Dynamic video processing method and device, equipment, medium and program

Abstract

The embodiment of the invention relates to the technical field of data processing, and discloses a dynamic video processing method and device, equipment, a medium and a program. The method comprises the following steps: acquiring a dynamic video, first interaction data and second interaction data; scoring the dynamic video according to the dynamic video and the first interaction data; scoring the uploading user according to the second interaction data; according to the score of the dynamic video and the score of the uploading user, a conversion strategy and an editing strategy are generated respectively, the conversion strategy comprises information indicating whether to convert the dynamic video or not, and the editing strategy comprises information indicating whether to add interaction elements in the dynamic video or not; and processing the dynamic video according to the conversion strategy and / or the editing strategy. A user can be helped to accurately mine the value of the dynamic video and realize effective conversion.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and in particular to a method, apparatus, device, medium, and program for processing dynamic video. Background Technology

[0002] Dynamic video is a content format distinct from both still images and video. Compared to still images, it can convey richer information, while reducing the creative burden compared to video. Therefore, dynamic video has gained widespread popularity among creators and viewers on user-generated content (UGC) platforms. Summary of the Invention

[0003] This application provides a method, apparatus, device, medium, and program for processing dynamic videos to address the problem of how to help users accurately extract the value of dynamic videos and achieve effective conversion.

[0004] According to some embodiments of this application, a method for processing dynamic video is provided, comprising: acquiring dynamic video, first interactive data, and second interactive data, wherein the dynamic video is uploaded by an uploading user, the first interactive data includes data generated by viewers interacting with the dynamic video, and the second interactive data includes data generated by the uploading user interacting with the video; rating the dynamic video based on the dynamic video and the first interactive data; rating the uploading user based on the second interactive data; generating a conversion strategy and an editing strategy based on the rating of the dynamic video and the rating of the uploading user, wherein the conversion strategy includes information indicating whether to convert the dynamic video, and the editing strategy includes information indicating whether to add interactive elements to the dynamic video; and processing the dynamic video according to the conversion strategy and / or the editing strategy.

[0005] According to some embodiments of this application, a dynamic video processing apparatus is also provided, comprising: an acquisition module for acquiring a dynamic video, first interactive data, and second interactive data, wherein the dynamic video is uploaded by an uploading user, the first interactive data includes data generated by a viewing user interacting with the dynamic video, and the second interactive data includes data generated by the uploading user interacting with the dynamic video; a first scoring module for scoring the dynamic video based on the dynamic video and the first interactive data; a second scoring module for scoring the uploading user based on the second interactive data; a strategy generation module for generating a conversion strategy and an editing strategy based on the scoring of the dynamic video and the scoring of the uploading user, wherein the conversion strategy includes information indicating whether to convert the dynamic video, and the editing strategy includes information indicating whether to add interactive elements to the dynamic video; and a video processing module for processing the dynamic video according to the conversion strategy and / or the editing strategy.

[0006] According to some embodiments of this application, an electronic device is also provided, including: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions, when executed by the at least one processor, can implement the dynamic video processing method described in the embodiments of this application.

[0007] According to some embodiments of this application, a computer-readable storage medium is also provided in this application embodiment. The computer-readable storage medium stores computer instructions, which, when executed by a processor, can implement the dynamic video processing method described in the embodiments of this application.

[0008] According to some embodiments of this application, a computer program product is also provided, including a computer program that, when at least a portion of the computer program is executed by a processor, can implement the dynamic video processing method described in the embodiments of this application.

[0009] The technical solution provided in this application not only quantifies the content value of dynamic videos by scoring them using the acquired dynamic videos and first interactive data, but also quantifies the added value brought by factors other than content by scoring the uploading users using the acquired second interactive data. This allows for multi-dimensional and accurate mining of the value of dynamic videos, and guides the processing of dynamic videos with conversion and editing strategies generated based on the scores of the dynamic videos and the uploading users, achieving effective conversion.

[0010] In some embodiments, generating a conversion strategy and an editing strategy based on the rating of the dynamic video and the rating of the uploading user, respectively, includes: weighting the rating of the dynamic video and the rating of the uploading user according to a first weight matrix to generate a first weighted rating of the dynamic video and a first weighted rating of the uploading user; generating the conversion strategy based on the first weighted rating and the second weighted rating; weighting the rating of the dynamic video and the rating of the uploading user according to a second weight matrix to generate a third weighted rating of the dynamic video and a fourth weighted rating of the uploading user; and generating the editing strategy based on the third weighted rating and the fourth weighted rating. This approach is simple and efficient.

