and B, A and C, B and C, and / or A, B, C, etc.).

[0118] Compared with the prior art, the present disclosure provides a video tampering detection method, apparatus, system, equipment, medium and product, the video tampering detection method comprising: receiving a video to be detected; automatically detecting the video to be detected using an ORB algorithm including feature point extraction and feature point description; when the automatic detection result shows that the video to be detected has been tampered with, sending the video to be detected and the detection result to a background maintenance subsystem, and having system maintenance personnel manually review the video to be detected in the background maintenance subsystem; when the automatic detection result shows that the video to be detected has not been tampered with, sending the video to be detected and the detection result to a background maintenance subsystem, and having the background maintenance subsystem store and register the detection result that the video to be detected has not been tampered with. The method disclosed herein can also receive detection parameter adjustment information for adjusting the detection parameters of the ORB algorithm, and use the ORB algorithm after adjusting the detection parameters to detect the video to be detected.

[0119] The video tampering detection method, apparatus, system, device, medium, and product provided by this disclosure can achieve video detection by combining automatic detection based on video frame feature point comparison with manual review to ensure video authenticity. It uses an algorithm that combines oFast and rBRIEF, which is faster than traditional SIFT and SURF algorithms. The rotational invariance of the ORB algorithm makes the algorithm more robust. It also supports dynamically adjusted video detection and allows customers to query detection results at any time, enabling continuous system optimization and providing a good user experience.

[0120] The following will be passed Figure 1-Figure 5 A detailed description is given of a blockchain-based confirmation letter transfer system, method, device, equipment, medium and product of an embodiment of the present disclosure.

[0121] Figure 1 The flowchart of a video tampering detection method according to an embodiment of the present disclosure is schematically shown.

[0122] like Figure 1 As shown, this embodiment provides a video tampering detection method, which includes operations S101 to S105, specifically as follows:

[0123] In operation S101 , a video to be detected is received.

[0124] Receiving the video to be tested specifically includes receiving the video to be tested from a backend maintenance subsystem, where the video to be tested is uploaded by a customer to the backend maintenance subsystem. The backend maintenance subsystem may be an e-commerce backend maintenance subsystem, and the video to be tested is a video of a product to be tested, and the e-commerce backend maintenance subsystem receives the video of the product to be tested uploaded by the customer.

[0125] In operation S102 , the video to be detected is automatically detected using an ORB algorithm including feature point extraction and feature point description.

[0126] The ORB algorithm includes two parts: oFast feature point extraction and rBRIEF feature point description. The automatic detection of the video to be detected using the ORB algorithm including feature point extraction and feature point description includes operations S1021-S1023, which are specifically as follows:

[0127] In operation S1021 , feature point extraction is performed using the oFast algorithm.

[0128] The oFast algorithm is used to extract feature points, specifically including:

[0129] (1) Extract preliminary feature points.

[0130] The extracting of preliminary feature points specifically includes: first roughly selecting a large number of detection points from the video to be detected, and then screening the preliminary feature points from the large number of detection points.

[0131] The screening out of the preliminary feature points from a large number of detection points specifically includes: selecting a point p from the roughly selected large number of detection points, and determining whether there are a certain number of adjacent pixel points within a range of three pixels centered on the point p whose grayscale values ​​are all greater than the point p or all smaller than the point p; if there are a certain number of adjacent pixel points whose grayscale values ​​are all greater than the point p or all smaller than the point p, then the point p is a preliminary feature point.

[0132] (2) Extracting the optimal feature points from the preliminary feature points.

[0133] The extracting of the optimal feature point from the preliminary feature points specifically includes: constructing a decision tree using the ID3 greedy algorithm of machine learning, inputting all pixel points within a range of 3 pixels around each point in the preliminary feature point into the decision tree to select the optimal feature point.

[0134] (3) Use the non-maximum suppression algorithm (NMS) to eliminate feature points at adjacent positions.

[0135] The non-maximum suppression algorithm (NMS) is used to eliminate feature points at adjacent positions, specifically including: calculating the response value of each feature point, where the response value is the sum of the absolute values ​​of the deviations of all pixels within the feature point and its surrounding 3 pixels, and then retaining the feature points with larger response values ​​among the adjacent feature points and deleting the feature points with smaller response values.

