Video watermark adding method and device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a video watermark adding method and device, electronic equipment and a storage medium, belongs to the technical field of video processing, and is applied to financial scenes and medical scenes. The method comprises the following steps: acquiring claim type information, field investigation information and a field investigation video; performing video blocking on the scene investigation video to obtain at least one key frame video block; performing key information detection on each key frame video block based on the field investigation information to obtain target detection information of each key frame video block; on the basis of each piece of target detection information, carrying out watermark region generation on the scene investigation information to obtain watermark region information of each key frame video block; and based on the claim type information, the field investigation information and the watermark area information of each key frame video block, carrying out watermark addition on each key frame video block to obtain a target watermark video. According to the embodiment of the invention, the interference of the watermark on key information can be reduced when the watermark is added.

Description

Technical Field

[0001] This application relates to the field of video processing technology and is applied to financial scenarios and medical scenarios. In particular, it relates to a method and device for adding video watermarks, an electronic device, and a storage medium. Background Art

[0002] Watermarks can be used in different scenarios such as images, texts, voices, videos, etc. Taking adding watermarks to videos as an example, watermark information is usually added to video frames. However, the added watermark information may affect the information in the original video frames, for example, blocking relevant information. For example, in the claims settlement scenario, some video frames may contain important information, and after adding watermarks, important information may be blocked, thus affecting the claims settlement judgment. Therefore, how to reduce the interference of watermarks on key information when adding watermarks has become an urgent problem to be solved. Summary of the Invention

[0003] The main purpose of the embodiments of this application is to propose a method and device for adding video watermarks, an electronic device, and a storage medium, aiming to reduce the interference of watermarks on key information when adding watermarks.

[0004] To achieve the above object, the first aspect of the embodiments of this application proposes a method for adding video watermarks, and the method includes:

[0005] Obtain a claims settlement report, where the claims settlement report includes claims settlement type information, on-site inspection information, and on-site inspection video;

[0006] Perform video segmentation on the on-site inspection video to obtain at least one key frame video block;

[0007] Based on the on-site inspection information, perform key information detection on each of the key frame video blocks to obtain target detection information for each of the key frame video blocks;

[0008] Based on each of the target detection information, generate a watermark area for the on-site inspection information to obtain watermark area information for each of the key frame video blocks;

[0009] Based on the claims settlement type information, the on-site inspection information, and the watermark area information of each of the key frame video blocks, add watermarks to each of the key frame video blocks to obtain a target watermark video.

[0010] In some embodiments, the step of based on the on-site inspection information, performing key information detection on each of the key frame video blocks to obtain target detection information for each of the key frame video blocks includes:

[0011] For each of the key frame video blocks, perform entity recognition on the key frame video block to obtain entity recognition information;

[0012] For each of the key-frame video blocks, calculate the entity criticality of the entity recognition information based on the on-site investigation information and the entity recognition information;

[0013] Based on the entity criticality, screen the entity recognition information to obtain the target detection information.

[0014] In some embodiments, calculating the entity criticality of the entity recognition information based on the on-site investigation information and the entity recognition information includes:

[0015] Perform entity recognition on the on-site investigation information to obtain at least one investigation information entity;

[0016] Based on the on-site investigation information, generate attention scores for each of the investigation information entities through an attention model;

[0017] Based on the attention scores of each of the investigation information entities, determine the entity criticality of the entity recognition information.

[0018] In some embodiments, determining the entity criticality of the entity recognition information based on the attention scores of each of the investigation information entities includes:

[0019] For each of the investigation information entities, confirm the matching degree between the investigation information entity and the entity recognition information based on the investigation information entity and the entity recognition information;

[0020] For each of the matching degrees, screen out target matching entities from each of the entity recognition information based on the matching degree;

[0021] Based on the target matching entities and the attention scores of the investigation information entities, determine the entity criticality of the entity recognition information.

[0022] In some embodiments, adding watermarks to each of the key-frame video blocks based on the claim type information, the on-site investigation information, and the watermark area information of each of the key-frame video blocks to obtain a target watermark video includes:

[0023] Perform information splitting on the on-site investigation information to obtain at least one sub-investigation information;

[0024] Based on the claim type information, sort each of the sub-investigation information to obtain an investigation information sequence;

[0025] Obtain the region boxes for each of the watermark region information, and based on each of the region boxes and the survey information sequence, screen each sub-survey information to determine the target survey information for each of the watermark region information;

[0026] Based on the on-site survey information and the target survey information for each of the watermark region information, add watermarks to each of the key frame video blocks to obtain the target watermark video.

[0027] In some embodiments, the adding watermarks to each of the key frame video blocks based on the on-site survey information and the target survey information for each of the watermark region information to obtain the target watermark video includes:

[0028] For the watermark region information of each of the key frame video blocks, generate explicit watermark data based on the target survey information corresponding to the watermark region information;

[0029] For each of the key frame video blocks, add watermarks to each of the key frame video blocks based on the explicit watermark data to obtain explicit watermark video frames;

[0030] Perform video frame splicing on each of the explicit watermark video frames to obtain a preliminary watermark video;

[0031] Perform implicit watermark addition to the preliminary watermark video based on the on-site survey information to obtain the target watermark video.

[0032] In some embodiments, the performing implicit watermark addition to the preliminary watermark video based on the on-site survey information to obtain the target watermark video includes:

[0033] Perform frequency domain conversion on the preliminary watermark video to obtain a frequency domain watermark video;

[0034] Perform implicit watermark addition to the frequency domain watermark video based on the on-site survey information to obtain a preliminary frequency domain video;

[0035] Perform inverse frequency domain conversion on the preliminary frequency domain video to obtain the target watermark video.

