Information verification method, device, electronic device and computer-readable medium

Abstract

Embodiments of the present disclosure disclose an information verification method, apparatus, electronic device, and computer-readable medium. A specific implementation of the method includes: obtaining video data captured by a camera within a preset time period as initial video data, where the initial video data includes: video data of the same environment at different times, and video data of the same time in different environments; generating an annotated video data set according to the initial video data; generating an object behavior detection result set according to the annotated video data set; inputting a first video feature set and a second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result; verifying the object motion trajectory information corresponding to the object feature prediction result to obtain verified trajectory information for the system to determine the object category. This implementation improves the security of the system and the response speed of the system.

Description

Technical Field

[0001] Embodiments of the present disclosure relate to the field of computer technologies, and particularly to information verification methods, apparatuses, electronic devices, and computer-readable media. Background Art

[0002] With the rapid development of information technologies, it has become increasingly common to perform behavior detection on moving objects in video data at the same time in different environments and at different times in the same environment. Conducting behavior detection on video data and verifying the motion trajectories of the detected motion behaviors can improve the security and response speed of the system. Information verification is a technology for verifying information. Currently, the commonly adopted method for information verification is to manually analyze the trajectory information of moving objects to achieve information verification.

[0003] However, when adopting the above method, the following technical problems often exist:

[0004] First, manually analyzing the trajectory information of moving objects will cause a large error in information verification and cannot correct the generated error in a timely manner, resulting in a slow response speed of the system and low security of the system.

[0005] Second, when annotating object information in video data, misrecognition may occur, resulting in a high error rate of the annotated video data and poor quality of the annotated video data. Due to the inability to determine the correlation between objects, the accuracy of annotation is low, resulting in poor quality of the annotated video data.

[0006] The above information disclosed in this background art section is only used to enhance the understanding of the background of the inventive concept, and thus, it may include information that does not form the prior art known to those of ordinary skill in the art in this country. Summary of the Invention

[0007] The content part of the present disclosure is used to introduce the concepts in a brief form, and these concepts will be described in detail in the subsequent detailed implementation part. The content part of the present disclosure is not intended to identify the key features or essential features of the claimed technical solution, nor is it intended to be used to limit the scope of the claimed technical solution.

[0008] Some embodiments of the present disclosure propose information verification methods, apparatuses, electronic devices, and computer-readable media to solve one or more of the technical problems mentioned in the above background art section.

[0009] In a first aspect, some embodiments of the present disclosure provide an information verification method, which includes: obtaining video data captured by a camera within a preset time period as initial video data, where the initial video data includes: video data of the same environment at different times, and video data of the same time in different environments; generating an annotated video data set according to the initial video data; generating an object behavior detection result set according to the annotated video data set; in response to determining that there is an object behavior detection result representing a normal object behavior detection result in the object behavior detection result set, determining the video feature set corresponding to at least one object behavior detection result representing the normal object behavior detection result as the first video feature set; in response to determining that there is an object behavior detection result representing an abnormal object behavior detection result in the object behavior detection result set, determining the video feature set corresponding to at least one object behavior detection result representing the abnormal object behavior detection result as the second video feature set; inputting the first video feature set and the second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result; verifying the object motion trajectory information corresponding to the object feature prediction result to obtain verified trajectory information for the system to determine the object category.

[0010] In a second aspect, some embodiments of the present disclosure provide an information verification device, which includes: an acquisition unit configured to obtain video data captured by a camera within a preset time period as initial video data, where the initial video data includes: video data of the same environment at different times, and video data of the same time in different environments; a first generation unit configured to generate an annotated video data set according to the initial video data; a second generation unit configured to generate an object behavior detection result set according to the annotated video data set; a first determination unit configured to, in response to determining that there is an object behavior detection result representing a normal object behavior detection result in the object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing the normal object behavior detection result as the first video feature set; a second determination unit configured to, in response to determining that there is an object behavior detection result representing an abnormal object behavior detection result in the object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing the abnormal object behavior detection result as the second video feature set; an input unit configured to input the first video feature set and the second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result; a verification unit configured to verify the object motion trajectory information corresponding to the object feature prediction result to obtain verified trajectory information for the system to determine the object category.

[0011] In a third aspect, some embodiments of the present disclosure provide an electronic device, including: one or more processors; a storage device storing one or more programs thereon, which, when executed by the one or more processors, cause the one or more processors to implement the method described in any implementation manner of the first aspect above.

[0012] In a fourth aspect, some embodiments of the present disclosure provide a computer-readable medium storing a computer program thereon, wherein the program, when executed by a processor, implements the method described in any implementation manner of the first aspect above.