[0136] (4) Realize the scale invariance of feature points and obtain oFast feature points.

[0137] The method of achieving scale invariance of feature points and obtaining oFast feature points specifically includes: using a pyramid algorithm to reduce the original image according to a scale factor to obtain a plurality of sets of feature point sets of graphics with different scales, and the sets are the oFast feature points of the image.

[0138] In operation S1022 , feature point description is performed using the rBRIEF algorithm.

[0139] The feature point description using the rBRIEF algorithm specifically includes:

[0140] (1) Perform noise reduction on the image.

[0141] The performing noise reduction processing on the image specifically includes: performing noise reduction processing on the image using a Gaussian filtering algorithm.

[0142] (2) Take a large neighborhood window of size S*S with the feature point as the center, randomly select two sub-windows of size 5*5 in the large window, compare the pixels in the sub-windows, and assign binary values ​​to them, where S represents the window length and the value of S is 31.

[0143] The binary assignment method is:

[0144]

[0145] Among them, p(x) and p(y) are the pixel values ​​of the two selected sub-windows.

[0146] (3) Randomly select N pairs of sub-windows in the large window, and perform binary assignment on the sub-windows to form a series of binary codes, where the binary codes are the feature point description values ​​of the feature points.

[0147] The value of N is 256.

[0148] The feature point description values ​​are as follows:

[0149]

[0150] Among them, 1≤n≤N.

[0151] In operation S1023 , similarity relationships between frames in the image are determined by comparing the feature point description values, and whether the video to be detected has been tampered with is determined based on the similarity relationships.

[0152] The similarity relationship between each frame in the image is determined by comparing the feature point description values, and whether the video to be detected has been tampered with is determined based on the similarity relationship. Specifically, after obtaining the feature point description value of each frame of the image, if there are basically the same feature point description values ​​between different frames, it can be determined that there is a copy or copy smear relationship between different frames, and then it is determined that the video to be detected is a tampered video.

[0153] See also Figure 2 , the feature point description value can be displayed graphically, that is, Figure 2 The number of feature points in Figure 2 It can be seen that the number of feature points in the 3rd and 6th frames is exactly the same, so there is a copy relationship between the two frames; and the number of feature points in the 10th and 13th frames is highly similar, so there is a copy and smear relationship between the frames, and it is determined that the video is a tampered video.

[0154] In operation S103, when the automatic detection result shows that the video to be detected has been tampered with, the video to be detected and the detection result are sent to the background maintenance subsystem, and the system maintenance personnel manually review the video to be detected in the background maintenance subsystem.

[0155] Furthermore, when the result of manual review of the video to be detected is that the video to be detected has not been tampered with, the detection result that the video to be detected has not been tampered with is stored and registered in the background maintenance subsystem.

[0156] In operation SI04, when the automatic detection result is that the video to be detected has not been tampered with, the video to be detected and the detection result are sent to the background maintenance subsystem, and the background maintenance subsystem stores and registers the detection result that the video to be detected has not been tampered with.

[0157] Furthermore, the background maintenance subsystem provides a query portal, and customers can query the detection results of the video through the query portal of the background maintenance subsystem.

[0158] In operation S105 , detection parameter adjustment information for adjusting detection parameters of the ORB algorithm is received, and the ORB algorithm after the detection parameters are adjusted is used to continue detecting the video to be detected.

[0159] System maintenance personnel can adjust the detection parameters of the ORB algorithm by sending parameter adjustment information to adjust the detection parameters of the ORB algorithm, and then use the ORB algorithm with adjusted parameters to perform video detection, thereby further optimizing video anti-tampering detection.

[0160] The adjusted parameters include:

[0161] 1) The rejection threshold of the non-maximum suppression algorithm in the oFast algorithm can be adjusted to adjust the result of feature point extraction.

[0162] 2) The number of sub-windows randomly selected from the large window in the rBRIEF algorithm can be adjusted to adjust the result of feature point description.

[0163] By adjusting these parameters, the system can continuously improve detection robustness. For example, if system maintenance personnel discover that the detection results have large errors, they can modify the number of random windows through this adjustment to see if subsequent detection results improve, gradually improving detection effectiveness.