[0011] In some embodiments, the weighting of the dynamic video rating and the uploader's rating based on a first weight matrix includes: obtaining the first weight matrix; updating the first weight matrix based on the dynamic video's status data and / or the uploader's protection coefficient, wherein the status data includes one or a combination of the following: publication duration and completion rate, the publication duration being the difference between the current time and the time when the uploader uploaded the dynamic video, and the protection coefficient being determined based on one or a combination of the following data: user level, creation duration, number of works, and creation frequency, the creation duration being the difference between the current time and the time when the uploader uploaded the first piece of content; and weighting the dynamic video rating and the uploader's rating based on the updated first weight matrix. By combining the current dynamic video and the uploader, the first weight matrix is ​​processed so that it varies with different dynamic videos and uploaders, thereby allowing the first weight matrix to better align with the performance of the current dynamic video and the uploader, resulting in more accurate conversion strategy generation and better conversion effects.

[0012] In some embodiments, before determining whether the score generated based on the third weighted score and the fourth weighted score exceeds a threshold, the method further includes: obtaining an initial threshold; generating a correction factor based on the content type of the dynamic video; and updating the initial threshold based on the correction factor to generate the threshold. By providing a correction factor associated with the content type of the dynamic video, the threshold is dynamically adjusted to make it more accurate, thereby making the editing strategy generated based on the threshold more accurate, and thus enabling more effective value conversion of the dynamic video.

[0013] In some embodiments, updating the initial threshold according to the correction factor includes: determining a content scarcity parameter for the dynamic video, wherein the content scarcity parameter describes the degree of scarcity of the content in the dynamic video; and updating the initial threshold according to the correction factor and the content scarcity parameter. Further introducing a parameter related to the degree of content scarcity allows for a more comprehensive adjustment of the threshold, particularly prioritizing the conversion of scarce content, which is beneficial for enriching the platform's content and attracting users. Attached Figure Description

[0014] One or more embodiments are illustrated by way of example with reference numerals in the accompanying drawings. These illustrations do not constitute a limitation on the embodiments. Elements with the same reference numerals in the drawings are denoted as similar elements. Unless otherwise stated, the figures in the drawings are not to be limited by scale.

[0015] Figure 1 This is a flowchart of a dynamic video processing method provided in one embodiment of this application; Figure 2 This is a schematic diagram of the structure of a dynamic video processing device provided in another embodiment of this application; Figure 3 This is a schematic diagram of the structure of an electronic device provided in another embodiment of this application. Detailed Implementation

[0016] Analysis revealed that while many UGC platforms currently support both animated and regular video submissions, there are isolations between their upload interfaces. For example, some platforms provide different upload buttons for animated and regular videos, while others use different receiving and storage architectures. This results in significant conversion limitations between animated and regular videos on UGC platforms, requiring users to manually re-upload animated videos using the regular video upload interface. Furthermore, some UGC platforms have time limits for maintaining animated videos, automatically archiving them after a certain period. This means that users may not realize the value of animated videos while they are still active, and by the time they realize their value and decide to convert and re-upload them as regular videos, the animated videos have become invalid, and their value has been lost. Additionally, some UGC platforms allow different interactive features for animated and regular videos, requiring users to manually edit animated videos to add interactive features, or even forgo such features altogether, further hindering the conversion of animated video value.

[0017] Based on this, the paper addresses the question of how to help users accurately extract the value of dynamic videos and achieve effective conversion. Furthermore, it provides a method, apparatus, device, medium, and program for processing dynamic videos. By scoring the dynamic video from two dimensions—the video itself and the uploading user—the value of the dynamic video is fully extracted and evaluated. Conversion and editing strategies generated based on the video's score and the uploading user's score guide the processing of the dynamic video, thereby achieving effective conversion of its value and solving the aforementioned problem.

[0018] It is understood that in the specific implementation of this application, the collection, use or processing of data is involved. When the embodiments of this application are applied to specific products or technologies, the permission or consent of the data subject is required. The collection, use or processing of related data must strictly comply with the relevant laws, regulations and standards of the data source country, implementation country and other relevant countries and regions. The de-identification technology ensures that the final data used is the de-identified data that has been securely processed, thereby protecting the rights and interests of the data subject and data security.

[0019] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the various embodiments of this application will be described in detail below with reference to the accompanying drawings. However, those skilled in the art will understand that many technical details have been presented in the various embodiments of this application to enable readers to better understand this application. However, the technical solutions claimed in this application can be implemented even without these technical details and various changes and modifications based on the following embodiments.

[0020] The division of the following embodiments is for ease of description and should not constitute any limitation on the specific implementation of this application. The various embodiments can be combined with and referenced by each other without contradiction.

[0021] This application provides a method for processing dynamic video, applicable to devices such as computers and servers. The following will be combined with... Figure 1 The flowchart shown illustrates the processing method for dynamic video.

[0022] In some embodiments, such as Figure 1 As shown, the process of processing dynamic video includes the following steps: In step 101, dynamic video, first interactive data, and second interactive data are obtained. The dynamic video is uploaded by the uploading user, the first interactive data includes data generated by the viewing user based on the dynamic video platform, and the second interactive data includes data generated by the uploading user's interaction.

[0023] Step 102: Score the dynamic video based on the dynamic video and the first interactive data.