[0013] The above-mentioned various embodiments of the present disclosure have the following beneficial effects: Through the information verification method of some embodiments of the present disclosure, the security of the system is improved, and the response speed of the system is increased. Specifically, the reasons for the slow response speed of the system and the low security of the system are as follows: Manually analyzing the trajectory information of moving objects will cause large errors in information verification, and the generated errors cannot be corrected in time, resulting in a slow response speed of the system and a low security of the system. Based on this, in the information verification method of some embodiments of the present disclosure, first, video data captured by a camera within a preset time period is obtained as initial video data, where the above-mentioned initial video data includes: video data of the same environment at different times, and video data of different environments at the same time. Thus, it can facilitate subsequent processing. Then, according to the above-mentioned initial video data, a labeled video data set is generated. Thus, the above-mentioned initial video data can be labeled, which is convenient for identifying the initial video data and is also beneficial for subsequent automatic analysis of the trajectory information of moving objects. After that, according to the above-mentioned labeled video data set, an object behavior detection result set is generated. Thus, object behavior detection can be performed on the above-mentioned labeled video data set, and the errors caused by manual operations can be reduced. Secondly, in response to determining that there is an object behavior detection result representing a normal result of object behavior detection in the above-mentioned object behavior detection result set, at least one video feature set corresponding to the object behavior detection result representing the normal result of object behavior detection is determined as the first video feature set. Thus, manual analysis of the trajectory information of moving objects can be avoided. Thirdly, in response to determining that there is an object behavior detection result representing an abnormal result of object behavior detection in the above-mentioned object behavior detection result set, at least one video feature set corresponding to the object behavior detection result representing the abnormal result of object behavior detection is determined as the second video feature set. Thus, manual analysis of the trajectory information of moving objects can be avoided, and the generated errors can be corrected in time. Therefore, the response speed of the system is increased, and the security of the system is improved. After that, the above-mentioned first video feature set and the above-mentioned second video feature set are input into a pre-trained object feature prediction model to obtain an object feature prediction result. Thus, the features of the trajectory information can be automatically predicted. Finally, the object motion trajectory information corresponding to the above-mentioned object feature prediction result is verified to obtain verified trajectory information for the system to determine the object category. Thus, the security of the system can be improved, which is helpful for determining the object category. Therefore, the security of the system is improved, and the response speed of the system is increased. BRIEF DESCRIPTION OF THE DRAWINGS

[0014] In combination with the accompanying drawings and with reference to the following specific embodiments, the above and other features, advantages and aspects of the various embodiments of the present disclosure will become more obvious. Throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic, and the elements and elements are not necessarily drawn to scale.

[0015] Figure 1 is a flowchart of some embodiments of the information verification method according to the present disclosure;

[0016] Figure 2 is a schematic structural diagram of some embodiments of the information verification apparatus according to the present disclosure;

[0017] Figure 3 is a schematic structural diagram of an electronic device suitable for implementing some embodiments of the present disclosure. Detailed Embodiments

[0018] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for illustrative purposes and are not used to limit the protection scope of the present disclosure.

[0019] In addition, it should be noted that for the sake of convenience of description, only parts related to the relevant invention are shown in the drawings. Without conflict, the embodiments in the present disclosure and the features in the embodiments can be combined with each other.

[0020] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or mutual dependence relationship of the functions performed by these devices, modules or units.

[0021] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0022] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0023] The present disclosure will be described in detail below with reference to the drawings and in combination with the embodiments.

[0024] Figure 1 is Process 100 of some embodiments of the information verification method of the present disclosure. The information verification method includes the following steps:

[0025] Step 101, obtain video data captured by a camera within a preset time period as initial video data.

[0026] In some embodiments, the execution subject of the information verification method (e.g., a computing device) may obtain video data captured by a camera within a preset time period through a wired connection or a wireless connection as initial video data, where the initial video data includes: video data of the same environment at different times and video data of the same time in different environments.

[0027] Here, the preset time period may refer to a preset time range. For example, the preset time period may be from 14:00:00 on April 20, 2024 to 12:00:00 on April 21, 2024.

[0028] It should be noted that the wireless connection method may include, but is not limited to, 3G / 4G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other currently known or future-developed wireless connection methods.

[0029] Step 102: Generate an annotated video data set based on the initial video data.

[0030] In some embodiments, the execution subject may generate an annotated video data set based on the initial video data.

[0031] Here, the annotated video data set may refer to the video data set obtained by labeling the initial video data. The label may be a label marked with "user-carrying items".

[0032] Optionally, the execution subject may generate an annotated video data set based on the initial video data through the following steps:

[0033] First step: Clean the initial video data to obtain cleaned video data.

[0034] Here, the data cleaning can be used to remove invalid and incorrect data from the initial video data.

[0035] Second step: Enhance the cleaned video data to obtain enhanced video data.

[0036] Here, the video enhancement may refer to enhancing the video contrast.

[0037] As an example, the execution subject may enhance the cleaned video data through sharpening to obtain enhanced video data.

[0038] Third step: Identify at least one moving object in the enhanced video data to obtain a set of identified moving objects.

[0039] Here, the above-mentioned set of recognized moving objects can refer to the set of recognized moving objects.