[0164] The video tampering detection method provided by this disclosure can achieve video detection by combining automatic detection based on video frame feature point comparison with manual review, ensuring video authenticity. It uses a combination of oFast and rBRIEF algorithms, which is faster than traditional SIFT and SURF algorithms. The rotational invariance of the ORB algorithm makes the algorithm more robust. It also supports dynamically adjusted video detection and allows customers to query detection results at any time, enabling continuous system optimization and providing a good user experience.

[0165] Based on the above Figure 1 、 Figure 2 The present disclosure also provides a video tampering detection device. Figure 3 The device is described in detail.

[0166] Figure 3 The following schematically shows a structural block diagram of a video tampering detection device according to an embodiment of the present disclosure.

[0167] like Figure 3 As shown, this embodiment provides a video tampering detection device 300 , which includes a video receiving module 301 , an automatic detection module 302 , a first sending module 303 , a second sending module 304 and a parameter adjustment module 305 .

[0168] The video receiving module 301 is used to receive the video to be detected.

[0169] Receiving the video to be tested specifically includes receiving the video to be tested from a backend maintenance subsystem, where the video to be tested is uploaded by a customer to the backend maintenance subsystem. The backend maintenance subsystem may be an e-commerce backend maintenance subsystem, and the video to be tested is a video of a product to be tested, and the e-commerce backend maintenance subsystem receives the video of the product to be tested uploaded by the customer.

[0170] The automatic detection module 302 is configured to automatically detect the video to be detected using an ORB algorithm including feature point extraction and feature point description.

[0171] The ORB algorithm includes two parts: oFast feature point extraction and rBRIEF feature point description. The automatic detection module 302 includes a feature point extraction module 3021 , a feature point description module 3022 and a tampering determination module 3023 .

[0172] The feature point extraction module 3021 is used to extract feature points using the oFast algorithm.

[0173] The oFast algorithm is used to extract feature points, specifically including:

[0174] (1) Extract preliminary feature points.

[0175] The extracting of preliminary feature points specifically includes: first roughly selecting a large number of detection points from the video to be detected, and then screening the preliminary feature points from the large number of detection points.

[0176] The screening out of the preliminary feature points from a large number of detection points specifically includes: selecting a point p from the roughly selected large number of detection points, and determining whether there are a certain number of adjacent pixel points within a range of three pixels centered on the point p whose grayscale values ​​are all greater than the point p or all smaller than the point p; if there are a certain number of adjacent pixel points whose grayscale values ​​are all greater than the point p or all smaller than the point p, then the point p is a preliminary feature point.

[0177] (2) Extracting the optimal feature points from the preliminary feature points.

[0178] The extracting of the optimal feature point from the preliminary feature points specifically includes: constructing a decision tree using the ID3 greedy algorithm of machine learning, inputting all pixel points within a range of 3 pixels around each point in the preliminary feature point into the decision tree to select the optimal feature point.

[0179] (3) Use the non-maximum suppression algorithm (NMS) to eliminate feature points at adjacent positions.

[0180] The non-maximum suppression algorithm (NMS) is used to eliminate feature points at adjacent positions, specifically including: calculating the response value of each feature point, where the response value is the sum of the absolute values ​​of the deviations of all pixels within the feature point and its surrounding 3 pixels, and then retaining the feature points with larger response values ​​among the adjacent feature points and deleting the feature points with smaller response values.

[0181] (4) Realize the scale invariance of feature points and obtain oFast feature points.

[0182] The method of achieving scale invariance of feature points and obtaining oFast feature points specifically includes: using a pyramid algorithm to reduce the original image according to a scale factor to obtain a plurality of sets of feature point sets of graphics with different scales, and the sets are the oFast feature points of the image.

[0183] The feature point description module 3022 is used to describe feature points using the rBRIEF algorithm.

[0184] The feature point description using the rBRIEF algorithm specifically includes:

[0185] (1) Perform noise reduction on the image.

[0186] The performing noise reduction processing on the image specifically includes: performing noise reduction processing on the image using a Gaussian filtering algorithm.

[0187] (2) Take a large neighborhood window of size S*S with the feature point as the center, randomly select two sub-windows of size 5*5 in the large window, compare the pixels in the sub-windows, and assign binary values ​​to them, where S represents the window length and the value of S is 31.