[0024] Step 103: Rate the uploading user based on the second interaction data.

[0025] Step 104: Based on the rating of the dynamic video and the rating of the uploading user, generate a conversion strategy and an editing strategy respectively. The conversion strategy includes information indicating whether to convert the dynamic video, and the editing strategy includes information indicating whether to add interactive elements to the dynamic video.

[0026] Step 105: Process the dynamic video according to the conversion and editing strategies.

[0027] exist Figure 1 In the illustrated embodiment, not only is the dynamic video scored using the acquired dynamic video and the first interactive data, thus quantifying the content value of the dynamic video, but the uploading user is also scored using the acquired second interactive data, thus quantifying the added value brought by factors other than the content. This allows for multi-dimensional and comprehensive exploration of the value of the dynamic video, and the conversion and editing strategies generated based on the dynamic video scores and the uploading user scores guide the processing of the dynamic video, achieving effective conversion.

[0028] It should be noted that, Figure 1 This is merely an example of the execution order of steps 101-105. In some embodiments, steps 102 and 103 can be executed concurrently, or step 103 can be executed before step 103, etc., which will not be listed here.

[0029] For ease of understanding, steps 101-105 will be explained below.

[0030] In step 101, the dynamic video can be any dynamic video. For example, it can be a dynamic video specified by the staff of the UGC platform or automatically selected by the platform according to certain rules. Alternatively, it can be a dynamic video specified by the uploading user, etc., which will not be listed here.

[0031] It should be noted that the embodiments of this application do not limit the data content of the first interactive data. It is understood that viewers can interact in various ways, such as uploading, playing, sharing, commenting, sending bullet comments, liking, and complaining. All of these interactions generate data, and all of this data is obtainable. However, based on different dynamic video rating processes and user rating processes, some of the aforementioned data may not be used subsequently. Therefore, when obtaining the first and second interactive data, this unused data may not be required; that is, the necessary data can be flexibly selected according to requirements.

[0032] It should also be noted that the same user can be both a viewer and an uploader. For example, if user 1 uploaded video 1 and watched video 2, user 1 is the uploader for video 1 and the viewer for video 2. This will not be elaborated further here.

[0033] In step 102, any dimension of the dynamic video associated with the dynamic video and the first interactive data can be scored. In some embodiments, the scoring of the dynamic video may include one or a combination of the following: semantic scoring, content scoring, and interaction scoring, wherein the semantic scoring describes the viewer's emotional tendency towards the dynamic video, the content scoring describes the image quality of the dynamic video, and the interaction scoring describes the viewer's interaction with the dynamic video. The following mainly describes the implementation process of scoring interactive videos in conjunction with the scoring provided in the above examples.

[0034] In some embodiments, semantic scores can be generated using the following expression: S = l × ρ; Where S is the semantic score, l is the proportion of positive keywords in the interactive text generated based on the first interactive data, and ρ is the popularity index of the dynamic video generated based on the first interactive data.

[0035] In some embodiments, content scores can be generated by using an image aesthetic evaluation model to perform aesthetic evaluation on keyframes in a dynamic video.

[0036] In some embodiments, the interaction rating can be generated using the following expression: E = ∑ (Ei × yi); Where E is the interaction score, Ei is the i-th interaction indicator generated based on the first interaction data, and yi is the preset coefficient corresponding to the i-th interaction indicator. The interaction indicators include at least one of the following indicators: like rate, share rate, completion rate, forwarding rate, comment rate, and click rate.

[0037] Of course, the above are just some examples of scoring processes for semantic scoring, content scoring, and interaction scoring for ease of understanding. In some embodiments, other methods can be used to score dynamic videos. For example, the keyframes of dynamic videos can be processed based on the Structure Similarity Index Measure (SSIM) algorithm to generate content scores, or the proportion of keywords with sentiment can be used as semantic scores, etc. These will not be listed here.

[0038] In step 103, the uploading user can be scored on any dimension associated with the second interaction data. In some embodiments, the uploading user is scored based on the second interaction data using the following expression: C = b × A; Where C is the rating of the uploading user, b is the average number of views of the content uploaded by the uploading user generated based on the second interaction data, and A is the fan activity level, which describes the interaction between the uploading user's fans and the uploading user.

[0039] Of course, the above are just examples. In some embodiments, the popularity of the uploading user's account can also be determined based on the second interaction data, which can be used as a rating for the uploading user, etc., which will not be listed here.

[0040] In step 104, in this embodiment, the conversion of dynamic video refers to converting it from dynamic video to ordinary video, or other video formats provided by the UGC platform. In some cases, the UGC platform may provide different video format requirements for uploaded dynamic videos and other video formats; therefore, the conversion of dynamic video will involve adjusting the video format. In other cases, the UGC platform may provide different upload interfaces for uploaded dynamic videos and other video formats; therefore, the conversion of dynamic video will involve uploading through another upload interface, etc., which will not be elaborated here. In this way, the limitations described above, such as setting two upload buttons and using different storage structures, which prevent dynamic videos from being directly presented to the audience as other content, can be automatically, accurately, and reliably resolved through corresponding conversion strategies, without requiring the uploading user to manually perform conversion and uploading.