[0040] As an example, the above-mentioned execution entity can perform at least one moving object recognition on the above-mentioned enhanced video data through an optical flow method to obtain a set of recognized moving objects.

[0041] Fourth step, perform label annotation on the above-mentioned set of recognized moving objects to obtain a set of labeled moving objects.

[0042] As an example, the above-mentioned execution entity can perform label annotation on the above-mentioned set of recognized moving objects through a labeling tool to obtain a set of labeled moving objects. For example, the above-mentioned labeling tool can refer to Prodigy.

[0043] Fifth step, perform moving object screening on the above-mentioned set of labeled moving objects to obtain a set of screened moving objects.

[0044] Here, the above-mentioned set of screened moving objects can refer to the set of objects selected from the set of labeled moving objects.

[0045] Sixth step, perform correlation analysis on the above-mentioned set of screened moving objects to obtain a result set after correlation analysis.

[0046] Here, the result after correlation analysis in the above-mentioned result set after correlation analysis can refer to the result of the objects with a correlation relationship in the set of screened moving objects. The above-mentioned correlation relationship can refer to "a user carrying a schoolbag", where the user is a moving object and the schoolbag is an object having a correlation relationship with the user.

[0047] As an example, the above-mentioned execution entity can perform correlation analysis on the above-mentioned set of screened moving objects through clustering analysis to obtain a result set after correlation analysis.

[0048] Seventh step, in response to determining that there is a result after correlation analysis in the above-mentioned result set after correlation analysis, where the corresponding threshold of the represented correlation is greater than or equal to a preset correlation threshold, determine the set of moving objects corresponding to at least one result after correlation analysis whose corresponding threshold of the represented correlation is greater than or equal to the preset correlation threshold as the labeled video data set.

[0049] Here, the above-mentioned preset correlation threshold can refer to a threshold for defining correlation set in advance. For example, the above-mentioned preset correlation threshold can refer to 0.8.

[0050] Eighth step, in response to determining that there is a result after correlation analysis in the above-mentioned result set after correlation analysis, where the corresponding threshold of the represented correlation is less than the above-mentioned preset correlation threshold, perform preset label annotation on the set of moving objects corresponding to at least one result after correlation analysis whose corresponding threshold of the represented correlation is less than the above-mentioned preset correlation threshold to obtain a labeled video data set.

[0051] Optionally, the above-mentioned execution entity may generate an annotated video dataset according to the above-mentioned initial video data through the following steps:

[0052] In the first step, perform abnormal data processing on the above-mentioned initial video data to obtain processed initial video data.

[0053] Here, the above-mentioned processed initial video data may refer to the initial video data after removing abnormal data.

[0054] In the second step, input the above-mentioned processed initial video data into an object recognition model to obtain an object information set of the recognized video data.

[0055] Here, the above-mentioned object recognition model may refer to a Long Short-Term Memory (LSTM) model and a Region Proposal Network (RPN) model. The above-mentioned Region Proposal Network refers to a network used for object behavior detection and generating candidate regions.

[0056] In the third step, perform similarity processing on each object information of the recognized video data object information set in the above to generate elements corresponding to the similarity matrix, and obtain a similarity matrix.

[0057] Here, the above-mentioned similarity processing may be cosine similarity. Here, the above-mentioned similarity matrix may be a matrix representing the similarity of each object information of the recognized video data.

[0058] In the fourth step, generate a similarity graph according to the above-mentioned similarity matrix, where the nodes in the above-mentioned similarity graph represent object information, and the edges in the above-mentioned similarity graph represent the corresponding similarities in the similarity matrix.

[0059] As an example, the above-mentioned execution entity may first traverse each element in the above-mentioned similarity matrix, then add the similarities whose corresponding similarities are greater than a preset similarity threshold to the graph, and then connect the corresponding nodes in the similarity matrix according to the similarity values greater than the preset similarity threshold to generate a similarity graph. Here, there is no limitation on the setting of the above-mentioned preset similarity threshold.

[0060] In the fifth step, perform a pruning operation on the above-mentioned similarity graph to obtain a pruned graph, where the above-mentioned pruned graph represents a graph after removing the corresponding redundant edges in the above-mentioned similarity graph.

[0061] As an example, the above-mentioned execution entity may refer to removing the similarities in the above similarity map that are lower than the similarity threshold of the preset map according to the similarity threshold of the preset map, and obtaining the map after removal as the pruned map. The above preset map may refer to a preset map.

[0062] Step 6: Perform index estimation processing on the above pruned map to obtain a clustering estimation result.

[0063] Here, the above index estimation processing may refer to fuzzy silhouette coefficient processing. The above clustering estimation result may refer to an evaluation value result used to measure the quality of the clustering result.

[0064] Step 7: In response to determining that the above clustering estimation result is greater than the preset estimation condition, determine each node in the similarity map corresponding to the above clustering estimation result that is greater than the preset estimation condition as the first type of label information group.