[0188] The binary assignment method is:

[0189]

[0190] Among them, p(x) and p(y) are the pixel values ​​of the two selected sub-windows.

[0191] (3) Randomly select N pairs of sub-windows in the large window, and perform binary assignment on the sub-windows to form a series of binary codes, where the binary codes are the feature point description values ​​of the feature points.

[0192] The value of N is 256.

[0193] The feature point description values ​​are as follows:

[0194]

[0195] Among them, 1≤n≤N.

[0196] The tampering determination module 3023 is configured to determine a similarity relationship between frames in an image by comparing the feature point description values, and determine whether the video to be detected has been tampered with based on the similarity relationship.

[0197] The similarity relationship between each frame in the image is determined by comparing the feature point description values, and whether the video to be detected has been tampered with is determined based on the similarity relationship. Specifically, after obtaining the feature point description value of each frame of the image, if there are basically the same feature point description values ​​between different frames, it can be determined that there is a copy or copy smear relationship between different frames, and then it is determined that the video to be detected is a tampered video.

[0198] See also Figure 2 , the feature point description value can be displayed graphically, that is, Figure 2 The number of feature points in Figure 2 It can be seen that the number of feature points in the 3rd and 6th frames is exactly the same, so there is a copy relationship between the two frames; and the number of feature points in the 10th and 13th frames is highly similar, so there is a copy and smear relationship between the frames, and it is determined that the video is a tampered video.

[0199] The first sending module 303 is used to send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result shows that the video to be detected has been tampered with, and the system maintenance personnel manually review the video to be detected in the background maintenance subsystem.

[0200] Furthermore, when the result of manual review of the video to be detected is that the video to be detected has not been tampered with, the detection result that the video to be detected has not been tampered with is stored and registered in the background maintenance subsystem.

[0201] The second sending module 304 is used to send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result shows that the video to be detected has not been tampered with, and the background maintenance subsystem stores and registers the detection result that the video to be detected has not been tampered with.

[0202] Furthermore, the background maintenance subsystem provides a query portal, and customers can query the detection results of the video through the query portal of the background maintenance subsystem.

[0203] The parameter adjustment module 305 is configured to receive parameter adjustment information for adjusting parameters of the ORB algorithm, and continue to detect the video to be detected using the ORB algorithm after adjusting the parameters.

[0204] System maintenance personnel can adjust the detection parameters of the ORB algorithm by sending parameter adjustment information to adjust the detection parameters of the ORB algorithm, and then use the ORB algorithm with adjusted parameters to perform video detection, thereby further optimizing video anti-tampering detection.

[0205] The adjusted parameters include:

[0206] 1) The rejection threshold of the non-maximum suppression algorithm in the oFast algorithm can be adjusted to adjust the result of feature point extraction.

[0207] 2) The number of sub-windows randomly selected from the large window in the rBRIEF algorithm can be adjusted to adjust the result of feature point description.

[0208] By adjusting these parameters, the system can continuously improve detection robustness. For example, if system maintenance personnel discover that the detection results have large errors, they can modify the number of random windows through this adjustment to see if subsequent detection results improve, gradually improving detection effectiveness.

[0209] The video tampering detection device provided by this disclosure can implement video detection by combining automatic detection based on video frame feature point comparison with manual review to ensure video authenticity. It uses a combination of oFast and rBRIEF algorithms, which is faster than traditional SIFT and SURF algorithms. The rotational invariance of the ORB algorithm makes the algorithm more robust. It also supports dynamically adjusted video detection and allows customers to query detection results at any time, enabling continuous system optimization and providing a good user experience.

[0210] Based on the above Figure 1 、 Figure 2 The present disclosure also provides a video tampering detection system. Figure 4 The system is described in detail.

[0211] Figure 4 The following schematically shows a block diagram of a video tampering detection system according to an embodiment of the present disclosure.

[0212] like Figure 4 As shown, this embodiment provides a video tampering detection system 400, which includes a background maintenance subsystem 401 and an automatic detection system 402, wherein:

[0213] The background maintenance subsystem 401 is used to receive the video to be detected uploaded by the customer 403, send the video to be detected to the automatic detection system 402 for detection, receive the video to be detected and the detection result returned by the automatic detection system 402, manually review the video to be detected that is detected to be tampered with, or store and register the detection result that the video to be detected has not been tampered with.