[0041] It should be noted that the embodiments of this application do not limit the generation method of the conversion strategy and editing strategy. In some embodiments, it can be implemented based on artificial intelligence, models, etc., for example, taking the rating of the dynamic video and the rating of the uploading user as input, and performing binary classification output on the questions of whether to convert the dynamic video and whether to add interactive elements to the dynamic video. In some embodiments, it can be judged by preset logical rules, etc.

[0042] It should also be noted that the embodiments of this application do not limit all interactive elements. In some embodiments, interactive elements may include consumption function components; in some embodiments, interactive elements may also include user evaluation / feedback function components, etc., which will not be listed here.

[0043] Furthermore, during the generation of conversion and editing strategies, preprocessing can be performed on the video's rating and the uploader's rating, including but not limited to data optimization and outlier removal. Taking weighted processing as an example, in some embodiments, conversion and editing strategies are generated based on the video's rating and the uploader's rating, respectively. This can be achieved as follows: The video's rating and the uploader's rating are weighted according to a first weight matrix to generate a first weighted rating for the video and the uploader; a conversion strategy is generated based on the first and second weighted ratings; the video's rating and the uploader's rating are weighted according to a second weight matrix to generate a third weighted rating for the video and a fourth weighted rating for the uploader; and an editing strategy is generated based on the third and fourth weighted ratings.

[0044] To facilitate understanding of the generation of the above strategy, further examples will be provided below, in conjunction with the weighted processing example above.

[0045] In some embodiments, weighting the rating of the dynamic video and the rating of the uploading user according to the first weight matrix can be achieved as follows: obtaining the first weight matrix; updating the first weight matrix based on the status data of the dynamic video and / or the protection coefficient of the uploading user, wherein the status data includes one or a combination of the following: publication duration and completion rate, the publication duration being the difference between the current moment and the moment the uploading user uploaded the dynamic video, and the protection coefficient including one or a combination of the following: user level, creation duration, number of works, and creation frequency, the creation duration being the difference between the current moment and the moment the uploading user uploaded the first piece of content; and weighting the rating of the dynamic video and the rating of the uploading user according to the updated first weight matrix. In other words, by combining the current dynamic video and the uploading user, the first weight matrix is ​​processed so that it varies with different dynamic videos and uploading users, thus allowing the first weight matrix to better align with the performance of the current dynamic video and the uploading user, resulting in more accurate conversion strategy generation and better conversion effects.

[0046] It should be noted that this application does not limit the method for determining the protection coefficient. For example, the current creation stage of the uploading user can be analyzed based on user level, creation duration, number of works, and creation frequency. Then, a corresponding protection coefficient can be matched based on the current creation stage of the uploading user. The creation stage can be divided in various ways, such as into a novice stage, a growth stage, and a transition stage. Alternatively, different protection coefficients can be set for different stages according to different strategies. Taking the creation stage as a novice stage, growth stage, and transition stage as an example, the protection coefficient is 0.3 for the novice stage, 0.2 for the growth stage, and 0.1 for the transition stage. This approach can improve the dynamic video rating of all uploading users, making it easier to support the conversion of their uploaded content to enrich their content and gain more attention. It also provides greater room for improvement in the dynamic video rating of uploading users whose abilities are not yet fully developed, helping these users grow. Of course, other values ​​or other methods of dividing the creation stage can also be set, which will not be listed here.

[0047] It should also be noted that the embodiments of this application do not limit the specific update strategy of the first weight matrix.

[0048] In some embodiments, updating the first weight matrix based on the status data of the dynamic video and / or the protection coefficient of the uploading user may include: updating the interaction weights in the first weight matrix according to the protection coefficient to generate an updated first weight matrix, wherein the interaction elements are elements in the first weight matrix used to weight the interaction-related scores (such as the aforementioned interaction scores) in the rating of the dynamic video. This supports local independent updates to the first weight matrix, making the update control of the first weight matrix more precise and accurate. The first weight matrix will be more closely aligned with the current dynamic video and the performance of the uploading user, resulting in more accurate conversion strategy generation and better conversion effects.

[0049] In some cases, the interaction weights in the first weight matrix are updated based on the protection coefficient, using the following expression: α' = α × (1 + x); Where α' is the interaction weight, α is the updated interaction weight, and x is the protection coefficient.

[0050] Of course, the above are just examples. In some embodiments, other expressions can be used as the update strategy for interaction weights, which will not be listed here.

[0051] In some embodiments, updating the first weight matrix based on the status data of the dynamic video and / or the protection coefficient of the uploading user may include updating the interaction weight based on the publication duration and a time decay function. This embodiment does not limit the time decay function; it can be set according to platform requirements, uploading user requirements, etc. This dynamically decays the interaction weight over time, ensuring that content with high past interaction volume does not interfere with current processing.

