Video generation rendering method and system based on AI technology, and storage medium
Through the video generation and rendering method based on AI technology, the audience user data is analyzed and videos are generated using AI models, which solves the problems of high labor intensity and insufficient update frequency of video creation, and improves video creation efficiency and market competitiveness.
Patent Information
- Application Number
- CN202510251747.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-05
- Publication Date
- 2025-05-09
AI Technical Summary
In the existing technology, video creation has a high labor intensity and insufficient video update frequency, resulting in video creators being at a disadvantage in market competition.
Using a video generation and rendering method based on AI technology, we generate reference data by analyzing the information data, preference data and loyalty data of audience users, and using AI model to render to generate output videos.
It shortens the video production time and video style selection time, improves video creation efficiency, enables video update efficiency to meet market demand, and avoids being eliminated by the market due to insufficient update efficiency.
Smart Images
Figure CN119967235A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of stereoscopic video systems, and specifically to a video generation and rendering method, system and storage medium based on AI technology. Background Art
[0002] The video creation industry is currently in rapid development. With the popularization of digital technology and the Internet, the demand and consumption of video content are growing. From individual current video authors to large production companies, various platforms and media are actively promoting the production and dissemination of creative content. New technologies such as real-time rendering, virtual reality and augmented reality are changing the way of creation. At the same time, current video authors are also facing the challenges of fierce market competition and content diversification.
[0003] However, with the continuous development of the video creation industry, market users have higher and higher requirements for the update frequency of current video authors. Current video authors who do not update frequently are often more likely to be eliminated by the market. In order to maintain or improve their popularity, current video authors need to continuously improve their output efficiency. However, for most video current video authors, especially individual current video authors, all related work needs to be completed by one person independently, and it is challenging to complete a high-quality video in a short period of time. Summary of the invention
[0004] The purpose of this application is to provide a video generation and rendering method, system and storage medium based on AI technology to solve the technical problem of high labor intensity in video creation in the prior art.
[0005] To achieve the above objectives, this application provides the following technical solutions: In a first aspect, the present application proposes a technical solution for a video generation and rendering method based on AI technology, and the video generation and rendering method based on AI technology includes: Based on big data technology, obtain information data of the audience users of the current video author; the information data includes the audience users' attention information data, click information data, like information data, viewing time data and viewing frequency data; Based on the information data of the audience user, obtaining the audience user's preference data; the preference data is at least used to characterize the audience user's preference for the emotion type, emotion intensity, scene change frequency and audio rhythm of the video; Based on the information data of the audience user, acquiring the loyalty data of the audience user; the loyalty data is at least used to characterize the degree of liking of the audience user for the current video author; acquiring reference data based on the preference data and the loyalty data; Based on the original material and the reference data, an output video is generated by rendering using an AI model; the original material and the AI model are acquired in advance.
[0006] As a specific solution in the technical solution of the present application, the method of obtaining the preference data of the audience user based on the information data of the audience user includes: Based on the information data of the audience users, a first user is obtained; the first user is any user among the audience users; Constructing a FP tree based on the information data of the first user; Based on the FP tree, the preference data of the first user is obtained.
[0007] As a specific solution in the technical solution of the present application, the method of obtaining the loyalty data of the audience user based on the information data of the audience user includes: Based on the information data of the audience user, a second user is obtained; the second user is any user among the audience users; Based on the information data of the audience user, obtaining the number of clicks by the second user on each video author and the number of likes by the second user on the current video author within a preset time period; Based on each number of clicks, a first number of clicks and a standard deviation of clicks are obtained; the first number of clicks is the number of clicks by the second user on the author of the current video; Obtaining a first coefficient based on the first number of clicks, the click standard deviation, and the number of likes; Based on the first coefficient, loyalty data of the second user is obtained.
[0008] As a specific solution in the technical solution of the present application, the step of obtaining the loyalty data of the second user based on the first coefficient includes: Based on the information data of the audience user, a first duration and a second duration are obtained; the first duration is the total duration of the second user watching the video within the preset time period; the second duration is the total duration of the second user watching the video of the author of the current video within the preset time period; Obtaining a second coefficient based on the first duration, the second duration, and the first coefficient; Based on the second coefficient, loyalty data of the second user is obtained.
[0009] As a specific solution in the technical solution of the present application, the step of obtaining the loyalty data of the second user based on the second coefficient includes: Based on the information data of the audience user, obtain all video authors followed by the second user; Based on all video authors, a second video author is obtained; the second video author is any video author among all video authors; Based on the author of the second video and the author of the current video, obtaining the similarity of the published videos; Based on the video similarity and the second coefficient, obtaining a third coefficient; Based on the third coefficient, loyalty data of the second user is obtained.
[0010] As a specific solution in the technical solution of the present application, the step of obtaining the loyalty data of the second user based on the third coefficient includes: Based on the information data of the audience user, the forwarding number is obtained; the forwarding number is the total number of times the second user forwards the video within the preset time period; Based on the number of forwarding times and the first duration, obtaining a profit coefficient; Loyalty data of the second user is obtained based on the profit coefficient and the third coefficient.