[0065] Here, the above preset estimation condition may be that the value corresponding to the clustering estimation result is greater than the preset estimation value. For example, the above preset estimation value may refer to 5.

[0066] Step 8: In response to determining that the above clustering estimation result is less than the preset estimation condition, determine each node in the similarity map corresponding to the above clustering estimation result that is less than the preset estimation condition as the second type of label information group.

[0067] Step 9: In response to determining that the above clustering estimation result is equal to the preset estimation condition, determine each node in the similarity map corresponding to the above clustering estimation result that is equal to the preset estimation condition as the third type of label information group.

[0068] Step 10: Determine the above first type of label information group, the above second type of label information group, and the above third type of label information group as the set of labeled label information groups.

[0069] Step 11: Integrate the video data corresponding to each labeled label information in the above set of labeled label information groups to obtain an integrated video data set as the labeled video data set.

[0070] As an example, the above execution entity may merge the video data corresponding to each labeled label information in the above set of labeled label information groups to obtain a merged video data set as the labeled video data set.

[0071] The above related content, as an inventive point of the present disclosure, solves the second technical problem mentioned in the background art: "When video data annotates object information, misrecognition may occur, resulting in a high error rate of the annotated video data, and thus the quality of the annotated video data is poor. Since the correlation between objects cannot be determined, the annotation accuracy is low, resulting in poor quality of the annotated video data." The factors that lead to a high error rate of the annotated video data and poor quality of the annotated video data are usually as follows: When video data annotates object information, misrecognition may occur, resulting in a high error rate of the annotated video data, and thus the quality of the annotated video data is poor. Since the correlation between objects cannot be determined, the annotation accuracy is low, resulting in poor quality of the annotated video data. If the above factors are solved, the efficiency of label annotation can be improved. To achieve this effect, first, perform abnormal data processing on the above initial video data to obtain the processed initial video data. Thus, abnormal data can be removed, and the quality of the annotated video data can be improved. Second, input the above processed initial video data into an object recognition model to obtain an object information set of the recognized video data. Thus, the situation where misrecognition may occur when video data annotates object information due to manual recognition can be avoided. Third, perform similarity processing on each object information of the recognized video data object information set in the above to generate elements corresponding to the similarity matrix, and obtain the similarity matrix. Thus, the correlation between objects can be determined. Fourth, generate a similarity graph according to the above similarity matrix, where the nodes in the above similarity graph represent object information, and the edges in the above similarity graph represent the similarities corresponding in the similarity matrix. Fifth, perform a pruning operation on the above similarity graph to obtain a pruned graph, where the above pruned graph represents the graph after removing the corresponding redundant edges in the above similarity graph. Sixth, perform index estimation processing on the above pruned graph to obtain a clustering estimation result. Seventh, in response to determining that the above clustering estimation result is greater than a preset estimation condition, determine each node in the similarity graph corresponding to the above clustering estimation result that is greater than the preset estimation condition as the first type of label information group. Eighth, in response to determining that the above clustering estimation result is less than the preset estimation condition, determine each node in the similarity graph corresponding to the above clustering estimation result that is less than the preset estimation condition as the second type of label information group. Ninth, in response to determining that the above clustering estimation result is equal to the preset estimation condition, determine each node in the similarity graph corresponding to the above clustering estimation result that is equal to the preset estimation condition as the third type of label information group. Tenth, determine the above first type of label information group, the above second type of label information group, and the above third type of label information group as the set of annotated label information groups. Eleventh, perform data integration on the video data corresponding to each annotated label information in the above set of annotated label information groups to obtain an integrated video data set as the annotated video data set.Therefore, the accuracy of annotation can be improved, thereby improving the quality of the annotated video data.

[0072] Step 103: Generate an object behavior detection result set according to the above-mentioned annotated video data set.

[0073] In some embodiments, the above-mentioned execution subject may generate an object behavior detection result set according to the above-mentioned annotated video data set.

[0074] Here, the object behavior detection results in the above-mentioned object behavior detection result set may represent the behavior trajectory results of the detected object motion. The object behavior detection results in the above-mentioned object behavior detection result set may represent normal object behavior detection results and abnormal object behavior detection results.

[0075] Optionally, the above-mentioned execution subject may generate an object behavior detection result set according to the above-mentioned annotated video data set through the following steps:

[0076] The first step: Denoise the above-mentioned annotated video data set to obtain a denoised video data set.

[0077] As an example, the above-mentioned execution subject may denoise the above-mentioned annotated video data set through a filter to obtain a denoised video data set.

[0078] The second step: Extract video features from the above-mentioned denoised video data set to obtain an extracted video feature set, where the above-mentioned video features are video static features and video dynamic features.

[0079] As an example, the above-mentioned execution subject may extract video features from the above-mentioned denoised video data set through computer vision technology to obtain an extracted video feature set.

[0080] The third step: Detect motion trajectory information for the video dynamic features corresponding to the above-mentioned extracted video feature set to obtain a motion trajectory information set.