[0052] In some embodiments, updating the first weight matrix based on the status data of the dynamic video and / or the protection coefficient of the uploading user may include: updating the elements in the first weight matrix used to weight the ratings related to the emotional tendencies of the viewing user (such as the semantic rating mentioned above) in the rating of the dynamic video based on the completion rate, and / or updating the elements in the first weight matrix used to weight the ratings related to the content (such as the content rating mentioned above) in the rating of the dynamic video.

[0053] For example, if the completion rate remains low, the element in the first weight matrix that is used to weight the ratings related to the emotional tendencies of the viewers in the rating of dynamic videos can be reduced, while the element in the first weight matrix that is used to weight the ratings related to the content in the rating of dynamic videos can be increased.

[0054] Of course, the above are just examples. In some embodiments, other strategies can be used to adjust the first weight matrix. Different strategies can be used based on different types of UGC platforms and / or different content of dynamic videos. For example, when the UGC platform is a short video type and the content of the dynamic video is entertainment-related, the elements in the first weight matrix used to weight the interaction rating can be increased. When the UGC platform is a knowledge type, the elements in the first weight matrix used to weight the content rating can be increased, etc.

[0055] Furthermore, in some embodiments, the first weight matrix can be dynamically determined based on the type of the UGC platform, which includes one or a combination of the following: entertainment, knowledge, social, short video, etc., so that the first weight matrix does not need to be updated subsequently. For example, when the UGC platform is entertainment, a first weight matrix with relatively larger values ​​for the elements used to weight interactive ratings and the elements used to weight interactive ratings can be determined; when the UGC platform is knowledge, a first weight matrix with relatively larger values ​​for the elements used to weight content dynamic ratings and the elements used to weight the ratings of uploading users can be determined, and so on.

[0056] To facilitate understanding of strategy generation, examples of editing strategy generation will be provided below. The generation of conversion strategies can be done in a similar manner to that of editing strategies, and will not be elaborated upon further.

[0057] In some embodiments, generating an editing strategy based on a third weighted score and a fourth weighted score can be achieved by: determining whether the score generated based on the third weighted score and the fourth weighted score exceeds a threshold; if it exceeds the threshold, determining to add interactive elements to the dynamic video to generate a corresponding editing strategy; if it does not exceed the threshold, determining not to add interactive elements to the dynamic video to generate a corresponding editing strategy.

[0058] It should be noted that the embodiments of this application do not limit the threshold; it can be a fixed value or a dynamically changing value.

[0059] Taking dynamic value acquisition as an example, in some embodiments, before determining whether the score generated based on the third weighted score and the fourth weighted score exceeds the threshold, the method further includes: obtaining an initial threshold; generating a correction factor based on the content type of the dynamic video; and updating the initial threshold based on the correction factor to generate a new threshold. In other words, by providing a correction factor associated with the content type of the dynamic video, the threshold is dynamically adjusted to make it more accurate. This leads to a more accurate editing strategy based on the threshold, thereby enabling more effective value conversion of the dynamic video.

[0060] It should be noted that the embodiments of this application do not limit the method of generating the correction factor. In some embodiments, the correction factor is generated according to the content type of the dynamic video, which can be achieved as follows: matching is performed on a preset parameter set according to the content type of the dynamic video, and the matched preset parameter is used as the correction factor. The preset parameter set includes at least two preset parameters associated with different content types. For example, the preset parameter associated with the two-dimensional tag is set to 0.2, and the preset parameter associated with the knowledge tag is set to 0.1, etc. Alternatively, in some embodiments, the correction factor can be calculated based on a preset expression instead of tag matching, etc., which will not be listed here.

[0061] Of course, the content type of the dynamic video is only one factor to consider. In some embodiments, updating the initial threshold based on a correction factor can be achieved as follows: determine a content scarcity parameter for the dynamic video, where the content scarcity parameter describes the degree of scarcity of the dynamic video's content; update the initial threshold based on the correction factor and the content scarcity parameter. In other words, further introducing parameters related to content scarcity allows for a more comprehensive adjustment of the threshold, especially prioritizing the conversion of scarce content, which helps enrich the platform's content and attract users to watch.

[0062] It should be noted that this application does not limit the specific update strategy of the initial threshold. In some examples, the initial threshold is updated according to the correction factor and the content scarcity parameter, and is determined by the following expression: Vth′=Vth×(1-Kt×(1-R)); where Vth′ is the threshold, Vth is the initial threshold, Kt is the correction factor, and R is the content scarcity parameter. Of course, the above is only an example, and other specific strategies can be used in some examples, which will not be listed here.