[0011] As a specific solution in the technical solution of the present application, the obtaining of reference data based on the preference data and the loyalty data includes: Based on the preference data, a first video type is acquired; the first video type is any video type in the preference data; Based on the preference data, obtaining a preference value of each user in the audience for the first video type; Based on the loyalty data, obtaining the loyalty value of each user in the audience user to the current video author; Based on the respective preference values and the respective loyalty values, reference data of the first video type is obtained.
[0012] As a specific solution in the technical solution of the present application, the calculation formula for obtaining the reference data of the first video type based on each preference value and each loyalty value is as follows:
[0013] in, e represents the reference data of the video type; x represents the number of audience users of the current video author; represents the loyalty value of the i-th audience user; Indicates the preference value of the i-th audience user for the video of type e.
[0014] In a second aspect, the present application proposes a technical solution of a video generation and rendering system based on AI technology, and the video generation and rendering system based on AI technology includes: A reader, for obtaining information data of audience users of the current video author based on big data technology; the information data includes attention information data, click information data, like information data, viewing time data and viewing frequency data of the audience users; The server obtains the preference data of the audience user based on the information data of the audience user; the preference data is used to at least characterize the audience user's preference for the emotion type, emotion intensity, scene change frequency and audio rhythm of the video; And, based on the information data of the audience user, acquiring the loyalty data of the audience user; the loyalty data is at least used to characterize the degree of liking of the audience user for the current video author; and, based on the preference data and the loyalty data, obtaining reference data; And, based on the original material and the reference data, an output video is generated by rendering using an AI model; the original material and the AI model are acquired in advance.
[0015] As a specific solution in the technical solution of the present application, the server is further used to obtain a first user based on the information data of the audience user; the first user is any user among the audience users; And, constructing a FP tree based on the information data of the first user; And, based on the FP tree, obtaining the preference data of the first user.
[0016] As a specific solution in the technical solution of the present application, the server is further used to obtain a second user based on the information data of the audience user; the second user is any user among the audience users; And, based on the information data of the audience user, obtaining the number of clicks by the second user on each video author and the number of likes by the second user on the current video author within a preset time period; And, based on each number of clicks, obtaining a first number of clicks and a click standard deviation; the first number of clicks is the number of clicks by the second user on the author of the current video; and, obtaining a first coefficient based on the first number of clicks, the click standard deviation, and the number of likes; And, based on the first coefficient, acquiring the loyalty data of the second user.
[0017] As a specific solution in the technical solution of the present application, the server is also used to obtain a first duration and a second duration based on the information data of the audience user; the first duration is the total duration of the second user watching the video within the preset time period; the second duration is the total duration of the second user watching the video of the author of the current video within the preset time period; and, based on the first duration, the second duration and the first coefficient, obtaining a second coefficient; And, based on the second coefficient, acquiring loyalty data of the second user.
[0018] As a specific solution in the technical solution of the present application, the server is also used to obtain all video authors followed by the second user based on the information data of the audience user; And, based on all video authors, a second video author is obtained; the second video author is any video author among all video authors; And, based on the author of the second video and the author of the current video, obtaining the similarity of the published videos; and, based on the video similarity and the second coefficient, obtaining a third coefficient; And, based on the third coefficient, acquiring the loyalty data of the second user.
[0019] As a specific solution in the technical solution of the present application, the server is further used to obtain the number of forwarding times based on the information data of the audience user; the number of forwarding times is the total number of times the second user forwards the video within the preset time period; And, based on the number of forwarding times and the first duration, obtaining a profit coefficient; And, based on the profit coefficient and the third coefficient, the loyalty data of the second user is obtained.
[0020] As a specific solution in the technical solution of the present application, the server is further used to obtain a first video type based on the preference data; the first video type is any video type in the preference data; And, based on the preference data, obtaining the preference level value of each user in the audience for the first video type; And, based on the loyalty data, obtaining the loyalty value of each user in the audience users to the current video author; And, based on the respective preference degree values and the respective loyalty values, obtaining reference data of the first video type.
[0021] As a specific solution in the technical solution of the present application, the server is further used to obtain the reference data of the first video type based on each preference value and each loyalty value according to the following calculation formula:
[0022] in, e represents the reference data of the video type; x represents the number of audience users of the current video author; represents the loyalty value of the i-th audience user; Indicates the preference value of the i-th audience user for the video of type e.
[0023] In a third aspect, the present application proposes a technical solution of a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements a video generation and rendering method based on AI technology as described in any one of the first aspects.