[0081] As an example, the above-mentioned execution subject may monitor the video dynamic features corresponding to the above-mentioned extracted video feature set and determine the motion trajectory information of each video dynamic feature to obtain a motion trajectory information set.

[0082] The fourth step: Perform object behavior detection on the above-mentioned motion trajectory information set to obtain an object behavior detection result set.

[0083] As an example, the above-mentioned execution subject may detect the object behavior corresponding to the above-mentioned motion trajectory information set to obtain an object behavior detection result set.

[0084] Step 104, in response to determining that there is an object behavior detection result representing a normal object behavior detection result in the above object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing a normal object behavior detection result as the first video feature set.

[0085] In some embodiments, the above execution subject may, in response to determining that there is an object behavior detection result representing a normal object behavior detection result in the above object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing a normal object behavior detection result as the first video feature set.

[0086] Here, the first video feature in the above first video feature set may refer to a video feature representing normality.

[0087] Step 105, in response to determining that there is an object behavior detection result representing an abnormal object behavior detection result in the above object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing an abnormal object behavior detection result as the second video feature set.

[0088] In some embodiments, the above execution subject may, in response to determining that there is an object behavior detection result representing an abnormal object behavior detection result in the above object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing an abnormal object behavior detection result as the second video feature set.

[0089] Here, the second video feature in the above second video feature set may refer to a video feature representing abnormality.

[0090] Step 106, input the above first video feature set and the above second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result.

[0091] In some embodiments, the above execution subject may input the above first video feature set and the above second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result.

[0092] Here, the above object feature prediction model can refer to a model that takes the above first video feature set and the above second video feature set as inputs and outputs object feature prediction results. The above object feature prediction model can include: a long short-term memory (LSTM) model and a graph convolutional network (GCN) model. The above long short-term memory network model is used to process the dynamic features in the time series corresponding to the above first video feature set and the above second video feature set. The above graph convolutional network model can obtain the complex relationships between dynamic features. The above object feature prediction model can refer to a model used to predict the motion trajectory of an object's features over a future period of time.

[0093] Optionally, the above object feature prediction model is trained through the following steps:

[0094] First step, obtain a video feature sample set.

[0095] Here, obtaining a video feature sample set means randomly obtaining a video feature sample set. Here, the video feature samples in the video feature sample set refer to the feature content of the video captured by the camera within a preset time period.

[0096] Second step, select video feature samples from the above video feature sample set.

[0097] Here, selecting video feature samples from the above video feature sample set means randomly selecting video feature samples.

[0098] Third step, input the above video feature samples into an initial convolutional neural network model to obtain object feature prediction test data.

[0099] Here, the initial convolutional neural network model can be a convolutional neural network model that has not undergone model training. Here, the object feature prediction test data refers to, for video feature samples, the corresponding object feature prediction test data means dividing the video features into multiple grid video features with the same ratio.

[0100] Fourth step, based on a preset video feature data loss function, determine the data difference value between the above object feature prediction test data and the video feature sample labels included in the above video feature samples.

[0101] Here, the above video feature data loss function can include but is not limited to: mean squared error loss function (MSE), cross-entropy loss function (CrossEntropy), etc.

[0102] Fifth step, in response to the above data difference value being greater than or equal to a preset data threshold, adjust the network parameters of the above object feature prediction model.

[0103] In some embodiments, the above-mentioned execution entity may adjust the network parameters of the above-mentioned initial convolutional operator prediction model in response to the above-mentioned data difference value being greater than or equal to a preset data threshold. Here, there is no limitation on the setting of the preset data threshold. For example, the difference between the data difference value and the preset data threshold can be calculated to obtain a loss difference. On this basis, methods such as backpropagation and stochastic gradient descent are used to forward the data difference value from the last layer of the model to adjust the parameters of each layer. Of course, according to needs, the method of network freezing (dropout) can also be adopted to keep the network parameters of some of the layers unchanged and not adjusted. In this regard, no specific limitation is made.

[0104] Sixth step, in response to the above-mentioned data difference value being less than the above-mentioned preset data threshold, determine the above-mentioned initial convolutional neural network model as the object feature prediction model.

[0105] In some embodiments, the above-mentioned execution entity may determine the above-mentioned initial convolutional neural network model as the initial convolutional operator prediction model in response to the above-mentioned data difference value being less than the above-mentioned preset data threshold. Here, there is no limitation on the setting of the preset data threshold.

[0106] Step 107, verify the object motion trajectory information corresponding to the above-mentioned object feature prediction result to obtain the verified trajectory information for the system to determine the object category.

[0107] In some embodiments, the above-mentioned execution entity may verify the object motion trajectory information corresponding to the above-mentioned object feature prediction result to obtain the verified trajectory information for the system to determine the object category.

[0108] As an example, the above-mentioned execution entity may perform trajectory information verification on the object motion trajectory information corresponding to the above-mentioned object feature prediction result, and then, through the verified trajectory information, the system can determine the object category.