[0063] It should also be noted that the embodiments of this application do not limit the generation method of the second weight matrix, whether updates are allowed, or the specific update method used when updates are allowed. In some embodiments, if the interactive element includes a consumption function component, the second weight matrix can be determined based on the consumption data so that the rating processed based on the second weight matrix can provide guidance from the perspective of consumption, so as to generate editing strategies more accurately and effectively. Of course, the second weight matrix can also be generated and updated in a similar way to the first weight matrix, etc., which will not be listed here.

[0064] Furthermore, this application does not limit whether the conversion strategy and editing strategy carry other information.

[0065] In some embodiments, the conversion strategy further includes scores generated based on a first weighted score and a second weighted score, and / or the editing strategy further includes scores generated based on a third weighted score and a fourth weighted score. This allows for the retention of score records, helping uploading users understand the basis for their processing, and also facilitates subsequent data analysis, model optimization, and problem tracing.

[0066] In some embodiments, the traffic and revenue generated by the converted dynamic video can also be predicted, wherein the traffic and revenue can be carried in the editing strategy (and / or conversion strategy) or recorded separately, wherein the traffic and revenue can be generated by: acquiring third-party interaction data and user preference data, wherein the third-party interaction data includes traffic and consumption data of content of the same type as the dynamic video, and the user preference data is data generated by viewing user interactions, or a knowledge graph describing viewing user preferences generated based on data generated by viewing user interactions; and predicting traffic and revenue based on the third-party interaction data and user preference data. Of course, the probability of traffic occurrence and the probability of revenue occurrence can also be carried in the prediction simultaneously.

[0067] Of course, the above is just one example. In some embodiments, other information can also be predicted or combined. For example, traffic, fan growth, and revenue can be predicted to present the following information to the uploading user through an interactive interface, facilitating the uploading user's decision on whether to accept the corresponding strategy: Conversion strategy: Recommended dynamic → normal (conversion score 86.7%), editing strategy: Suggest adding paid interactive elements (editing score 90.9%). Expected results: Increased traffic by 120,000 to 180,000 | Increased followers by 300 to 500 | Increased revenue by $8 to $350.

[0068] In step 106, this embodiment of the application does not exclude the combination of processing methods other than conversion strategies and editing strategies. Taking intelligent screen expansion as an example, a scene semantic segmentation map associated with the dynamic video can be obtained, and then the original vertical screen frames in the dynamic video are compared with the scene semantic segmentation map. Figure 1 The input is PGGAN (Progressive Generative Adversarial Network), which outputs landscape frames of a preset size (e.g., 1920×1080 pixels). Of course, this is just an example; other image processing methods can also be used, which will not be listed here.

[0069] It should be noted that the embodiments of this application do not limit the interactive elements. For example, they can be product mounting slot elements, payment elements, multi-page segmented prompt elements, etc.

[0070] For ease of understanding Figure 1 The illustrated embodiment will be provided as an example below. The following primarily illustrates the conversion of dynamic video through uploading via different interfaces.

[0071] In some embodiments, a processing request from the uploading user is first received to determine the dynamic video to be processed, and dynamic video data, first interaction data, and second interaction data are acquired. Then, simultaneously: the composition and / or sharpness of keyframes in the dynamic video data are quantified to obtain a content score Q, so that professionally shot videos receive a higher score, such as 0.9, while blurry, casually shot videos receive a lower score, such as 0.3; the text content in the first interaction data is acquired, and expressions indicating high semantic value, such as "learned something," are counted as positive keywords and their proportion in the text is multiplied by the popularity of the dynamic video to obtain a semantic score S; based on the first interaction data, the like rate, share rate, and completion rate of the dynamic video are calculated and weighted summed to obtain... Interaction score E, where the like rate, share rate, and completion rate can vary depending on the type of UGC platform. For example, when the UGC platform is a short video / entertainment type, the weights for share rate, completion rate, and like rate are 0.6, 0.3, and 0.1 respectively, making it more likely to spread interesting content to viewers and increase the platform's attractiveness to them. Or, when the UGC platform is a knowledge type, the weights for completion rate, like rate, and share rate are 0.5, 0.3, and 0.1 respectively, etc.; according to the second interaction data statistics of uploading users The average number of views and fan activity of uploaded content on the UGC platform are multiplied to obtain the uploader's score W. Then, the content score Q, semantic score S, interaction score E, and uploader's score W are weighted and summed using the first weight matrix [α1, β1, γ1, δ1] to obtain the first score V1 (i.e., V1 = α1 × Q + β1 × S + γ1 × E + δ1 × W). It is then checked whether V1 exceeds 60% to determine whether to upload the dynamic video through the corresponding other interface, generating a conversion strategy. Finally, the second weight matrix [α2, δ1] is used... The content score Q, semantic score S, interaction score E, and uploader's score W are weighted and summed to obtain the first score V2 (i.e., V2 = α2 × Q + β2 × S + γ2 × E + δ2 × W). It is then checked whether V2 exceeds Vth′ to determine whether to add interactive elements to the dynamic video to generate an editing strategy. Then, the conversion strategy and editing strategy are presented to the uploader in an understandable way. Finally, when the uploader confirms acceptance of the conversion strategy and editing strategy, the dynamic video is processed according to the conversion strategy and / or editing strategy.