[0036] To achieve the above object, a second aspect of the embodiments of the present application proposes a video watermark addition device, and the device includes:

[0037] A data acquisition module, configured to acquire a claim report, where the claim report includes claim type information, on-site survey information, and an on-site survey video;

[0038] A video segmentation module, configured to segment the on-site survey video to obtain at least one key frame video block;

[0039] An information monitoring module, configured to perform key information detection on each of the key frame video blocks based on the on-site inspection information, so as to obtain the target detection information of each of the key frame video blocks;

[0040] A region generation module, configured to generate a watermark region for the on-site inspection information based on each of the target detection information, so as to obtain the watermark region information of each of the key frame video blocks;

[0041] A watermark addition module, configured to add a watermark to each of the key frame video blocks based on the claim type information, the on-site inspection information, and the watermark region information of each of the key frame video blocks, so as to obtain a target watermark video.

[0042] To achieve the above object, a third aspect of the embodiments of the present application provides an electronic device, where the electronic device includes a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect above is implemented.

[0043] To achieve the above object, a fourth aspect of the embodiments of the present application provides a computer-readable storage medium, where the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect above is implemented.

[0044] A video watermark addition method, device, electronic device, and storage medium provided by the present application obtain claim type information, on-site inspection information, and an on-site inspection video, and perform video segmentation on the on-site inspection video, thereby processing the video in a refined manner and improving the accuracy of adding a watermark to the video; further, key information detection is performed using the on-site inspection information, so that the target detection information accurately corresponds to each key frame video block; then, a video watermark region for each key frame is generated based on the target detection information, avoiding important content in the video frame and preventing the watermark information from obscuring key claim content; finally, watermark addition is achieved by combining the claim type information, the on-site inspection information, and the watermark region information, reducing the possibility of the watermark interfering with claim judgment and realizing reducing the interference of the watermark on key information when adding the watermark. Description of the Drawings

[0045] Figure 1 is a flowchart of the video watermark addition method provided by the embodiments of the present application;

[0046] Figure 2 is Figure 1 a flowchart of step S102 in

[0047] Figure 3 is Figure 2 a flowchart of step S205 in

[0048] Figure 4 is Figure 3 the flowchart of step S303 in

[0049] Figure 5 is Figure 1 the flowchart of step S105 in

[0050] Figure 6 is Figure 5 the flowchart of step S504 in

[0051] Figure 7 is Figure 6 the flowchart of step S604 in

[0052] Figure 8 the structural schematic diagram of the video watermark adding device provided by the embodiments of the present application;

[0053] Figure 9 the hardware structural schematic diagram of the electronic device provided by the embodiments of the present application. Detailed implementation manners

[0054] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0055] It should be noted that although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be executed in a different order from the module division in the device or the order in the flowchart. Terms such as "first" and "second" in the specification, claims and the above drawings are used to distinguish similar objects, and do not necessarily need to describe a specific order or sequence.

[0056] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0057] First, several nouns involved in the present application are analyzed:

[0058] Frequency domain transformation: Frequency domain transformation is a mathematical method used to convert a signal or function from a time-domain representation to a frequency-domain representation. This technique is the foundation of signal processing and system analysis and belongs to the fields of electrical engineering and computer science. Frequency domain transformation is achieved through methods such as the Fourier transform or the Laplace transform. In engineering and scientific research, frequency domain transformation is widely applied in fields such as communication systems, acoustics, vibration analysis, image processing, and medical imaging. By analyzing the signals in the frequency domain, filters can be designed more effectively, the frequency response characteristics of signals can be analyzed, and the dynamic behavior of systems can be diagnosed. Frequency domain transformation not only provides a powerful tool for processing and interpreting complex signal data but also improves the efficiency and accuracy of signal processing.

[0059] Inverse frequency domain transformation: Inverse frequency domain transformation is a mathematical method that converts a signal or function from a frequency-domain representation back to a time-domain representation. This technique is an important part of signal processing and system analysis and belongs to the fields of electrical engineering and computer science. By applying methods such as the Inverse Fourier Transform or the Inverse Laplace Transform, inverse frequency domain transformation reconstructs the time-series signal from its frequency components. This process has wide applications in various fields such as communication systems, audio processing, image restoration, medical imaging, and data analysis in scientific research. Inverse frequency domain transformation is not only crucial for signal synthesis and restoration but also the basis for understanding and operating the results of frequency domain analysis.

[0060] Watermarks can be used in different scenarios such as images, text, speech, videos, etc. Taking adding watermarks to videos as an example, watermark information is usually added to video frames. However, the added watermark information may affect the information in the original video frames, such as obscuring relevant information. For example, in a claims settlement scenario, some video frames may contain important information, and after adding the watermark, the important information may be obscured, thus affecting the claims settlement judgment. Therefore, how to reduce the interference of the watermark on key information when adding the watermark has become an urgent problem to be solved.

[0061] Based on this, the embodiments of this application provide a video watermark adding method, device, electronic device, and storage medium, aiming to reduce the interference of the watermark on key information when adding the watermark.

[0062] The video watermark adding method, device, electronic device, and storage medium provided by the embodiments of this application are specifically described through the following embodiments. First, the video watermark adding method in the embodiments of this application is described.

[0063] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, Artificial Intelligence (AI) is the theory, method, technology, and application system that uses digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use knowledge to obtain the best results.