[0109] The above embodiments of the present disclosure have the following beneficial effects: Through the information verification method of some embodiments of the present disclosure, the security of the system is improved, and the response speed of the system is increased. Specifically, the reasons for the slow response speed of the system and the low security of the system are as follows: Analyzing the trajectory information of moving objects manually will cause large errors in information verification, and the generated errors cannot be corrected in time, resulting in a slow response speed of the system and a low security of the system. Based on this, in some embodiments of the information verification method of the present disclosure, first, video data captured by a camera within a preset time period is obtained as initial video data, where the above initial video data includes: video data of the same environment at different times, and video data of the same time in different environments. Thus, it can facilitate subsequent processing. Then, according to the above initial video data, a labeled video data set is generated. Thus, the above initial video data can be labeled, which is convenient for identifying the initial video data and is also beneficial for subsequent automatic analysis of the trajectory information of moving objects. After that, according to the above labeled video data set, an object behavior detection result set is generated. Thus, object behavior detection can be performed on the above labeled video data set, and the errors caused by manual operations can be reduced. Secondly, in response to determining that there is an object behavior detection result representing a normal result of object behavior detection in the above object behavior detection result set, at least one video feature set corresponding to the object behavior detection result representing the normal result of object behavior detection is determined as the first video feature set. Thus, manual analysis of the trajectory information of moving objects can be avoided. Thirdly, in response to determining that there is an object behavior detection result representing an abnormal result of object behavior detection in the above object behavior detection result set, at least one video feature set corresponding to the object behavior detection result representing the abnormal result of object behavior detection is determined as the second video feature set. Thus, manual analysis of the trajectory information of moving objects can be avoided, and the generated errors can be corrected in time. Therefore, the response speed of the system is increased, and the security of the system is improved. After that, the above first video feature set and the above second video feature set are input into a pre-trained object feature prediction model to obtain an object feature prediction result. Thus, the features of the trajectory information can be automatically predicted. Finally, the object motion trajectory information corresponding to the above object feature prediction result is verified to obtain verified trajectory information for the system to determine the object category. Thus, the security of the system can be improved, which is helpful for determining the object category. Therefore, the security of the system is improved, and the response speed of the system is increased.

[0110] Further referring to Figure 2 , as an implementation of the methods shown in the above figures, the present disclosure provides some embodiments of an information verification method. These apparatus embodiments correspond to Figure 1 the method embodiments shown, and the apparatus can be specifically applied to various electronic devices.

[0111] As Figure 2 shown, the information verification device 200 of some embodiments includes: an acquisition unit 201, a first generation unit 202, a second generation unit 203, a first determination unit 204, a second determination unit 205, an input unit 206, and a verification unit 207. Among them, the acquisition unit 201 is configured to acquire video data captured by a camera within a preset time period as initial video data, where the initial video data includes: video data of the same environment at different times, and video data of different environments at the same time; the first generation unit 202 is configured to generate an annotated video data set according to the initial video data; the second generation unit 203 is configured to generate an object behavior detection result set according to the annotated video data set; the first determination unit 204 is configured to, in response to determining that there is an object behavior detection result representing a normal result of object behavior detection in the object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing a normal result of object behavior detection as the first video feature set; the second determination unit 205 is configured to, in response to determining that there is an object behavior detection result representing an abnormal result of object behavior detection in the object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing an abnormal result of object behavior detection as the second video feature set; the input unit 206 is configured to input the first video feature set and the second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result; the verification unit 207 is configured to verify the object motion trajectory information corresponding to the object feature prediction result to obtain verified trajectory information for the system to determine the object category.

[0112] It can be understood that the various units described in the device 200 correspond to the respective steps in the method described with reference to Figure 1 Therefore, the operations, features, and beneficial effects described above for the method also apply to the device 200 and the units included therein, and will not be elaborated here.

[0113] Next, refer to Figure 3 , which shows a schematic structural diagram of an electronic device (such as a computing device) 300 suitable for implementing some embodiments of the present disclosure. Figure 3 The electronic device shown is only an example and should not impose any limitation on the functions and usage scopes of the embodiments of the present disclosure.

[0114] As Figure 3As shown, the electronic device 300 may include a processing device (such as a central processing unit, a graphics processing unit, etc.) 301, which may perform various appropriate actions and processes according to a program stored in the read-only memory (ROM) 302 or a program loaded from the storage device 308 into the random access memory (RAM) 304. In the RAM 303, various programs and data required for the operation of the electronic device 300 are also stored. The processing device 301, the ROM 302, and the RAM 304 are connected to each other through the bus 304. The input / output (I / O) interface 305 is also connected to the bus 304.