[0024] Compared with the prior art, the beneficial effects of this application are: This application analyzes and determines the video styles that audience users like, and reasonably uses AI generation technology to shorten the video production time and video style selection time based on the original materials provided by the creators, thereby helping video creators improve their video creation efficiency, so that the creators' video update efficiency can meet market demand and avoid being gradually eliminated by the market due to insufficient video update efficiency. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 A flowchart of a video generation and rendering method based on AI technology proposed in an embodiment of the present application; Figure 2 This is a structural diagram of a video generation and rendering system based on AI technology proposed in an embodiment of the present application. DETAILED DESCRIPTION
[0026] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0027] The terms "first", "second", etc. in the specification of the embodiment of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence, such as the first user and the second user proposed below, which belong to different users. It should be understood that the users used in this way can be interchanged when appropriate, so that the embodiments described here can be implemented in an order other than the content illustrated or described here. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device including a series of steps or modules need not be limited to those steps or modules clearly listed, but may include other steps or modules that are not clearly listed or inherent to these processes, methods, products or devices. The division of the modules that appear in the embodiment of the present application is only a logical division. There may be other division methods when implemented in practical applications, such as multiple modules can be combined or integrated into another system, or some features can be ignored, or not executed. In addition, the coupling or direct coupling or communication connection between each other shown or discussed can be through some interfaces, and the indirect coupling or communication connection between modules can be electrical or other similar forms, which are not limited in the embodiment of the present application. Furthermore, the modules or sub-modules described as separate components may or may not be physically separated, may or may not be physical modules, or may be distributed in multiple circuit modules, and some or all of the modules may be selected according to actual needs to achieve the purpose of the embodiment of the present application.
[0028] Before understanding the embodiments of the present application, it should be clear that video rendering refers to the process of generating the final video image by processing a computer-generated 3D model or 2D image sequence through lighting, coloring, shadows, etc. using computer graphics technology. In the rendering process, many factors need to be considered, such as lighting model, material properties, camera angle, etc., to ensure that the generated image can accurately present the designer's intention. This is a mature technology and will not be introduced in detail here.
[0029] In order to solve the technical problem of high labor intensity in video creation in the prior art mentioned in the background technology, this application proposes a video generation and rendering method based on AI technology, such as Figure 1 As shown, the video generation and rendering method based on AI technology includes steps S100 to S500.
[0030] Step S100: Based on big data technology, obtain information data of the audience users of the current video author.
[0031] It should be clear that big data technology specifically refers to the general term for technologies that process, analyze and apply big data through various big data platforms, applications, and technology indexes. Since big data technology is a mature technology, it will not be described in detail here. The information data described in the embodiments of the present application include but are not limited to the audience user's attention information data, click information data, like information data, viewing time data, and viewing frequency data. After obtaining the information data of the audience user based on big data, the collected data is cleaned and desensitized, and missing values and outliers are processed (deleted or corrected) to ensure the availability of the obtained data, and the processed data is stored in the database as the basis for subsequent analysis.
[0032] Step S200: Based on the information data of the audience users, the preference data of the audience users is obtained.
[0033] It should be clear that the preference data described in this embodiment is at least used to characterize the audience user's preference for the emotional type, emotional intensity, scene change frequency and audio rhythm of the video. Specifically, the emotional type of the video includes plot, animation, single-person narration, video / picture carousel or quotation, etc.; the emotional intensity of the video includes happiness, satisfaction, curiosity, anger, anxiety, disappointment, etc.; the scene change frequency of the video is usually determined by the frame rate of the video, that is, the number of frames displayed per second; the audio rhythm of the video refers to the fast and slow changes of the audio elements in the video (including background music, sound effects, etc.), and the coordination between these changes and the video picture and other audio elements. That is to say, in the embodiment of the present application, the emotional type, emotional intensity, scene change frequency and audio rhythm of the video are some artificially defined video parameters. The audience user's preference data is the audience user's preference for these video parameters. For example, if a user's preference for plot videos is 0.8 and the preference for animation videos is 0.1; it means that the user prefers plot videos more than animation videos. In the subsequent video production process, you can produce a plot video that the user likes more to attract the user's attention.
[0034] In the embodiment of the present application, the preference data of the audience user can be obtained based on the information data of the audience user in any manner. For example, in a specific embodiment of the present application, step S200, based on the information data of the audience user, obtains the preference data of the audience user, including steps S210 to S230.
[0035] Step S210: Acquire the first user based on the information data of the audience user.
[0036] It should be clear that the first user is any user among the audience users. That is to say, in this embodiment, it is necessary to obtain the preference data of all audience users of the current video author. In the embodiment of the present application, all audience users of the current video author may refer to all users who have only followed the current video author, or all users who have liked or commented on the current video author, or all users who have clicked on any video work of the current video author.
[0037] Step S220: constructing an FP tree based on the information data of the first user.
[0038] It should be clear that the FP-tree algorithm is an algorithm for data mining, which is mainly used to find frequent item sets in a data set. It is based on a part of the FP-Growth algorithm. By building an FP tree, the data in the data set is mapped to the tree, and then all frequent item sets are found according to the tree. The FP-tree algorithm solves the problem that the Apriori algorithm will generate a large number of candidate sets during the processing process. By only scanning the data set twice, unnecessary calculations are avoided and execution efficiency is improved. In other words, the embodiment of the present application can obtain the degree of preference of each user for various parameters of the video through the FP-tree.
[0039] Step S230: Based on the FP tree, obtain the preference data of the first user.