[0064] Artificial intelligence basic technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, and mechatronics. Artificial intelligence software technologies mainly include several major directions such as computer vision technology, robotics, biometric technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0065] The video watermark adding method provided by the embodiments of the present application relates to the field of video processing technology and is applied to financial scenarios and medical scenarios. The video watermark adding method provided by the embodiments of the present application can be applied to a terminal, or to a server side, or can be software running on a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or can be configured as a server cluster or distributed system composed of multiple physical servers, or can be configured as a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the video watermark adding method, etc., but is not limited to the above forms.

[0066] The present application can be used in many general-purpose or special-purpose computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet-type devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and so on. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment, where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0067] It should be noted that in each specific embodiment of the present application, when it comes to relevant processing based on data related to the user's identity or characteristics, such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first. Moreover, the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiments of the present application need to obtain the user's sensitive personal information, the user's separate permission or separate consent will be obtained through methods such as pop-up windows or redirecting to a confirmation page. After clearly obtaining the user's separate permission or separate consent, the necessary user-related data for the normal operation of the embodiments of the present application will be obtained.

[0068] Figure 1 is an optional flowchart of the video watermark addition method provided by the embodiments of the present application. Figure 1 The method in may include but is not limited to steps S101 to S105.

[0069] Step S101, obtain a claim settlement report, where the claim settlement report includes claim settlement type information, on-site investigation information, and on-site investigation videos;

[0070] Step S102, perform video segmentation on the on-site investigation video to obtain at least one key-frame video block;

[0071] Step S103, based on the on-site investigation information, perform key information detection on each key-frame video block to obtain the target detection information of each key-frame video block;

[0072] Step S104, based on each target detection information, generate a watermark area for the on-site investigation information to obtain the watermark area information of each key-frame video block;

[0073] Step S105, based on the claim settlement type information, on-site investigation information, and the watermark area information of each key-frame video block, add watermarks to each key-frame video block to obtain a target watermark video.

[0074] Steps S101 to S105 shown in the embodiments of the present application obtain claim type information, on-site investigation information, and on-site investigation videos, and perform video segmentation on the on-site investigation videos, thereby processing the videos in a refined manner and improving the accuracy of adding watermarks to the videos; further, key information detection is performed using the on-site investigation information, so that the target detection information accurately corresponds to each key frame video block; then, based on the target detection information, video watermark regions for each key frame are generated, avoiding important content in the video frames and preventing the watermark information from obscuring the key content of the claim; finally, watermark addition is achieved by combining the claim type information, on-site investigation information, and watermark region information, reducing the possibility of the watermark interfering with the claim judgment, and realizing the reduction of the interference of the watermark on the key information when adding the watermark.

[0075] In step S101 of some embodiments, the claim report is a comprehensive report generated based on accident investigation, evidence collection, and review processes, containing various information related to claims, and is mainly used to evaluate the reasonableness of insurance payouts. The claim type information refers to the category of the claim case, such as vehicle insurance claims, medical claims, property damage claims, etc. The on-site investigation information is the information collected by the investigation personnel during the investigation of the accident scene, including the accident occurrence time, location, weather conditions, damaged parts, personnel injury conditions, liability division analysis, etc. The on-site investigation video refers to the accident scene video taken by the accident investigator or monitoring equipment. The acquisition method can be obtained through various means. For example, the claim report can be obtained by querying from a predetermined database or uploaded by the claim adjuster.

[0076] In step S102 of some embodiments, the key frame video block is a part of the on-site investigation video, and all the key frame video blocks can be combined to restore the on-site investigation video. The determination of the key frame video block can be achieved in the following ways. Calculate the color histogram difference between adjacent frames or frames with a time interval. If the difference exceeds the set threshold, it is marked as a key frame. Or it can be achieved in the following way. Use a convolutional neural network (CNN) to extract the visual features of the frame, and learn the temporal information through a long short-term memory network (LSTM) to automatically select the key frames.

[0077] Please refer to Figure 2 , in some embodiments, step S103 may include but is not limited to steps S201 to S203:

[0078] Step S201, for each key frame video block, perform entity recognition on the key frame video block to obtain entity recognition information;

[0079] Step S202, for each key frame video block, calculate the entity key degree of the entity recognition information based on the on-site investigation information and the entity recognition information;

[0080] Step S203: Based on the entity criticality, filter the entity recognition information to obtain the target detection information.

[0081] Steps S201 to S203 illustrated in the embodiments of the present application extract entity information in the video by performing entity recognition on each key-frame video block, enabling subsequent processing to analyze based on specific entity features rather than relying solely on the overall image information, thereby improving the recognition ability of key content. Then, calculate the entity criticality by combining the on-site investigation information and the recognized entity information to evaluate the importance of each entity in the claim settlement scenario, so as to distinguish between key entities and secondary entities. Finally, filter the entity recognition information based on the entity criticality, and retain the target detection information that has practical value for claim settlement judgment, which determines which areas cannot be covered when adding watermarks later.

[0082] In step S201 of some embodiments, the entity recognition information is a structured data, including entity category, position coordinates, attribute information, etc. The attribute information is the attributes of the entity. For example, for a vehicle, the attribute information includes license plate number, vehicle brand, etc. When settling a claim, it is often necessary to determine the final claim settlement result based on the attribute information. For example, in the scenario of vehicle insurance, the license plate number helps to determine the two parties involved in the accident, the vehicle brand can be used to judge whether the damage to the vehicle conforms to the insurance terms, and the vehicle color helps to compare with other evidence (such as witness descriptions or surveillance videos) to confirm the identity of the accident vehicle. In addition, by combining the position coordinates in the entity recognition information, the relative position of the vehicle at the accident scene can be determined, thereby assisting in analyzing the division of accident liability. For example, if the accident occurs at an intersection, the position coordinate information can help determine which vehicle is in a straight-ahead state, which vehicle is in a left-turn or right-turn state, to judge whether there are violations such as running a red light or failing to yield as required. At the same time, for a multi-vehicle chain collision accident, by combining the entity recognition information of multiple key-frame video blocks, the sequence of the accident can be restored, so as to more accurately determine liability and process claim settlements. Therefore, through the structured entity recognition information, the claim settlement system can analyze the accident scene more accurately, improving the automation level of the claim settlement process and the accuracy of decision-making.