[0115] Generally, the following devices may be connected to the I / O interface 305: an input device 306 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.; an output device 307 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 308 including, for example, a magnetic tape, a hard disk, etc.; and a communication device 309. The communication device 309 may allow the electronic device 300 to communicate with other devices wirelessly or wiredly to exchange data. Although Figure 3 the electronic device 300 with various devices is shown, it should be understood that it is not required to implement or have all the shown devices. More or fewer devices may be implemented or had alternatively. Figure 3 Each block shown in may represent one device or, as needed, multiple devices.

[0116] In particular, according to some embodiments of the present disclosure, the processes described above with reference to the flowcharts may be implemented as computer software programs. For example, some embodiments of the present disclosure include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program codes for executing the methods shown in the flowcharts. In such some embodiments, the computer program may be downloaded and installed from the network through the communication device 309, or installed from the storage device 308, or installed from the ROM 302. When the computer program is executed by the processing device 301, the functions defined in the methods of some embodiments of the present disclosure are executed.

[0117] It should be noted that the computer-readable media described in some embodiments of the present disclosure may be a computer-readable signal medium, a computer-readable storage medium, or any combination of the two. A computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples of the computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, 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), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination thereof. In some embodiments of the present disclosure, the computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In some embodiments of the present disclosure, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium may be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.

[0118] In some embodiments, the client and the server can communicate using any currently known or future-developed network protocol such as HTTP (Hyper Text Transfer Protocol), and can be interconnected with digital data communication in any form or medium (e.g., a communication network). Examples of communication networks include local area networks ("LANs"), wide area networks ("WANs"), the Internet (e.g., the Internet), and end-to-end networks (e.g., ad hoc end-to-end networks), as well as any currently known or future-developed networks.

[0119] The above computer-readable medium may be included in the above electronic device; or it may exist independently and not be assembled into the electronic device. The above computer-readable medium carries one or more programs. When the above one or more programs are executed by the electronic device, the electronic device is caused to: obtain video data captured by a camera within a preset time period as initial video data, where the initial video data includes: video data of the same environment at different times, and video data of the same time in different environments; generate an annotated video data set based on the initial video data; generate an object behavior detection result set based on the annotated video data set; in response to determining that there is an object behavior detection result representing a normal object behavior detection result in the object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing the normal object behavior detection result as the first video feature set; in response to determining that there is an object behavior detection result representing an abnormal object behavior detection result in the object behavior detection result set, determine the video feature set corresponding to at least one object behavior detection result representing the abnormal object behavior detection result as the second video feature set; input the first video feature set and the second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result; verify the object motion trajectory information corresponding to the object feature prediction result to obtain verified trajectory information for the system to determine the object category.

[0120] Computer program code for performing the operations of some embodiments of the present disclosure may be written in one or more programming languages or combinations thereof. The programming languages include object-oriented programming languages such as Java, Smalltalk, C++, and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code may execute entirely on the user's computer, partially on the user's computer, execute as a stand-alone software package, execute partially on the user's computer and partially on a remote computer, or execute entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., by using an Internet service provider to connect through the Internet).

[0121] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flowchart or block diagram may represent a module, a segment of a program, or a part of code that contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions marked in the blocks may occur in an order different from that marked in the accompanying drawings. For example, two consecutive blocks shown may actually be executed substantially in parallel, and they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each block in the block diagram and / or flowchart, as well as combinations of blocks in the block diagram and / or flowchart, can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.

[0122] The units described in some embodiments of the present disclosure can be implemented in software or in hardware. The described units can also be provided in a processor. For example, it can be described as: a processor includes: an acquisition unit, a first generation unit, a second generation unit, a first determination unit, a second determination unit, an input unit, and a verification unit. Among them, the names of these units do not constitute a limitation on the unit itself in some cases. For example, the verification unit can also be described as "a unit that verifies the object motion trajectory information corresponding to the object feature prediction result to obtain the verified trajectory information for the system to determine the object category".

[0123] The functions described above can be at least partially performed by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that can be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on a chip (SOCs), complex programmable logic devices (CPLDs), and so on.

[0124] The above description is only some preferred embodiments of the present disclosure and an explanation of the technical principles applied above. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by the specific combination of the above technical features, and should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above inventive concept. For example, technical solutions formed by mutually replacing the above features with technical features (but not limited to) having similar functions disclosed in the embodiments of the present disclosure.