[0214] The automatic detection system 402 is used to receive the video to be detected sent by the background maintenance subsystem 401, and use the ORB algorithm including feature point extraction and feature point description to automatically detect the video to be detected. When the automatic detection result is that the video to be detected has been tampered with or not tampered with, the video to be detected and the detection result are sent to the background maintenance subsystem.

[0215] Furthermore, the backend maintenance subsystem 401 is further configured to receive a video to be tested uploaded by a client 403. The video to be tested is uploaded to the backend maintenance subsystem by the client 403. The backend maintenance subsystem may be an e-commerce backend maintenance subsystem, and the video to be tested is a video of a product to be tested, and the e-commerce backend maintenance subsystem receives the video of the product to be tested uploaded by the client 403.

[0216] Furthermore, the background maintenance subsystem 401 is also used to store and register the detection results of the videos to be detected that are manually reviewed as not tampered with.

[0217] Furthermore, the background maintenance subsystem 401 is also used to provide a query entry to receive customer queries on video detection results.

[0218] Furthermore, the automatic detection system 402 is further configured to receive detection parameter adjustment information for adjusting detection parameters of the ORB algorithm, and use the ORB algorithm with the adjusted detection parameters to detect the video to be detected.

[0219] Furthermore, the background maintenance subsystem 401 is further configured to receive detection parameter adjustment information from the system maintenance personnel 404 for adjusting the detection parameters of the ORB algorithm, and send the detection parameter adjustment information to the automatic detection system 402 .

[0220] The functions performed by the automatic detection system 402 are the same as those described above. Figure 1 The functions of the corresponding embodiment of the video tampering detection method are the same, and its content can be fully introduced into the embodiment of this video tampering detection system, and will not be repeated here.

[0221] Figure 5 A block diagram of an electronic device suitable for implementing a video tampering detection method according to an embodiment of the present disclosure is schematically shown.

[0222] like Figure 5 As shown, the electronic device 500 according to an embodiment of the present disclosure includes a processor 501, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 502 or a program loaded from a storage part 508 into a random access memory (RAM) 503. The processor 501 may include, for example, a general-purpose microprocessor (e.g., a CPU), an instruction set processor and / or a related chipset and / or a special-purpose microprocessor (e.g., an application-specific integrated circuit (ASIC)), etc. The processor 501 may also include an onboard memory for caching purposes. The processor 501 may include a single processing unit or multiple processing units for performing different actions of the method flow according to an embodiment of the present disclosure.

[0223] Various programs and data required for the operation of the electronic device 500 are stored in the RAM 503. The processor 501, ROM 502, and RAM 503 are connected to each other via a bus 504. The processor 501 executes the various operations of the method flow according to the embodiment of the present disclosure by executing the programs in the ROM 502 and / or RAM 503. It should be noted that the programs may also be stored in one or more memories other than the ROM 502 and RAM 503. The processor 501 may also execute the various operations of the method flow according to the embodiment of the present disclosure by executing the programs stored in the one or more memories.

[0224] According to an embodiment of the present disclosure, the electronic device 500 may further include an input / output (I / O) interface 505, which is also connected to the bus 504. The electronic device 500 may further include one or more of the following components connected to the I / O interface 505: an input portion 506 including a keyboard, a mouse, etc.; an output portion 507 including devices such as a cathode ray tube (CRT), a liquid crystal display (LCD), and speakers; a storage portion 508 including a hard disk; and a communication portion 509 including a network interface card such as a LAN card or a modem. The communication portion 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to the I / O interface 505 as needed. A removable medium 511, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed in the drive 510 as needed, so that a computer program read therefrom can be installed into the storage portion 508 as needed.

[0225] The present disclosure also provides a computer-readable storage medium, which may be included in the device / apparatus / system described in the above embodiments, or may exist independently and not be incorporated into the device / apparatus / system. The computer-readable storage medium carries one or more programs, and when executed, implements the method according to the embodiments of the present disclosure.

[0226] According to an embodiment of the present disclosure, a computer-readable storage medium may be a non-volatile computer-readable storage medium, for example, it may include but is not limited to: 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 portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In the present disclosure, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus, or device. For example, according to an embodiment of the present disclosure, a computer-readable storage medium may include the ROM 502 and / or RAM 503 described above and / or one or more memories other than ROM 502 and RAM 503.

