Story text generation method and device, equipment, storage medium and program product

Abstract

The invention provides a story text generation method and device, equipment, a storage medium and a program product, and relates to the technical field of video processing. The method comprises the following steps: acquiring first preference information of a target user for historical video content; screening a target story model matched with the first preference information in a pre-constructed story model knowledge base; wherein a plurality of candidate story models are pre-stored in the story model knowledge base; and generating a story text corresponding to the target user based on the target story model and the first preference information. According to the embodiment of the invention, on the basis of the existing story model, the story text which better accords with the preference tendency of the user can be generated in combination with the preference of the user, and is not limited to deconstruction, re-splicing and the like of the existing story text, so that on one hand, more adaptive content expansion can be realized, and on the other hand, the method can better fit the preference of the user, and the user experience is improved. The personalized requirements of the user are met, the accuracy of the generated story text is improved, and the user experience is further improved.

Description

Technical Field

[0001] The present disclosure relates to the field of computer technology, specifically to the field of video processing technology, and in particular to a story text generation method, device, equipment, storage medium and program product. Background Art

[0002] Existing video recommendation technologies mainly recommend existing video content based on user behaviors (such as likes, viewing time, etc.), but cannot generate new content that targets user preferences. In addition, in related technologies, video generation methods usually require a lot of manual participation, making it difficult to achieve accurate content creation based on user preferences. Summary of the invention

[0003] The embodiments of the present disclosure provide a story text generation method, device, equipment, storage medium and program product.

[0004] In a first aspect, an embodiment of the present disclosure proposes a method for generating a story text, comprising: obtaining first preference information of a target user for historical video content; screening a target story template that matches the first preference information in a pre-constructed story template knowledge base; wherein a plurality of candidate story templates are pre-stored in the story template knowledge base; and generating a story text corresponding to the target user based on the target story template and the first preference information.

[0005] In a second aspect, an embodiment of the present disclosure proposes a story text generation device, comprising: a preference information acquisition module, configured to obtain first preference information of a target user for historical video content; a story template acquisition module, configured to screen a target story template that matches the first preference information in a pre-built story template knowledge base; wherein a plurality of candidate story templates are pre-stored in the story template knowledge base; and a story text generation module, configured to generate a story text corresponding to the target user based on the target story template and the first preference information.

[0006] In a third aspect, an embodiment of the present disclosure provides an electronic device, comprising: 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 are executed by the at least one processor so that when the at least one processor executes, it is able to implement the story text generation method described in any implementation method in the first aspect.

[0007] In a fourth aspect, an embodiment of the present disclosure provides a non-transitory computer-readable storage medium storing computer instructions, which are used to enable a computer to implement the story text generation method described in any implementation method of the first aspect when executed.

[0008] In a fifth aspect, an embodiment of the present disclosure provides a computer program product including a computer program, which, when executed by a processor, can implement the story text generation method described in any implementation manner in the first aspect.

[0009] The story text generation method, device, equipment, storage medium and program product proposed in the embodiments of the present disclosure obtain the first preference information of the historical videos watched by the target user, and select the target story template that meets the target user's preferences in the story template knowledge base according to the first preference information of the target user, and generate a story text corresponding to the target user in the framework of the target story template in combination with the first preference information, so that the story text can better meet the preferences of the target user. In this way, it is possible to generate a story text that is more in line with the user's preferences based on the existing story templates and in combination with the user's preferences, without being limited to the deconstruction and re-splicing of the existing story texts. On the one hand, it can achieve more adaptive content expansion, and on the other hand, it can also be more in line with the user's preferences and meet the user's personalized needs, thereby improving the accuracy of the generated story text and further improving the user experience.

[0010] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present disclosure, nor is it intended to limit the scope of the present disclosure. Other features of the present disclosure will become easily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Other features, objects and advantages of the present disclosure will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0012] Figure 1 is an exemplary system architecture in which the present disclosure may be applied;

[0013] Figure 2 A flowchart of a story text generation method provided by an embodiment of the present disclosure;

[0014] Figure 3 A flowchart of another story text generation method provided by an embodiment of the present disclosure;

[0015] Figure 4 A structural block diagram of a story text generation device provided by an embodiment of the present disclosure;

[0016] Figure 5 A structural block diagram of another story text generation device provided by an embodiment of the present disclosure;

[0017] Figure 6 A schematic diagram of the structure of an electronic device suitable for executing a story text generation method provided in an embodiment of the present disclosure. DETAILED DESCRIPTION

[0018] The following is a description of exemplary embodiments of the present disclosure in conjunction with the accompanying drawings, including various details of the embodiments of the present disclosure to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present disclosure. Similarly, for clarity and conciseness, descriptions of well-known functions and structures are omitted in the following description. It should be noted that the embodiments in the present disclosure and the features in the embodiments may be combined with each other without conflict.