[0040] In an embodiment of the present application, based on the FP tree, the steps of obtaining the preference data of the first user are as follows: input the annotated video and audio into the neural network, and obtain the video sentiment analysis tool through model training. The collected data includes the video information that the user likes on the platform, and the sentiment analysis tool is used to extract the key emotional information of the video, including the emotion type, emotion intensity, scene change frequency and audio rhythm information, and the AI algorithm is used to detect and analyze the video content and identify the basic type of the video, extract key frames from the video, and identify the scene changes and main objects in the video.
[0041] The corresponding information of the videos liked by a certain user is counted, and the minimum support is set (for example: 5 or 6). The FP tree is constructed based on the video information, and the constructed FP tree is mined. The association confidence between the main objects of each video and other key information is established through the frequent item set results. The association confidence result is regarded as the user's preference information, which represents the specific video rhythm type that the current user prefers when watching a certain type of video. It should be clear that obtaining the preference data of each user based on the FP tree is a mature technology and will not be elaborated here.
[0042] Step S300: Based on the information data of the audience users, the loyalty data of the audience users is obtained.
[0043] It should be clear that in the embodiments of the present application, the loyalty data is at least used to characterize the audience user's liking for the current video author or the video produced by the current video author. It is easy to understand that if a user likes the current video author or the video produced by the current video author more, it means that the user's loyalty to the current video author is higher, and in the subsequent video production process, it is more necessary to produce videos that the user likes. When creating a video based on the audience user, the loyalty data of the audience user can be referred to, and then the video that the audience user likes can be generated based on the existing materials. It is easy to understand that the value of each user to the current video author is different, so the reference value of each user's preference information to the current video author is different, and it is necessary to calculate the loyalty data of each user to the current video author, and then determine what style of video is more popular with most audience users.
[0044] In an embodiment of the present application, the audience user's loyalty data can be obtained based on the audience user's information data in any manner. For example, the audience user's likes, comments, or reposts can be used as the audience user's loyalty data. In a specific embodiment of the present application, step S300, based on the audience user's information data, obtains the audience user's loyalty data, including steps S310 to S350.
[0045] Step S310: Acquire a second user based on the information data of the audience user.
[0046] It should be clear that in the embodiment of the present application, the loyalty data of each audience user needs to be calculated, so in this embodiment, the second user is any user among the audience users.
[0047] Step S320: Based on the information data of the audience user, the number of clicks by the second user on each video author and the number of likes by the second user on the current video author within a preset time period are obtained.
[0048] It should be clear that for the current video author, the more frequently a user clicks on the current video author's videos, the more videos he likes, and the more balanced the distribution of clicks on different videos is, the more it can be explained that the user does not like a certain video of the current video author, and the higher the user's loyalty to the current video author is, and the more likely the user is to be a stable audience of the current video author.
[0049] In the embodiments of the present application, the preset time period can be selected according to actual needs. For example, if the current video author wants to retain users who like him / her more, the preset time period can be longer, such as one month or one year; if the current video author wants to attract the attention of new users, the preset time period can be shorter, such as one day or one week.
[0050] Step S330: Based on each number of clicks, obtain the first number of clicks and the standard deviation of the clicks.
[0051] In an embodiment of the present application, the first number of clicks is the number of clicks by the second user on the author of the current video within a preset time period. The click standard deviation is the standard deviation of clicks by the second user on different videos of the author of the current video within a preset time period.
[0052] Step S340: Obtain a first coefficient based on the first number of clicks, the click standard deviation, and the number of likes.
[0053] It is easy to understand that the user's preference for the current video author is directly proportional to the number of clicks and likes, and inversely proportional to the click standard deviation. Therefore, in one embodiment of the present application, the calculation formula of the first coefficient can be as follows:
[0054] in, represents the first coefficient of the second user; C represents the number of clicks; L represents the number of likes; S represents the standard deviation of clicks, and N represents the anti-zero coefficient to prevent the denominator from being 0. N can be any non-zero positive integer, for example, 1 or 2.
[0055] In another embodiment of the present application, the calculation formula of the first coefficient may be as follows:
[0056] in, represents the first coefficient of the second user; C represents the number of clicks; L represents the number of likes; S represents the standard deviation of clicks; N represents the anti-zero coefficient to prevent the denominator from being 0; D represents the first constant. The anti-zero coefficient N can be any non-zero positive integer, for example, 1 or 2. The first constant D can also be any non-zero positive integer, for example, 100 or 200.
[0057] Step S350: Based on the first coefficient, obtain the loyalty data of the second user.
[0058] In an embodiment of the present application, the first coefficient corresponding to the second user can be directly regarded as the loyalty data of the second user. It should be noted that the above process does not take into account whether some users watch the video after clicking it. For the current video author, the longer the total time the user watches the video of the current video author, the more attractive the video content of the current video author is to the user, and the higher the user's loyalty to the current video author. Based on this, in an embodiment of the present application, step S350, based on the first coefficient, obtains the loyalty data of the second user, including steps S360 to S380.
[0059] Step S360: Based on the information data of the audience user, obtain the first duration and the second duration.