[0083] Perform entity recognition on key-frame video blocks to obtain entity recognition information, which can be achieved through the following methods. First, based on a pre-trained image entity category discrimination model, determine the category of the entity. For example, determine whether the entity is a vehicle, a person, a building, etc. Then, based on the corresponding entity, use the look-up table method to determine the attribute information that needs to be recognized. For example, for a building, information such as the house number, number of floors, and building use (such as residential, commercial, office building, etc.) can be queried to further determine its relevance in the accident scenario. For example, in a car accident claim scenario, if a vehicle collides with a certain building, identifying the specific information of the building helps to evaluate the damage situation and determine whether third-party property damage compensation is involved. For the category of people, information such as the person's posture, clothing characteristics, and whether they are wearing safety equipment (such as safety helmets, reflective vests) can be recognized to assist in determining accident liability. For example, in a construction site accident or a traffic accident, if it is recognized that the relevant personnel are not wearing safety equipment, it may affect the liability division of the claim. For the vehicle category, key attributes such as license plate number, brand, model, and color can be further recognized, and combined with the on-site investigation information, match the vehicle information in the insurance company's database to confirm whether it is the insured vehicle. In addition, by identifying the damaged parts of the vehicle, it is possible to assist in judging the collision direction and the degree of damage, and compare it with the accident description to improve the accuracy of claim review.

[0084] Please refer to Figure 3 , in some embodiments, step S202 may include but is not limited to steps S301 to S303:

[0085] Step S301, perform entity recognition on the on-site investigation information to obtain at least one investigation information entity;

[0086] Step S302, based on the on-site investigation information, generate attention scores for each investigation information entity through an attention model;

[0087] Step S303, based on the attention scores of each investigation information entity, determine the entity key degree of the entity recognition information.

[0088] Steps S301 to S303 shown in the embodiments of the present application perform entity recognition on the on-site investigation information, extract investigation information entities, and thus refine objects with practical significance, such as vehicles, pedestrians, traffic signals, or accident traces, etc., so as to provide structured information for subsequent processing. Then, use the attention model to calculate the attention scores of each investigation information entity based on the on-site investigation information, dynamically measure the importance of different entities in claim analysis, so that key entities can receive higher attention, while the influence of irrelevant or secondary entities is weakened. Finally, according to the attention scores of each investigation information entity, determine the entity key degree of the entity recognition information to achieve a quantitative evaluation of the importance of different entities.

[0089] In step S301 of some embodiments, the survey information entity and the entity recognition information in step 201 belong to the same structural data, which will not be elaborated here. Entity recognition is performed on the on-site survey information. To obtain the survey information entity, a pre-trained text entity category discrimination model can be used to determine the category of the entity. For example, it can be determined that the entity is a vehicle, a person, a building, etc. Then, based on the natural language model, the position coordinates and attribute information of the survey information entity are extracted from the on-site survey information.

[0090] In step S302 of some embodiments, the attention model is a deep network model that can assign different weights according to different features of the input data, making the model pay more attention to the information crucial to the task while weakening the irrelevant content. In this embodiment, the attention model calculates the importance scores of each survey information entity based on the on-site survey information to highlight the entity information crucial for claim settlement judgment. The attention model can be a deep network model known in the art, and the present application does not make specific limitations.

[0091] Please refer to Figure 4 , in some embodiments, step S303 may include but is not limited to steps S401 to S403:

[0092] Step S401, for each survey information entity, based on the survey information entity and the entity recognition information, confirm the matching degree between the survey information entity and the entity recognition information;

[0093] Step S402, for each matching degree, based on the matching degree, screen out the target matching entity from each entity recognition information;

[0094] Step S403, based on the target matching entity and the attention score of the survey information entity, determine the entity criticality of the entity recognition information.

[0095] Steps S401 to S403 illustrated in the embodiments of the present application calculate the matching degree between the survey information entity and the entity recognition information, then screen out the target matching entity from each entity recognition information based on the matching degree, and finally determine the entity criticality of the entity recognition information based on the target matching entity and the attention score of the survey information entity, so as to quantify the entity that exists both in the survey information entity and in the entity recognition information and determine the criticality of the entity.

[0096] In step S401 of some embodiments, the matching degree is a numerical value, which is represented in the form of a decimal between 0 and 1 or in percentage form, and is used to measure the similarity between the exploration information entity and the entity recognition information. The greater the matching degree, the greater the similarity. The method of determining the exploration information entity and the entity recognition information can be to vectorize the exploration information entity and the entity recognition information respectively, and then calculate the cosine similarity between the vector of the exploration information entity and the vector of the entity recognition information. It can also be a weighted sum according to the similarity value between the exploration information entity and the entity recognition information according to the preset weight. For example: set a weight of 0.5 for the entity category, a weight of 0 for the location coordinates, and a weight of 0.5 for the attribute information. At the same time, set corresponding weights for each sub-attribute information within the attribute information. For example, for an entity of a car, the entity category is a car, the attribute information includes the license plate number and the vehicle brand, the weight of the license plate number is 0.25, and the weight of the vehicle brand is 0.25. The entity category of the exploration information entity is a car, the vehicle brand is A, and the license plate is B. The entity category of the entity recognition information is a car, the brand is A, and the license plate is C, then the similarity is 0.5 + 0.25 + 0 = 0.75.