[0227] The embodiments of the present disclosure also include a computer program product, which includes a computer program containing program code for executing the method shown in the flowchart. When the computer program product is executed in a computer system, the program code is used to cause the computer system to implement the item recommendation method provided by the embodiments of the present disclosure.

[0228] The computer program executes the above functions defined in the system / device of the embodiment of the present disclosure when the computer program is executed by the processor 501. According to the embodiment of the present disclosure, the system, device, module, unit, etc. described above can be implemented by a computer program module.

[0229] In one embodiment, the computer program may be stored on a tangible storage medium such as an optical storage device or a magnetic storage device. In another embodiment, the computer program may be transmitted and distributed in the form of a signal on a network medium, downloaded and installed via the communication portion 509, and / or installed from a removable medium 511. The program code contained in the computer program may be transmitted using any appropriate network medium, including but not limited to wireless, wired, or any suitable combination thereof.

[0230] In such an embodiment, the computer program can be downloaded and installed from a network via the communication section 509, and / or installed from a removable medium 511. When the computer program is executed by the processor 501, the above-described functions defined in the system of the embodiment of the present disclosure are performed. According to the embodiment of the present disclosure, the systems, devices, means, modules, units, etc. described above can be implemented by computer program modules.

[0231] According to an embodiment of the present disclosure, the program code for executing the computer program provided by the embodiment of the present disclosure can be written in any combination of one or more programming languages. Specifically, these computer programs can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. Programming languages ​​include, but are not limited to, languages ​​such as Java, C++, python, "C" or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, partially on a remote computing device, or entirely on a remote computing device or server. In cases involving a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, using an Internet service provider to connect via the Internet).

[0232] The flowcharts and block diagrams in the accompanying drawings illustrate the possible implementation architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each box in the flowchart or block diagram can represent a module, program segment, or a part of code, and the above-mentioned module, program segment, or a part of code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession 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 or flowchart, and the combination of boxes in the block diagram or flowchart, can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0233] Those skilled in the art will appreciate that the features described in the various embodiments and / or claims of this disclosure may be combined and / or coupled in various ways, even if such combinations and / or couplings are not explicitly described in this disclosure. In particular, the features described in the various embodiments and / or claims of this disclosure may be combined and / or coupled in various ways without departing from the spirit and teachings of this disclosure. All such combinations and / or couplings are intended to fall within the scope of this disclosure.

[0234] The embodiments of the present disclosure are described above. However, these embodiments are for illustrative purposes only and are not intended to limit the scope of the present disclosure. Although each embodiment has been described separately above, this does not mean that the measures in each embodiment cannot be used in combination to advantage. The scope of the present disclosure is defined by the appended claims and their equivalents. Without departing from the scope of the present disclosure, those skilled in the art may make various substitutions and modifications, which should all fall within the scope of the present disclosure.

Claims

1. A video tampering detection method, characterized in that: The video tampering detection method comprises: Receive the video to be detected; Automatically detect the video to be detected using the ORB algorithm including feature point extraction and feature point description, wherein in the ORB algorithm, the oFast algorithm is used for feature point extraction and the rBRIEF algorithm is used for feature point description; If the automatic detection result shows that the video to be detected has been tampered with, the video to be detected and the detection result are sent to the background maintenance subsystem, and the system maintenance personnel manually review the video to be detected in the background maintenance subsystem to determine whether the video to be detected in the background maintenance subsystem has been tampered with. The video tampering detection method further includes: receiving detection parameter adjustment information for adjusting detection parameters of the ORB algorithm, detecting the video to be detected using the ORB algorithm after adjusting the detection parameters, and modifying the detection parameters based on the detection result of the video to be detected, The adjusted parameters include: The rejection threshold of the non-maximum suppression (NMS) algorithm in the oFast algorithm can be adjusted to adjust the result of feature point extraction; or The number of sub-windows randomly selected in the large window in the rBRIEF algorithm can be adjusted to adjust the result of feature point description. The automatic detection of the video to be detected using the ORB algorithm including feature point extraction and feature point description specifically includes: The similarity relationship between each frame in the image is determined by comparing the feature point description values, and whether the video to be detected has been tampered with is determined based on the similarity relationship.