[0060] In an embodiment of the present application, the first duration is the total duration of the second user watching videos within the preset time period. The second duration is the total duration of the second user watching videos of the author of the current video within the preset time period. It is easy to understand that the larger the ratio of the second duration to the first duration, the more the second user likes the video produced by the author of the current video.
[0061] Step S370: Obtain a second coefficient based on the first duration, the second duration and the first coefficient.
[0062] In one embodiment of the present application, the calculation formula of the second coefficient is as follows:
[0063] in, a second coefficient representing a second user; A first coefficient representing a second user; represents a second duration of a second user; Indicates the first duration of the second user.
[0064] Step S380: Based on the second coefficient, obtain the loyalty data of the second user.
[0065] In the embodiment of the present application, the second coefficient of the second user may be directly used as the loyalty data of the second user.
[0066] It should be noted that in addition to the stable user group with high loyalty to the current video author, the attractiveness of the video to potential audiences should also be considered in the embodiments of the present application. The greater the probability of a user being converted into a high-loyalty user and the higher the platform usage of the user, the more attention the current video author should pay to the user's preference information. Therefore, when the similarity between other current video authors followed by a user and the current video author is higher, it means that the user is more likely to become a loyal audience of the current video author. At the same time, the above formula should be used to calculate the user's loyalty to each current video author in the follow-up list to more accurately evaluate the user's conversion possibility. Therefore, in one embodiment of the present application, step S380, based on the second coefficient, obtains the loyalty data of the second user, including steps S381 to S385.
[0067] Step S381: Based on the information data of the audience user, obtain all video authors followed by the second user.
[0068] It should be clear that obtaining some fixed data among certain data is a mature technology and will not be elaborated here.
[0069] Step S382: Based on all video authors, obtain the second video author.
[0070] It should be clear that the second video author is any video author among all video authors.
[0071] Step S383: Based on the second video author and the current video author, obtain the similarity of the published videos.
[0072] It should be clear that obtaining the similarity between two videos or between multiple videos is a mature technology. For example, in an embodiment of the present application, computer vision and natural language processing technology can be used to extract visual and text features from the video library of the current video author and the second video author, including scene recognition, object detection, color histogram, emotional tags, and keywords, etc.; then, audio analysis tools are used to extract video audio features, including rhythm, pitch, volume, etc.; then, machine learning algorithms are used to convert these features into multidimensional feature vectors to characterize the style and theme of each creator; then, the cosine similarity between these feature vectors is calculated to obtain the video similarity between the creators.
[0073] Step S384: Obtain a third coefficient based on the video similarity and the second coefficient.
[0074] It should be clear that if most of the video authors followed by the second user have similar video similarities to the current video author, it means that the second user is more loyal to the current video author. In this embodiment, step S384, based on the video similarity and the second coefficient, the calculation formula for obtaining the third coefficient is as follows:
[0075] in, represents the third coefficient of the second user; w represents the number of video authors followed by the second user; The second coefficient representing the jth video author (i.e., the second video author); Indicates the video similarity between the jth video author and the current video author. In the embodiment of the present application, the second video author may be the current video author. It is easy to understand that if the second video author is the current video author, the video similarity is equal to 1.
[0076] Step S385: Based on the third coefficient, obtain the loyalty data of the second user.
[0077] In an embodiment of the present application, the third coefficient of the second user can be directly used as the loyalty data of the second user. It should be clear that the purpose of the author who creates the video is to obtain income. Therefore, in an embodiment of the present application, the income that each user can bring to the video author can also be considered. It is easy to understand that the higher the income a user brings to the video author, the more the video author should need to get closer to the user's preferences when making the video in order to attract the user and strive for higher income. It is easy to understand that the longer a user watches the video and the more times the video is forwarded, the higher the income the user can bring to the video author. In one embodiment of the present application, step S385, based on the third coefficient, obtains the loyalty data of the second user, including steps S386 to S388.
[0078] Step S386: Based on the information data of the audience user, obtain the number of forwarding times.
[0079] It is easy to understand that the forwarding times are the total number of times the second user forwarded the video within the preset time period. It is easy to understand that if the second user is a user who likes to forward, the total number of times he forwards within the preset time period is greater. That is, after attracting this user, the income that this user can bring to the current video author is also higher.
[0080] Step S387: Obtain a profit coefficient based on the number of forwarding times and the first duration.
[0081] As can be seen from the foregoing, the first duration is the total duration of the second user watching the video within the preset time period. In an embodiment of the present application, the profit coefficient can be obtained in any manner based on the number of forwarding times and the first duration. For example, in one embodiment of the present application, step S387, based on the number of forwarding times and the first duration, the calculation formula for obtaining the profit coefficient is as follows:
[0082] in, represents the profit coefficient of the second user; represents the first duration; q represents the number of forwarding times; represents a norm function, which is used to convert the first duration and the number of forwarding times q into a space vector.