[0097] In step S402 of some embodiments, the target matching entity is the entity in the entity recognition information that is similar to the exploration information entity. For example, on the street, the entity recognition information identifies 20 people, and the exploration information entity includes three people, namely Grandma Zhang, Uncle Li, and Aunt Wang. Then, based on these three people in the exploration information entity, it is confirmed which of these 20 people on the street are the people who appear in the exploration information entity, that is, which person is Grandma Zhang, which person is Uncle Li, and which person is Aunt Wang among these 20 people. The screening method can be to obtain the top N entities with the largest matching degree, or to compare with a predetermined threshold.

[0098] The reason for screening the target matching entity from the exploration entity information based on the matching degree is as follows: The entities that appear in the exploration entity information are identified from the on-site exploration information. Therefore, the exploration information entity may be an important entity in the claim judgment. The entity recognition information is from the video, and not all of the multiple entities in the video are important entities. Therefore, through the matching degree, the entities that match the entity recognition information are screened from each entity recognition information based on the matching degree, so as to determine which entities in the on-site exploration video are important and determine the areas that cannot be covered by subsequent watermarks.

[0099] In step S403 of some embodiments, the entity criticality is a numerical value representing the criticality of entity recognition information. The numerical value is a decimal number, and the larger the value, the more critical the entity recognition information is. Based on the attention scores of the target matching entity and the survey information entity, the entity criticality of the entity recognition information can be determined in the following manner: When the target matching entity and the entity recognition information are the same entity, the attention score of the survey information entity is used as the entity criticality of the entity recognition information. When the target matching entity and the entity recognition information are not the same entity, the entity criticality is 0.

[0100] For example, the target matching entities include Grandma Zhang and Aunt Wang. In the survey information entity, the attention score of Grandma Zhang is 1.3, and the attention score of Aunt Wang is 3.4. The entity recognition information identifies 20 people, among which the third person is Grandma Zhang and the sixth person is Aunt Wang. Then the entity criticality of the third person is 1.3, the entity criticality of Aunt Wang is 3.4, and the entity criticality of the remaining people is 0.

[0101] In step S203 of some embodiments, the target detection information is a structured data formed by adding a label to an entity on the basis of the entity. As described above, the entity is a structured data including entity category, position coordinates, attribute information, etc. Then the target detection information is entity category, position coordinates, label, attribute information, etc. The target detection information is used to indicate which entities in the on-site survey video are important and cannot be covered by watermarks. When the target detection information is "1", it means it cannot be covered. When the target detection information is "0", it means it can be covered. Based on the entity criticality, the entity recognition information is screened to obtain the target detection information, which can be achieved by screening the top N entities or by comparing with a preset threshold.

[0102] In step S104 of some embodiments, the watermark area information is a set of pixel coordinates, indicating the pixel coordinate area that can be used to add watermarks. Based on each target detection information, the watermark area information of each key frame video block can be determined in the following manner: For a key frame video block, according to the target detection information corresponding to the key frame video block, obtain the target detection information with the label "1", and then based on the position coordinates in the target detection information, determine the smallest box that can enclose the target detection information. All the pixel coordinates within this smallest box are used as the non-watermarkable area, and all areas of the video frame except the non-watermarkable area are used as the coverable area. The coverable area is used as the watermark area information. Repeat the process described above until all key frame video blocks are processed.

[0103] Please refer to Figure 5 , in some embodiments, step S105 includes but is not limited to steps S501 to S504:

[0104] Step S501: Split the on-site survey information to obtain at least one sub-survey information.

[0105] Step S502: Based on the claim type information, sort each sub-survey information to obtain a survey information sequence.

[0106] Step S503: Obtain the region frames of each watermark region information, and based on each region frame and the survey information sequence, screen each sub-survey information to determine the target survey information of each watermark region information.

[0107] Step S504: Based on the on-site survey information and the target survey information of each watermark region information, add watermarks to each key-frame video block to obtain a target watermark video.

[0108] Steps S501 to S504 shown in the embodiments of the present application, by splitting the on-site survey information, refine the complex survey data into multiple sub-survey information, and then, based on the claim type information, sort each sub-survey information to form a survey information sequence, so as to determine the priority of each sub-survey information as a watermark. Further, obtain the region frames of the watermark region information, and screen the sub-survey information in combination with the survey information sequence to ensure that when generating the watermark subsequently, the watermark is important and does not exceed the watermark region. Finally, based on the on-site survey information and the target survey information of each watermark region information, add watermarks to each key-frame video block to obtain a target watermark video, realizing reducing the interference of the watermark on the key information when adding the watermark.

[0109] In step S501 of some embodiments, the sub-survey information is the sub-information in the on-site survey information, for example: information such as date, weather, location, claims adjuster, etc. The information splitting can split the on-site survey information according to a preset template, or can split the on-site survey information through a pre-trained natural language model.

[0110] In step S502 of some embodiments, the survey information sequence is a sequence obtained by sorting the sub-information according to the information importance of each sub-information. The information importance refers to which information is important for the sub-information under this claim type. The way to determine the information importance of the sub-information can be set in advance, or can be dynamically calculated by a data-driven method. For example, based on the data of historical claim cases, statistically analyze the influence degree of each sub-information on the final claim decision under different claim types, and use machine learning or statistical analysis methods to calculate the information importance.