[0072] The steps of the various methods described above are only for clarity. In practice, they can be combined into one step or some steps can be split into multiple steps. As long as they include the same logical relationship, they are all within the scope of protection of this patent. Adding insignificant modifications or introducing insignificant designs to the algorithm or process, but without changing the core design of the algorithm and process, are also within the scope of protection of this patent.

[0073] This application also provides a dynamic video processing apparatus, such as... Figure 2 As shown, it includes: The acquisition module is used to acquire dynamic video, first interactive data and second interactive data. The dynamic video is uploaded to the UGC platform by the uploading user. The first interactive data includes data generated by the viewing users of the UGC platform based on the dynamic video and interacting with the UGC platform. The second interactive data includes data generated by the uploading user and interacting with the UGC platform. The first scoring module is used to score the dynamic video based on the dynamic video and the first interactive data; The second rating module is used to rate the uploading user based on the second interaction data; The strategy generation module is used to generate conversion strategies and editing strategies based on the rating of the dynamic video and the rating of the uploading user, respectively. The conversion strategy includes information indicating whether to convert the dynamic video, and the editing strategy includes information indicating whether to add interactive elements to the dynamic video. The video processing module is used to process dynamic videos according to conversion and / or editing strategies.

[0074] It is not difficult to see that this embodiment is a device embodiment corresponding to the method embodiment, and this embodiment can be implemented in conjunction with the method embodiment. The relevant technical details mentioned in the method embodiment are still valid in this embodiment, and will not be repeated here to reduce repetition. Correspondingly, the relevant technical details mentioned in this embodiment can also be applied to the method embodiment.

[0075] It is worth mentioning that all modules involved in this embodiment are logical modules. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this application, this embodiment does not introduce units that are not closely related to solving the technical problems proposed in this application; however, this does not mean that other units are absent in this embodiment.

[0076] This application also provides an electronic device, such as... Figure 3 As shown, it includes: at least one processor 301; and a memory 302 communicatively connected to at least one processor 301; wherein the memory 302 stores instructions executable by at least one processor 301, the instructions being executed by at least one processor 301 to enable at least one processor 301 to perform the dynamic video processing method described in any of the above method embodiments.

[0077] The memory 302 and processor 301 are connected via a bus, which can include any number of interconnecting buses and bridges. The bus connects various circuits of one or more processors 301 and memory 302 together. The bus can also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver can be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 301 is transmitted over a wireless medium via an antenna, which further receives data and transmits it to processor 301.

[0078] Processor 301 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. Memory 302 can be used to store data used by processor 301 during operation.

[0079] This application also provides a computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the above-described method embodiments.

[0080] That is, those skilled in the art will understand that all or part of the steps in the methods of the above embodiments can be implemented by a program instructing related hardware. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as a USB flash drive, a portable hard drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk.

[0081] This application also provides a computer program product, including a computer program that, when at least a portion of the computer program is executed by a processor, can implement the video content adjustment method as described in any of the preceding claims.

[0082] Those skilled in the art will understand that the above embodiments are specific embodiments for implementing this application, and in practical applications, various changes can be made to them in form and detail without departing from the spirit and scope of this application.

Claims

1. A method for processing dynamic video, characterized in that, include: Acquire dynamic video, first interactive data, and second interactive data, wherein the dynamic video is uploaded by the uploading user, the first interactive data includes data generated by the viewing user based on the dynamic video interaction, and the second interactive data includes data generated by the uploading user interaction; The dynamic video is scored based on the dynamic video and the first interactive data; The uploading user is rated based on the second interaction data; Based on the rating of the dynamic video and the rating of the uploading user, a conversion strategy and an editing strategy are generated respectively. The conversion strategy includes information indicating whether to convert the dynamic video, and the editing strategy includes information indicating whether to add interactive elements to the dynamic video. The dynamic video is processed according to the conversion strategy and / or the editing strategy.

2. The method for processing dynamic video according to claim 1, characterized in that, The step of generating a conversion strategy and an editing strategy based on the rating of the dynamic video and the rating of the uploading user includes: Based on the first weight matrix, the rating of the dynamic video and the rating of the uploading user are weighted to generate a first weighted rating of the dynamic video and a first weighted rating of the uploading user. The conversion strategy is generated based on the first weighted score and the second weighted score; Based on the second weight matrix, the rating of the dynamic video and the rating of the uploading user are weighted to generate a third weighted rating of the dynamic video and a fourth weighted rating of the uploading user. The editing strategy is generated based on the third weighted score and the fourth weighted score.