[0083] In another embodiment of the present application, in step S387, based on the number of forwarding times and the first duration, a calculation formula for obtaining a revenue coefficient is as follows:
[0084] in, represents the profit coefficient of the second user; represents the first duration; q represents the number of forwarding times; represents a norm function, which is used to convert the first duration and the number of forwarding times q into a space vector.
[0085] Step S388: Based on the profit coefficient and the third coefficient, obtain the loyalty data of the second user.
[0086] In an embodiment of the present application, the sum or product of the profit coefficient and the third coefficient may be used as the loyalty data of the second user.
[0087] Step S400: Acquire reference data based on the preference data and the loyalty data.
[0088] In the embodiments of the present application, the reference data is mainly used for reference when the current video author or the AI model described below generates a video work. That is to say, in the embodiments of the present application, the reference data is at least used to characterize the emotional type, emotional intensity, scene change frequency, and audio rhythm trend of the generated video. As can be seen from the foregoing, if the emotional type of the video only includes plot and animation; and the reference data for producing the plot is greater than the reference data for the animation, it means that when producing this video, you should choose to produce a video with the emotional type of the plot, so as to ensure that the produced video can be liked by the audience and can obtain greater benefits.
[0089] It should be clear that in the embodiments of the present application, in order to avoid redundancy, it is impossible to enumerate all application scenarios. Therefore, in the embodiments of the present application, the emotional type of the video is taken as an example. It is assumed that the emotional type of the video in the embodiments of the present application includes n types (for example: the plot type, animation type, single-person narration type, video / picture carousel type or quotation type mentioned above). In this embodiment, step S400, based on the preference data and the loyalty data, obtains reference data, including steps S410 to S440.
[0090] Step S410: Based on the preference data, obtain a first video type.
[0091] It should be clear that the first video type is any video type in the preference data. Taking the emotion type of the video as an example, in this embodiment, the first video type can be any type of plot type, animation type, single-person narration type, video / picture carousel type or quotation type.
[0092] Step S420: Based on the preference data, obtain the preference level value of each user in the audience for the first video type.
[0093] It should be clear that obtaining some fixed data among certain data is a mature technology and will not be elaborated here.
[0094] Step S430: Based on the loyalty data, obtain the loyalty value of each user in the audience to the current video author.
[0095] It should be clear that obtaining some fixed data among certain data is a mature technology and will not be elaborated here.
[0096] Step S440: Based on the preference values and the loyalty values, obtain reference data of the first video type.
[0097] Specifically, in this embodiment, in step S440, based on each preference value and each loyalty value, the calculation formula for obtaining the reference data of the first video type is as follows:
[0098] in, e represents the reference data of the video type (i.e., the first video type); x represents the number of audience users of the current video author; represents the loyalty value of the i-th audience user; Indicates the preference value of the i-th audience user for the video of video type e. Step S500: Based on the original material and the reference data, the output video is generated by rendering using the AI model.
[0099] It should be clear that the original material and the AI model are acquired in advance. In this embodiment, the original material includes images, video clips, text descriptions, etc. When in use, the original material is input into the AI model, and then the AI model is used to analyze and extract the key information in the original material, and the AI model is used to select the appropriate style and color palette to match the emotional atmosphere according to the emotion type and emotion intensity specified by the creator. At the same time, combined with the scene change frequency and audio rhythm, the AI model is used to generate or select the corresponding transition effects and background music to ensure the rhythm and smoothness of the video. Then, the AI model is used to merge the material and the selected style with the help of a deep learning mechanism to create new video content. Finally, the AI model is used for post-processing, including color correction, audio synchronization, etc., to generate the final output video.
[0100] It should be noted that during the video generation process, creators are allowed to fine-tune and provide feedback at each stage to ensure that the final output video meets the creative vision, and collect feedback from creators, and use the feedback to self-optimize the AI model to minimize the process of manual adjustments by creators in the subsequent video generation process.
[0101] It should be clear that the video generation and rendering method based on AI technology proposed in this application analyzes and determines the video styles that the audience users like, and reasonably uses AI generation technology to shorten the video production time and video style selection time based on the original materials provided by the creator, thereby helping video creators improve video creation efficiency, so that the creator's video update efficiency can meet market demand and avoid being gradually eliminated by the market due to insufficient video update efficiency.
[0102] After introducing an embodiment of a video generation and rendering method based on AI technology proposed in this application, the following introduces a video generation and rendering system based on AI technology proposed in an embodiment of this application. Figure 2 As shown, the video generation and rendering system 10 based on AI technology includes: The reader 11 is used to obtain information data of the audience users of the current video author based on big data technology; the information data includes attention information data, click information data, like information data, viewing time data and viewing frequency data of the audience users; The server 12 obtains the preference data of the audience user based on the information data of the audience user; the preference data is at least used to characterize the audience user's preference for the emotion type, emotion intensity, scene change frequency and audio rhythm of the video; And, based on the information data of the audience user, acquiring the loyalty data of the audience user; the loyalty data is at least used to characterize the degree of liking of the audience user for the current video author; and, based on the preference data and the loyalty data, obtaining reference data; And, based on the original material and the reference data, an output video is generated by rendering using an AI model; the original material and the AI model are acquired in advance.