[0111] The reason for sorting each sub-survey information based on the claim type information is that different claim types have different focuses on survey information, and sorting can ensure that the most critical information will be displayed in the watermark area. For example, in the claim for a vehicle collision accident, the damaged parts of the vehicle, the information about the responsible party for the accident, and the traffic signal conditions may be the most important.

[0112] In step S503 of some embodiments, the region box refers to the effective region in the watermark region information that can be used to generate a watermark. Since the watermark region information is the pixel coordinate data of some specific regions in the image, these pixel coordinates may present irregular shapes and cannot be directly used to add a watermark. Therefore, the watermark region is divided by multiple rectangles, and the irregular watermark region is approximately represented as several regular rectangular regions as the region box. Specifically, for a key-frame video block, first, according to the pixel coordinate distribution of the watermark region, the watermark region is divided into multiple smaller rectangular units using a bounding box or a minimum bounding rectangle. Repeat the above description until the region box of the watermark region information of each key-frame video block is determined.

[0113] The target survey information is a watermark that can be displayed within the region box, and the source of the watermark is generated from some sub-survey information. The way to determine the target survey information can be achieved through the following methods.

[0114] Generate its corresponding watermark based on each sub-survey information;

[0115] Compare the watermark corresponding to the sub-survey information with the region box according to the order of the sub-survey information in the sequence. If the watermark corresponding to the sub-survey information is smaller than the region box, then take this sub-survey information as the target survey information, and subtract the watermark region corresponding to this sub-survey information from the region box, and use the obtained region to update the region box; if the watermark corresponding to the sub-survey information is larger than the region box, then skip this sub-survey information;

[0116] Obtain the next sub-survey information according to the survey information sequence until there is no sub-survey information in the survey information sequence or the region box is smaller than the preset threshold.

[0117] Please refer to Figure 6 , in some embodiments, step S504 includes but is not limited to steps S601 to S604:

[0118] Step S601, for the watermark region information of each key-frame video block, generate explicit watermark data based on the target survey information corresponding to the watermark region information;

[0119] Step S602, for each key-frame video block, add a watermark to each key-frame video block based on the explicit watermark data to obtain an explicit watermark video frame;

[0120] Step S603: Perform video frame stitching on each explicit watermark video frame to obtain a preliminary watermark video.

[0121] Step S604: Add an implicit watermark to the preliminary watermark video based on the on-site investigation information to obtain the target watermark video.

[0122] Steps S601 to S604 shown in the embodiments of the present application generate explicit watermark data based on the target investigation information corresponding to the watermark area information. Then, use the generated explicit watermark data to add a watermark to the key frame video block to obtain an explicit watermark video frame, which reduces the interference of the watermark on key information when adding the watermark. Further, by performing video frame stitching on each explicit watermark video frame, a preliminary watermark video is obtained, and then an implicit watermark is further added to the preliminary watermark video based on the on-site investigation information to achieve deep protection of the video content, so that even if the explicit watermark is tampered with or removed, the integrity and authenticity of the video can still be verified through the implicit watermark. Through the dual protection mechanism of combining explicit watermark and implicit watermark, it reduces the interference of the watermark on key information when adding the watermark, and at the same time has strong security and anti-tampering capabilities.

[0123] In step S601 of some embodiments, the explicit watermark data refers to directly visible watermark information, and the generation of the display watermark data can be achieved by using image processing technology to generate a semi-transparent watermark layer.

[0124] In step S602 of some embodiments, the explicit watermark video frame is a video block with a display watermark added to the key frame video block. The explicit watermark video frame can be obtained by covering the generated watermark data within the specified watermark area.

[0125] In step S603 of some embodiments, the preliminary watermark video is a video that has an additional semi-transparent watermark layer compared to the on-site investigation video. The preliminary watermark video can be generated in the following way. First, obtain the order of the key frame video blocks in the on-site investigation video, and in accordance with this order, perform video frame stitching on each explicit watermark video frame to obtain the preliminary watermark video.

[0126] Please refer to Figure 7 , in some embodiments, step S604 may include but is not limited to steps S701 to S703:

[0127] Step S701: Perform frequency domain conversion on the preliminary watermark video to obtain a frequency domain watermark video.

[0128] Step S702: Add an implicit watermark to the frequency domain watermark video based on the on-site investigation information to obtain a preliminary frequency domain video.

[0129] Step S703: Perform inverse frequency-domain transformation on the preliminary frequency-domain video to obtain the target watermarked video.

[0130] Steps S701 to S703 illustrated in the embodiments of the present application, through frequency-domain transformation of the preliminary watermarked video, convert the video data from the time domain to the frequency domain. Then, based on the on-site investigation information, add an implicit watermark to the frequency-domain watermarked video, perform inverse frequency-domain transformation on the preliminary frequency-domain video, and restore it to the time-domain video, so that the target watermarked video not only retains the information security characteristics of the implicit watermark but also can maintain the same visual quality as the original video. This enables the watermark to provide higher security and anti-tampering capabilities without affecting the visibility of the video, while ensuring the authenticity of the claim settlement video. Even if it undergoes common video editing operations, the integrity of the video can still be verified through the watermark, thereby enhancing the reliability and security of claim settlement review.

[0131] In step S701 of some embodiments, the frequency-domain watermarked video refers to the frequency-domain components obtained after the video frame data is mapped to the frequency-domain space through frequency-domain transformation. The method of frequency-domain transformation can be discrete Fourier transform or discrete cosine transform, etc.