3. The method for processing dynamic video according to claim 2, characterized in that, The step of weighting the rating of the dynamic video and the rating of the uploading user according to the first weight matrix includes: Obtain the first weight matrix; The first weight matrix is ​​updated based on the status data of the dynamic video and / or the protection coefficient of the uploading user, wherein the status data includes one or a combination of the following: publication duration and completion rate, the publication duration being the difference between the current time and the time when the uploading user uploaded the dynamic video, and the protection coefficient being determined based on one or a combination of the following data: user level, creation duration, number of works, and creation frequency, the creation duration being the difference between the current time and the time when the uploading user uploaded the first piece of content; The ratings of the dynamic video and the ratings of the uploading user are weighted according to the updated first weight matrix.

4. The method for processing dynamic video according to claim 3, characterized in that, The step of updating the first weight matrix based on the status data of the dynamic video and / or the protection coefficient of the uploading user includes: Based on the protection coefficient, the interaction weights in the first weight matrix are updated to generate the updated first weight matrix, wherein the interaction elements are elements in the first weight matrix used to weight the interaction-related scores in the scoring of the dynamic video. And / or, The interaction weight is updated based on the publication duration and time decay function.

5. The method for processing dynamic video according to any one of claims 2 to 4, characterized in that, The step of generating the editing strategy based on the third weighted score and the fourth weighted score includes: Determine whether the score generated based on the third weighted score and the fourth weighted score exceeds the threshold; If the threshold is exceeded, the interactive element is added to the dynamic video to generate the corresponding editing strategy; If the threshold is not exceeded, it is determined not to add the interactive elements to the dynamic video in order to generate the corresponding editing strategy.

6. The method for processing dynamic video according to claim 5, characterized in that, Before determining whether the score generated based on the third weighted score and the fourth weighted score exceeds the threshold, the method further includes: Obtain the initial threshold; A correction factor is generated based on the content type of the dynamic video; The initial threshold is updated based on the correction factor to generate the threshold.

7. The method for processing dynamic video according to claim 6, characterized in that, The step of updating the initial threshold according to the correction factor includes: Determine the content scarcity parameter of the dynamic video, wherein the content scarcity parameter is used to describe the degree of scarcity of the content of the dynamic video; The initial threshold is updated based on the correction factor and the content scarcity parameter.

8. The method for processing dynamic video according to claim 6 or 7, characterized in that, The step of generating a correction factor based on the content type of the dynamic video includes: Based on the content type of the dynamic video, a preset parameter set is matched to use the matched preset parameter as the correction factor. The preset parameter set includes at least two preset parameters associated with different content types.

9. The method for processing dynamic video according to any one of claims 2 to 8, characterized in that, The conversion strategy also includes a score generated based on the first weighted score and the second weighted score, and / or the editing strategy also includes a score generated based on the third weighted score and the fourth weighted score.

10. The method for processing dynamic video according to any one of claims 1 to 9, characterized in that, The rating of the dynamic video includes one or a combination of the following: semantic rating, content rating, and interaction rating, wherein the semantic rating is used to describe the emotional tendency of the viewing user towards the dynamic video, the content rating is used to describe the image quality of the dynamic video, and the interaction rating is used to describe the interaction of the viewing user with the dynamic video.

11. The method for processing dynamic video according to any one of claims 1 to 10, characterized in that, The method further includes: Acquire third-party interaction data and user preference data, wherein the third-party interaction data includes traffic data and consumption data of content with the same type as the dynamic video, and the user preference data is data generated by viewing user interaction, or a knowledge graph generated based on the data generated by viewing user interaction to describe viewing user preferences; Based on the third interaction data and the user preference data, predict the traffic generated by the converted dynamic video, the probability of the traffic occurring, the revenue, and the probability of the corresponding revenue occurring.

12. A dynamic video processing apparatus, characterized in that, include: The acquisition module is used to acquire dynamic video, first interactive data and second interactive data, wherein the dynamic video is uploaded by the uploading user, the first interactive data includes data generated by the viewing user based on the dynamic video interaction, and the second interactive data includes data generated by the uploading user interaction. The first scoring module is used to score the dynamic video based on the dynamic video and the first interactive data; The second rating module is used to rate the uploading user based on the second interaction data; The strategy generation module is used to generate a conversion strategy and an editing strategy based on the rating of the dynamic video and the rating of the uploading user, respectively. The conversion strategy includes information indicating whether to re-upload the dynamic video, and the editing strategy includes information indicating whether to add interactive elements to the dynamic video. A video processing module is used to process the dynamic video according to the conversion strategy and / or the editing strategy.

13. An electronic device, characterized in that, include: At least one processor; as well as, A memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, which, when executed by the at least one processor, can implement the method for processing dynamic video as described in any one of claims 1 to 11.

14. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer instructions, which, when executed by a processor, enable the processing method for dynamic video as described in any one of claims 1 to 11.

15. A computer program product, characterized in that, It includes a computer program, which, when at least a portion of the computer program is executed by a processor, enables the processing method of dynamic video as described in any one of claims 1 to 11.

Images