Claims

1. An information verification method, comprising: Acquire video data captured by a camera within a preset time period as initial video data, wherein the initial video data includes: video data of the same environment at different times, and video data of the same time at different environments; Generate a labeled video dataset based on the initial video data, wherein generating a labeled video dataset based on the initial video data includes: Performing data cleaning on the initial video data to obtain cleaned video data; Performing video enhancement on the cleaned video data to obtain enhanced video data; Performing at least one moving object recognition on the enhanced video data to obtain a set of recognized moving objects; Labeling the identified moving object set to obtain a labeled moving object set; Performing moving object screening on the marked moving object set to obtain a screened moving object set; Performing correlation analysis on the screened moving object set to obtain a correlation analysis result set; In response to determining that there is a correlation analysis result in the correlation analysis result set whose corresponding threshold for representing correlation is greater than or equal to a preset correlation threshold, determining a moving object set corresponding to at least one correlation analysis result corresponding to the corresponding threshold for representing correlation is greater than or equal to the preset correlation threshold as a labeled video data set; In response to determining that there is a correlation analysis result in the correlation analysis result set whose correlation corresponding threshold is less than the preset correlation threshold, annotating a moving object set corresponding to at least one correlation analysis result corresponding to the correlation corresponding threshold less than the preset correlation threshold with a preset label to obtain a labeled video data set; Generating an object behavior detection result set according to the annotated video data set; In response to determining that there is an object behavior detection result representing a normal object behavior detection result in the object behavior detection result set, determining a video feature set corresponding to at least one object behavior detection result representing a normal object behavior detection result as a first video feature set; In response to determining that there is an object behavior detection result representing an abnormal object behavior detection result in the object behavior detection result set, determining a video feature set corresponding to at least one object behavior detection result representing an abnormal object behavior detection result as a second video feature set; Inputting the first video feature set and the second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result; The object motion trajectory information corresponding to the object feature prediction result is verified to obtain verified trajectory information for the system to determine the object category.

2. The method according to claim 1, wherein: The step of generating an object behavior detection result set according to the annotated video data set includes: Performing denoising processing on the annotated video data set to obtain a denoised video data set; Performing video feature extraction on the denoised video data set to obtain an extracted video feature set, wherein the video features include video static features and video dynamic features; Performing motion trajectory information detection on the video dynamic features corresponding to the extracted video feature set to obtain a motion trajectory information set; Object behavior detection is performed on the motion trajectory information set to obtain an object behavior detection result set.

3. The method according to claim 1, wherein: The object feature prediction model is trained by the following steps: Obtain a video feature sample set; Selecting a video feature sample from the video feature sample set; Inputting the video feature samples into an initial convolutional neural network model to obtain object feature prediction test data; Based on a preset video feature data loss function, determining a data difference value between the object feature prediction test data and the video feature sample label included in the video feature sample; In response to the data difference value being greater than or equal to a preset data threshold, adjusting a network parameter of the object feature prediction model.

4. The method according to claim 3, wherein: The method further comprises: In response to the data difference value being less than the preset data threshold, the initial convolutional neural network model is determined as an object feature prediction model.

5. An information verification device, comprising: An acquisition unit is configured to acquire video data captured by a camera within a preset time period as initial video data, wherein the initial video data includes: video data of the same environment at different times, and video data of the same time at different environments; A first generating unit is configured to generate a labeled video data set according to the initial video data; The second generating unit is configured to generate an object behavior detection result set according to the labeled video data set, wherein the generating of the labeled video data set according to the initial video data includes: performing data cleaning on the initial video data to obtain cleaned video data; performing video enhancement on the cleaned video data to obtain enhanced video data; performing at least one moving object recognition on the enhanced video data to obtain a recognized moving object set; performing label marking on the recognized moving object set to obtain a labeled moving object set; performing moving object screening on the labeled moving object set to obtain a screened moving object set; and performing correlation analysis on the screened moving object set. analysis to obtain a result set after correlation analysis; in response to determining that there is a result after correlation analysis in the result set after correlation analysis whose corresponding threshold value representing correlation is greater than or equal to a preset correlation threshold, determining a moving object set corresponding to at least one result after correlation analysis corresponding to the corresponding threshold value representing correlation is greater than or equal to the preset correlation threshold as a labeled video data set; in response to determining that there is a result after correlation analysis in the result set after correlation analysis whose corresponding threshold value representing correlation is less than the preset correlation threshold, annotating a moving object set corresponding to at least one result after correlation analysis corresponding to the corresponding threshold value representing correlation is less than the preset correlation threshold with a preset label, to obtain a labeled video data set; A first determining unit is configured to, in response to determining that there is an object behavior detection result representing a normal object behavior detection result in the object behavior detection result set, determine a video feature set corresponding to at least one object behavior detection result representing a normal object behavior detection result as a first video feature set; A second determining unit is configured to, in response to determining that there is an object behavior detection result representing an abnormal object behavior detection result in the object behavior detection result set, determine a video feature set corresponding to at least one object behavior detection result representing an abnormal object behavior detection result as a second video feature set; An input unit, configured to input the first video feature set and the second video feature set into a pre-trained object feature prediction model to obtain an object feature prediction result; The verification unit is configured to verify the object motion trajectory information corresponding to the object feature prediction result to obtain verified trajectory information for the system to determine the object category.

6. An electronic device comprising: one or more processors; a storage device having one or more programs stored thereon; When the one or more programs are executed by the one or more processors, the one or more processors implement the method according to any one of claims 1 to 4.

7. A computer readable medium having a computer program stored thereon, wherein: When the program is executed by a processor, the method according to any one of claims 1 to 4 is implemented.

Images