[0132] In step S702 of some embodiments, the preliminary frequency-domain video refers to the frequency-domain components with an implicit watermark embedded on the basis of the frequency-domain watermarked video. The method of adding the implicit watermark can be frequency-domain coefficient modulation or phase modulation, or other known methods of adding implicit watermarks in the art.

[0133] The reason for adding an implicit watermark in addition to the explicit watermark is that the explicit watermark performs necessary screening and streamlining on the on-site investigation information during superposition, and may not be able to completely cover all key information. To ensure that all on-site investigation information can be completely retained and verified and restored when needed, the method of using an implicit watermark is adopted to embed all on-site investigation information into the video, thereby enhancing the integrity and security of the information without affecting the readability of the video.

[0134] In step S703 of some embodiments, the target watermarked video refers to the final video that contains both the explicit watermark and the embedded implicit watermark and is stored and displayed in the time domain after inverse frequency-domain transformation. Performing inverse frequency-domain transformation on the preliminary frequency-domain video can be achieved through inverse discrete Fourier transform, inverse discrete cosine transform, or other known inverse transformation methods in the art.

[0135] It should be noted that during the process of frequency-domain transformation and inverse frequency-domain transformation, the methods used are corresponding. For example, if discrete Fourier transform is used during frequency-domain transformation, then inverse discrete Fourier transform is used during inverse frequency-domain transformation.

[0136] Please refer to Figure 8, embodiments of the present application further provide a video watermark adding device, which can implement the above video watermark adding method. The device includes:

[0137] A data acquisition module 801, configured to acquire a claim report, where the claim report includes claim type information, on-site investigation information, and on-site investigation video;

[0138] A video segmentation module 802, configured to segment the on-site investigation video to obtain at least one key frame video block;

[0139] An information monitoring module 803, configured to perform key information detection on each key frame video block based on the on-site investigation information to obtain target detection information of each key frame video block;

[0140] A region generation module 804, configured to generate a watermark region for the on-site investigation information based on each target detection information to obtain watermark region information of each key frame video block;

[0141] A watermark adding module 805, configured to add a watermark to each key frame video block based on the claim type information, the on-site investigation information, and the watermark region information of each key frame video block to obtain a target watermark video.

[0142] The specific implementation manner of this video watermark adding device is basically the same as that of the above video watermark adding method, and will not be elaborated here.

[0143] Embodiments of the present application further provide an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the above video watermark adding method. The electronic device can be any intelligent terminal including a tablet computer, an in-vehicle computer, etc.

[0144] Please refer to Figure 9 , Figure 9 illustrates the hardware structure of an electronic device according to another embodiment. The electronic device includes:

[0145] A processor 901, which can be implemented in a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, etc., and is configured to execute relevant programs to implement the technical solutions provided by the embodiments of the present application;

[0146] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM), etc. The memory 902 can store an operating system and other application programs. When implementing the technical solutions provided in the embodiments of this specification through software or firmware, the relevant program codes are stored in the memory 902 and are called by the processor 901 to execute the video watermark addition method of the embodiments of this application;

[0147] The input / output interface 903 is used to implement information input and output;

[0148] The communication interface 904 is used to implement communication and interaction between this device and other devices. It can achieve communication through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.);

[0149] The bus 905 transmits information between the various components of the device (such as the processor 901, the memory 902, the input / output interface 903, and the communication interface 904);

[0150] Among them, the processor 901, the memory 902, the input / output interface 903, and the communication interface 904 are communicatively connected to each other inside the device through the bus 905.

[0151] The embodiments of this application also provide a computer-readable storage medium. The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, it implements the above-mentioned video watermark addition method.

[0152] As a non-transitory computer-readable storage medium, the memory can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include high-speed random access memory, and can also include non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state storage devices. In some embodiments, the memory optionally includes a memory remotely provided with respect to the processor, and these remote memories can be connected to the processor through a network. Examples of the above-mentioned network include but are not limited to the Internet, an enterprise intranet, a local area network, a mobile communication network, and combinations thereof.

[0153] The video watermark addition method, video watermark addition device, electronic device, and storage medium provided by the embodiments of the present application obtain claim type information, on-site investigation information, and on-site investigation videos, and perform video segmentation on the on-site investigation videos, thereby processing the videos in a refined manner and improving the accuracy of adding watermarks to the videos. Further, key information detection is performed using the on-site investigation information, so that the target detection information accurately corresponds to each key frame video block. Then, based on the target detection information, video watermark regions for each key frame are generated, avoiding important content in the video frames and preventing the watermark information from blocking key claim content. Finally, watermark addition is achieved by combining the claim type information, on-site investigation information, and watermark region information, reducing the possibility of the watermark interfering with claim judgment and realizing the reduction of watermark interference with key information when adding watermarks.

[0154] The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art know that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of the present application are equally applicable to similar technical problems.

[0155] Those skilled in the art can understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than those shown in the figures, or combine certain steps, or different steps.

[0156] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, that is, they may be located in one place, or may be distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0157] Those of ordinary skill in the art can understand that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, and their appropriate combinations.

[0158] In the description of this application and the above-mentioned accompanying drawings, terms such as "first", "second", "third", "fourth", etc. (if any) are used to distinguish similar objects and do not necessarily describe a specific order or sequence. It should be understood that the data used in this way can be interchanged under appropriate circumstances so that the embodiments of this application described here can be implemented in an order different from those illustrated or described here. In addition, the terms "comprising" and "having" and any variations thereof are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or device that comprises a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but may include other steps or units not clearly listed or inherent to these processes, methods, products, or devices.