2. The video tampering detection method according to claim 1, characterized in that: The video tampering detection method further includes: When the automatic detection result shows that the video to be detected has not been tampered with, the video to be detected and the detection result are sent to the background maintenance subsystem, and the background maintenance subsystem stores and registers the detection result that the video to be detected has not been tampered with.

3. The video tampering detection method according to claim 1, wherein: The oFast algorithm is used to extract feature points, specifically including: Extract preliminary feature points; Extracting optimal feature points from the preliminary feature points; Use non-maximum suppression algorithm to remove feature points at adjacent positions; Realize the scale invariance of feature points and obtain oFast feature points.

4. The video tampering detection method according to claim 3, characterized in that: The extracting of preliminary feature points specifically includes: First, a large number of detection points are roughly selected from the video to be detected, and then the preliminary feature points are screened out from the large number of detection points.

5. The video tampering detection method according to claim 4, characterized in that: The step of selecting the preliminary feature points from a large number of detection points specifically includes: Select a point p from the roughly selected large number of detection points; Determine whether there are a certain number of adjacent pixels within a range of three pixels centered at point p whose grayscale values ​​are all greater than point p or all less than point p; When the grayscale values ​​of a certain number of adjacent pixel points are all greater than or all less than point p, point p is determined to be a preliminary feature point.

6. The video tampering detection method according to claim 3, characterized in that: The extracting the optimal feature point from the preliminary feature points specifically includes: A decision tree is constructed using the ID3 greedy algorithm of machine learning, and all pixel points within a range of 3 pixels around each of the preliminary feature points are input into the decision tree, thereby selecting the optimal feature point.

7. The video tampering detection method according to claim 3, characterized in that: The method of using the non-maximum suppression algorithm to eliminate feature points at adjacent positions specifically includes: Calculate the response value of each feature point, where the response value is the sum of the absolute values ​​of the deviations of all pixels within the feature point and its surrounding 3 pixels; Among the adjacent feature points, the ones with larger response values ​​are retained and the ones with smaller response values ​​are deleted.

8. The video tampering detection method according to claim 3, wherein: The method of achieving scale invariance of feature points and obtaining oFast feature points specifically includes: By using the pyramid algorithm to reduce the original image according to the scale factor, multiple sets of feature point sets of graphics with different scales are obtained. This set is the oFast feature point of the image.

9. The video tampering detection method according to claim 1, wherein: The feature point description using the rBRIEF algorithm specifically includes: Perform noise reduction on the image; Take a large neighborhood window of size S*S with the feature point as the center, randomly select two sub-windows of size 5*5 in the large window, compare the pixels in the sub-windows, and assign binary values ​​to them; N pairs of sub-windows are randomly selected from the large window, and binary values ​​are assigned to the sub-windows to form a series of binary codes, where the binary codes are the feature point description values ​​of the feature points.

10. The video tampering detection method according to claim 9, characterized in that: The performing noise reduction processing on the image specifically includes: performing noise reduction processing on the image using a Gaussian filtering algorithm.

11. The video tampering detection method according to claim 1, wherein: Determining the similarity relationship between frames in the image by comparing the feature point description values, and determining whether the video to be detected has been tampered with according to the similarity relationship specifically includes: After obtaining the feature point description values ​​of each frame image, if substantially the same feature point description values ​​exist between different frames, it is determined that there is a duplication or copy smear relationship between the different frames, and furthermore, it is determined that the video to be detected is a tampered video.

12. The video tampering detection method according to claim 1, wherein: The receiving of the video to be detected specifically includes: Receive the video to be detected sent by the background maintenance subsystem, where the video to be detected is uploaded to the background maintenance subsystem by the customer.

13. The video tampering detection method according to claim 1, wherein: The method further comprises: When the result of manual review of the video to be detected is that the video to be detected has not been tampered with, the detection result that the video to be detected has not been tampered with is stored and registered in the background maintenance subsystem.

14. The video tampering detection method according to claim 1 or 2, characterized in that: The method further comprises: The backend maintenance subsystem provides a query portal, and the customer queries the detection result of the video through the query portal of the backend maintenance subsystem.