[0103] As a specific embodiment of the present application, the server 12 is further used to obtain a first user based on the information data of the audience user; the first user is any user among the audience users; And, constructing a FP tree based on the information data of the first user; And, based on the FP tree, obtaining the preference data of the first user.
[0104] As a specific embodiment of the present application, the server 12 is further used to obtain a second user based on the information data of the audience user; the second user is any user among the audience users; And, based on the information data of the audience user, obtaining the number of clicks by the second user on each video author and the number of likes by the second user on the current video author within a preset time period; And, based on each number of clicks, obtaining a first number of clicks and a click standard deviation; the first number of clicks is the number of clicks by the second user on the author of the current video; and, obtaining a first coefficient based on the first number of clicks, the click standard deviation, and the number of likes; And, based on the first coefficient, acquiring the loyalty data of the second user.
[0105] As a specific embodiment of the present application, the server 12 is further used to obtain a first duration and a second duration based on the information data of the audience user; the first duration is the total duration of the second user watching the video within the preset time period; the second duration is the total duration of the second user watching the video of the current video author within the preset time period; and, based on the first duration, the second duration and the first coefficient, obtaining a second coefficient; And, based on the second coefficient, acquiring loyalty data of the second user.
[0106] As a specific embodiment of the present application, the server 12 is further used to obtain all video authors followed by the second user based on the information data of the audience user; And, based on all video authors, a second video author is obtained; the second video author is any video author among all video authors; And, based on the author of the second video and the author of the current video, obtaining the similarity of the published videos; and, based on the video similarity and the second coefficient, obtaining a third coefficient; And, based on the third coefficient, acquiring the loyalty data of the second user.
[0107] As a specific embodiment of the present application, the server 12 is further used to obtain the number of forwarding times based on the information data of the audience user; the number of forwarding times is the total number of times the second user forwards the video within the preset time period; And, based on the number of forwarding times and the first duration, obtaining a profit coefficient; And, based on the profit coefficient and the third coefficient, the loyalty data of the second user is obtained.
[0108] As a specific embodiment of the present application, the server 12 is further configured to obtain a first video type based on the preference data; the first video type is any video type in the preference data; And, based on the preference data, obtaining the preference level value of each user in the audience for the first video type; And, based on the loyalty data, obtaining the loyalty value of each user in the audience users to the current video author; And, based on the respective preference degree values and the respective loyalty values, obtaining reference data of the first video type.
[0109] As a specific embodiment of the present application, the server 12 is further configured to obtain reference data of the first video type based on each preference value and each loyalty value using the following calculation formula: in, e represents the reference data of the video type; x represents the number of audience users of the current video author; represents the loyalty value of the i-th audience user; Indicates the preference value of the i-th audience user for the video of type e.
[0110] It should be clear that the video generation and rendering system based on AI technology proposed in this application analyzes and determines the video styles that audience users like, and reasonably uses AI generation technology to shorten the video production time and video style selection time based on the original materials provided by the creators, thereby helping video creators improve video creation efficiency, so that the creators' video update efficiency can meet market demand and avoid being gradually eliminated by the market due to insufficient video update efficiency.
[0111] After introducing an embodiment of a video generation and rendering system based on AI technology proposed in the present application, a computer-readable storage medium proposed in an embodiment of the present application is introduced below, on which a computer program is stored. When the computer program is executed by a processor, a video generation and rendering method based on AI technology as described in any of the above embodiments is implemented.
[0112] It should be clear that the computer-readable storage media in this application include permanent and non-permanent, removable and non-removable media that can be used to store information by any method or technology. Information can be computer-readable instructions, data structures, modules of programs or other data. Examples of computer storage media include, but are not limited to, phase change memory, static random access memory, dynamic random access memory, other types of random access memory, read-only memory, electrically erasable programmable read-only memory, flash memory or other memory technology, read-only compact disc read-only memory, digital versatile disc or other optical storage, magnetic cassettes, disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer-readable media does not include temporary computer-readable media, such as modulated data signals and carrier waves.
[0113] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.
[0114] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the methods, devices and equipment described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0115] In the several embodiments provided in the embodiments of the present application, it should be understood that the disclosed devices, apparatuses and methods can be implemented in other ways. For example, the device embodiments described above are merely schematic. For example, the division of the modules is only a logical function division. There may be other division methods in actual implementation, such as multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or modules, which can be electrical, mechanical or other forms.
[0116] The modules described as separate components may or may not be physically separated, and the components shown as modules may or may not be physical modules, that is, they may be located in one place or distributed on multiple network modules. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.
[0117] In addition, each functional module in each embodiment of the present application can be integrated into a processing module, or each module can exist physically separately, or two or more modules can be integrated into one module. The above-mentioned integrated module can be implemented in the form of hardware or in the form of a software functional module. If the integrated module is implemented in the form of a software functional module and sold or used as an independent product, it can be stored in a computer-readable storage medium.
[0118] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.