[0159] It should be understood that in this application, "at least one (item)" means one or more, and "a plurality" means two or more. "And / or" is used to describe the association relationship of associated objects and indicates that three relationships may exist. For example, "A and / or B" may mean: only A exists, only B exists, and both A and B exist at the same time. Among them, A and B can be singular or plural. The character " / " generally means that the associated objects before and after are in an "or" relationship. "At least one (one) of the following" or its similar expression means any combination of these items, including any combination of single item (one) or plural items (ones). For example, at least one (one) of a, b, or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0160] In several embodiments provided by this application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the above-mentioned division of units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed coupling or direct coupling or communication connection to each other can be through some interfaces. The indirect coupling or communication connection of devices or units can be in an electrical, mechanical, or other form.

[0161] The units described above as separate components may or may not be physically separated. The components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0162] In addition, in each embodiment of the present application, each functional unit can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0163] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods in each embodiment of the present application. The foregoing storage medium includes: various media that can store programs, such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs.

[0164] The preferred embodiments of the embodiments of the present application have been described above with reference to the accompanying drawings. However, this does not limit the scope of the rights of the embodiments of the present application. Any modifications, equivalent replacements, and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall fall within the scope of the rights of the embodiments of the present application.

Claims

1. A method for adding video watermark, characterized in that, The method includes: Obtaining a claim settlement report, where the claim settlement report includes claim settlement type information, on-site inspection information, and on-site inspection videos; Performing video segmentation on the on-site inspection videos to obtain at least one key-frame video block; Based on the on-site inspection information, performing key information detection on each of the key-frame video blocks to obtain the target detection information of each of the key-frame video blocks; Based on each of the target detection information, generating a watermark area for the on-site inspection information to obtain the watermark area information of each of the key-frame video blocks; Based on the claim settlement type information, the on-site inspection information, and the watermark area information of each of the key-frame video blocks, adding a watermark to each of the key-frame video blocks to obtain a target watermark video.

2. The method according to claim 1, wherein The performing key information detection on each of the key-frame video blocks based on the on-site inspection information to obtain the target detection information of each of the key-frame video blocks includes: For each of the key-frame video blocks, performing entity recognition on the key-frame video block to obtain entity recognition information; For each of the key-frame video blocks, calculating the entity key degree of the entity recognition information based on the on-site inspection information and the entity recognition information; Based on the entity key degree, screening the entity recognition information to obtain the target detection information.

3. The method according to claim 2, wherein The calculating the entity key degree of the entity recognition information based on the on-site inspection information and the entity recognition information includes: Performing entity recognition on the on-site inspection information to obtain at least one inspection information entity; Generating an attention score for each of the inspection information entities based on the on-site inspection information through an attention model; Based on the attention scores of each of the inspection information entities, determining the entity key degree of the entity recognition information.

4. The method according to claim 3, wherein The determining the entity key degree of the entity recognition information based on the attention scores of each of the inspection information entities includes: For each of the inspection information entities, confirming the matching degree between the inspection information entity and the entity recognition information based on the inspection information entity and the entity recognition information; For each of the matching degrees, screening out target matching entities from each of the entity recognition information based on the matching degree; Based on the target matching entities and the attention scores of the inspection information entities, determining the entity key degree of the entity recognition information.

5. The method according to claim 1, characterized in that, The adding a watermark to each of the key-frame video blocks based on the claim settlement type information, the on-site inspection information, and the watermark area information of each of the key-frame video blocks to obtain a target watermark video includes: Performing information splitting on the on-site inspection information to obtain at least one sub-inspection information; Based on the claim settlement type information, sorting each of the sub-inspection information to obtain an inspection information sequence; Obtaining the region frames of each of the watermark area information, and based on each of the region frames and the inspection information sequence, screening each of the sub-inspection information to determine the target inspection information of each of the watermark area information; Based on the target survey information of the on-site survey information and each watermark area information, watermarking is added to each key-frame video block to obtain the target watermark video.

6. The method according to claim 5, wherein The adding watermark to each key-frame video block based on the target survey information of the on-site survey information and each watermark area information to obtain the target watermark video includes: For the watermark area information of each key-frame video block, explicit watermark data is generated based on the target survey information corresponding to the watermark area information; For each key-frame video block, watermarking is added to each key-frame video block based on the explicit watermark data to obtain an explicit watermark video frame; Video frame splicing is performed on each explicit watermark video frame to obtain a preliminary watermark video; Based on the on-site survey information, implicit watermarking is added to the preliminary watermark video to obtain the target watermark video.

7. The method according to claim 6, characterized in that, The adding implicit watermark to the preliminary watermark video based on the on-site survey information to obtain the target watermark video includes: Performing frequency domain conversion on the preliminary watermark video to obtain a frequency domain watermark video; Based on the on-site survey information, implicit watermarking is added to the frequency domain watermark video to obtain a preliminary frequency domain video; Performing inverse frequency domain conversion on the preliminary frequency domain video to obtain the target watermark video.

8. A video watermark adding device, characterized in that, The device includes: A data acquisition module for acquiring a claim report, where the claim report includes claim type information, on-site survey information, and an on-site survey video; A video segmentation module for segmenting the on-site survey video to obtain at least one key-frame video block; An information monitoring module for performing key information detection on each key-frame video block based on the on-site survey information to obtain the target detection information of each key-frame video block; A region generation module for generating a watermark area for the on-site survey information based on each target detection information to obtain the watermark area information of each key-frame video block; A watermark addition module for adding watermark to each key-frame video block based on the claim type information, the on-site survey information, and the watermark area information of each key-frame video block to obtain a target watermark video.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the video watermark addition method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the video watermark addition method according to any one of claims 1 to 7.

Images