15. A video tampering detection device, characterized in that: The device includes a video receiving module, an automatic detection module and a first sending module, wherein: The video receiving module is used to receive the video to be detected; The automatic detection module is used to automatically detect the video to be detected using the ORB algorithm including feature point extraction and feature point description. The automatic detection module includes a feature point extraction module and a feature point description module, wherein the feature point extraction module is used to extract feature points using the oFast algorithm; the feature point description module is used to describe feature points using the rBRIEF algorithm; The first sending module is used to send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result shows that the video to be detected has been tampered with, and the system maintenance personnel manually review the video to be detected in the background maintenance subsystem to determine whether the video to be detected in the background maintenance subsystem has been tampered with. The device further includes a parameter adjustment module, the parameter adjustment module being configured to receive detection parameter adjustment information for adjusting detection parameters of the ORB algorithm, detect the video to be detected using the ORB algorithm after adjusting the detection parameters, and modify the detection parameters based on the detection result of the video to be detected. The adjusted parameters include: The rejection threshold of the non-maximum suppression algorithm in the oFast algorithm can be adjusted to adjust the result of feature point extraction; or The number of sub-windows randomly selected in the large window in the rBRIEF algorithm can be adjusted to adjust the result of feature point description. Wherein, the automatic detection module further includes: The tampering determination module is used to determine the similarity relationship between each frame in the image by comparing the feature point description values, and determine whether the video to be detected has been tampered with based on the similarity relationship.

16. The video tampering detection device according to claim 15, characterized in that: The device also includes a second sending module, which is used to send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result is that the video to be detected has not been tampered with, and the background maintenance subsystem stores and registers the detection result that the video to be detected has not been tampered with.

17. A video tampering detection system, characterized in that: The system includes a background maintenance subsystem and an automatic detection system, wherein: The backend maintenance subsystem is configured to send the video to be detected to the automatic detection system for detection, receive the video to be detected and the detection result returned by the automatic detection system, and manually review the video to be detected that is detected as tampered with to determine whether the video to be detected has been tampered with; The automatic detection system is used to receive the video to be detected sent by the background maintenance subsystem, automatically detect the video to be detected using the ORB algorithm including feature point extraction and feature point description, and send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result shows that the video to be detected has been tampered with. Among them, in the ORB algorithm, the oFast algorithm is used to extract feature points, and the rBRIEF algorithm is used to describe feature points. The background maintenance subsystem is further configured to receive detection parameter adjustment information for adjusting the detection parameters of the ORB algorithm, and send the detection parameter adjustment information to the automatic detection system. The automatic detection system uses the ORB algorithm after adjusting the detection parameters to detect the video to be detected. Based on the detection result of the video to be detected, the background maintenance subsystem receives the information for modifying the detection parameters. The adjusted parameters include: The rejection threshold of the non-maximum suppression algorithm in the oFast algorithm can be adjusted to adjust the result of feature point extraction; or The number of sub-windows randomly selected in the large window in the rBRIEF algorithm can be adjusted to adjust the result of feature point description. Wherein, the automatic detection system is also used for: The similarity relationship between each frame in the image is determined by comparing the feature point description values, and whether the video to be detected has been tampered with is determined based on the similarity relationship.

18. The video tampering detection system according to claim 17, wherein: The automatic detection system is further configured to send the video to be detected and the detection result to the background maintenance subsystem when the automatic detection result indicates that the video to be detected has not been tampered with.

19. The video tampering detection system according to claim 18, wherein: The background maintenance subsystem is further used to store and register the detection result that the video to be detected has not been tampered with.

20. The video tampering detection system according to any one of claims 17 to 19, characterized in that: The background maintenance subsystem is also used to receive videos to be tested uploaded by customers.

21. The video tampering detection system according to any one of claims 17 to 19, characterized in that: The background maintenance subsystem is further used to store and register the detection results of the videos to be detected that have been manually reviewed as not tampered with.

22. The video tampering detection system according to claim 21, wherein: The background maintenance subsystem is also used to provide a query entry to receive customer queries on video detection results.

23. An electronic device comprising: one or more processors; a storage device for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors are enabled to perform the method according to any one of claims 1 to 14.

24. A computer-readable storage medium having executable instructions stored thereon, which, when executed by a processor, causes the processor to perform the method according to any one of claims 1 to 14.

25. A computer program product comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements the method according to any one of claims 1 to 14.

Images