[0119] The computer program product includes one or more computer instructions. When the computer program is loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can store or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a digital video disk), or a semiconductor medium (e.g., a solid-state drive (SSD)).
[0120] Although the embodiments of the present application have been shown and described, it will be appreciated by those skilled in the art that many changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles of the present application.
Claims
1. A video generation and rendering method based on AI technology, characterized in that: include: Based on big data technology, obtain information data of the audience users of the current video author; the information data includes the audience users' attention information data, click information data, like information data, viewing time data and viewing frequency data; Based on the information data of the audience user, obtaining the audience user's preference data; the preference data is at least used to characterize the audience user's preference for the emotion type, emotion intensity, scene change frequency and audio rhythm of the video; Based on the information data of the audience user, acquiring the loyalty data of the audience user; the loyalty data is at least used to characterize the degree of liking of the audience user for the current video author; acquiring reference data based on the preference data and the loyalty data; Based on the original material and the reference data, an output video is generated by rendering using an AI model; the original material and the AI model are acquired in advance.
2. The video generation and rendering method based on AI technology according to claim 1, characterized in that: The step of obtaining the preference data of the audience user based on the information data of the audience user includes: Based on the information data of the audience users, a first user is obtained; the first user is any user among the audience users; Constructing a FP tree based on the information data of the first user; Based on the FP tree, the preference data of the first user is obtained.
3. The video generation and rendering method based on AI technology according to claim 2 is characterized in that: The step of obtaining the loyalty data of the audience user based on the information data of the audience user includes: Based on the information data of the audience user, a second user is obtained; the second user is any user among the audience users; Based on the information data of the audience user, obtaining the number of clicks by the second user on each video author and the number of likes by the second user on the current video author within a preset time period; Based on each number of clicks, a first number of clicks and a standard deviation of clicks are obtained; the first number of clicks is the number of clicks by the second user on the author of the current video; Obtaining a first coefficient based on the first number of clicks, the click standard deviation, and the number of likes; Based on the first coefficient, loyalty data of the second user is obtained.
4. The video generation and rendering method based on AI technology according to claim 3 is characterized in that: The acquiring the loyalty data of the second user based on the first coefficient includes: Based on the information data of the audience user, a first duration and a second duration are obtained; the first duration is the total duration of the second user watching the video within the preset time period; the second duration is the total duration of the second user watching the video of the author of the current video within the preset time period; Obtaining a second coefficient based on the first duration, the second duration, and the first coefficient; Based on the second coefficient, loyalty data of the second user is obtained.
5. The video generation and rendering method based on AI technology according to claim 4 is characterized in that: The acquiring the loyalty data of the second user based on the second coefficient includes: Based on the information data of the audience user, obtain all video authors followed by the second user; Based on all video authors, a second video author is obtained; the second video author is any video author among all video authors; Based on the author of the second video and the author of the current video, obtaining the similarity of the published videos; Based on the video similarity and the second coefficient, obtaining a third coefficient; Based on the third coefficient, loyalty data of the second user is obtained.
6. The video generation and rendering method based on AI technology according to claim 5, characterized in that: The acquiring the loyalty data of the second user based on the third coefficient includes: Based on the information data of the audience user, the forwarding number is obtained; the forwarding number is the total number of times the second user forwards the video within the preset time period; Based on the number of forwarding times and the first duration, obtaining a profit coefficient; Loyalty data of the second user is obtained based on the profit coefficient and the third coefficient.
7. The video generation and rendering method based on AI technology according to any one of claims 1 to 6, characterized in that: The acquiring of reference data based on the preference data and the loyalty data comprises: Based on the preference data, a first video type is acquired; the first video type is any video type in the preference data; Based on the preference data, obtaining a preference value of each user in the audience for the first video type; Based on the loyalty data, obtaining the loyalty value of each user in the audience user to the current video author; Based on the respective preference values and the respective loyalty values, reference data of the first video type is obtained.
8. The video generation and rendering method based on AI technology according to claim 7, characterized in that: The calculation formula for obtaining the reference data of the first video type based on each preference value and each loyalty value is as follows: ； in, e represents the reference data of the video type; x represents the number of audience users of the current video author; represents the loyalty value of the i-th audience user; Indicates the preference value of the i-th audience user for the video of type e.
9. A video generation and rendering system based on AI technology, characterized in that: include: A reader, for obtaining information data of audience users of the current video author based on big data technology; the information data includes attention information data, click information data, like information data, viewing time data and viewing frequency data of the audience users; The server obtains the preference data of the audience user based on the information data of the audience user; the preference data is used to at least characterize the audience user's preference for the emotion type, emotion intensity, scene change frequency and audio rhythm of the video; And, based on the information data of the audience user, acquiring the loyalty data of the audience user; the loyalty data is at least used to characterize the degree of liking of the audience user for the current video author; and, based on the preference data and the loyalty data, obtaining reference data; And, based on the original material and the reference data, an output video is generated by rendering using an AI model; the original material and the AI model are acquired in advance.
10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, a video generation and rendering method based on AI technology as described in any one of claims 1 to 8 is implemented.