[0019] In the technical solution of the present disclosure, the collection, storage, use, processing, transmission, provision and disclosure of user personal information involved are in compliance with the provisions of relevant laws and regulations and do not violate public order and good morals.

[0020] Figure 1 An exemplary system architecture 100 is shown to which embodiments of the story text generation method, apparatus, electronic device, and computer-readable storage medium disclosed herein can be applied.

[0021] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0022] The user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages, etc. The terminal devices 101, 102, 103 and the server 105 can be installed with various applications for realizing information communication between the two, such as short video platform applications, video platform applications, instant messaging applications, etc.

[0023] Terminal devices 101, 102, 103 and server 105 can be hardware or software. When terminal devices 101, 102, 103 are hardware, they can be various electronic devices with display screens, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, etc.; when terminal devices 101, 102, 103 are software, they can be installed in the electronic devices listed above, which can be implemented as multiple software or software modules, or as a single software or software module, and no specific limitation is made here. When server 105 is hardware, it can be implemented as a distributed server cluster composed of multiple servers, or as a single server; when the server is software, it can be implemented as multiple software or software modules, or as a single software or software module, and no specific limitation is made here.

[0024] It should be noted that, in addition to being obtained from the terminal devices 101, 102, 103 through the network 104, information such as video content and story templates can also be pre-stored locally in the server 105 in various ways. Therefore, when the server 105 detects that such data has been stored locally, it can choose to directly obtain such data locally. In this case, the exemplary system architecture 100 may also not include the terminal devices 101, 102, 103 and the network 104.

[0025] Since processing a video may require more computing resources and stronger computing power, the story text generation method provided in the subsequent embodiments of the present disclosure is generally performed by a server 105 with stronger computing power and more computing resources. Accordingly, the story text generation device is generally also arranged in the server 105. However, it should also be pointed out that when the terminal devices 101, 102, and 103 also have computing power and computing resources that meet the requirements, the terminal devices 101, 102, and 103 can also complete the various operations handed over to the server 105 by components such as the intelligent agent installed thereon, and then output the same result as the server 105, wherein the intelligent agent refers to an artificial intelligence system that uses a neural network model as a core component to build a system capable of handling complex perception, decision-making, and planning tasks (for example, ChatGPT, as a large language model, can be embedded in various intelligent agents, giving the intelligent agents the ability to understand and generate language. This enables the intelligent agent to conduct natural language dialogues and provide information and services to users). Especially when there are multiple terminal devices with different computing capabilities, but when the intelligent agent and other components determine that the terminal device has stronger computing capabilities and more remaining computing resources, the terminal device can be allowed to perform the above-mentioned operations, thereby appropriately reducing the computing pressure of the server 105. Accordingly, the story text generation device can also be set in the terminal devices 101, 102, and 103. In this case, the exemplary system architecture 100 may also not include the server 105 and the network 104.

[0026] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0027] Please refer to Figure 2 , Figure 2 A flowchart of a story text generation method provided by an embodiment of the present disclosure, wherein process 200 includes the following steps:

[0028] Step 201: Obtain the first preference information of the target user for historical video content.

[0029] This step is intended to be performed by the execution subject of the story text generation method (e.g. Figure 1The server 105 shown in the figure obtains the first preference information of the historical videos watched by the target user. Among them, the historical video refers to the video content watched by the target user in a period of time (for example, one day, one week or one month). The first preference information refers to the information used to characterize the preferences of the target user for the video watched. For example, video type (martial arts, romance, action, suspense, etc.), plot characteristics (reversal plot, high-burning scene, touching moment, etc.), character setting (hero, villain, "sweeping monk", etc.), etc. The first preference information can be obtained by analyzing and processing by other servers and terminal devices with analysis functions, and pre-stored in a server or cloud connected to the execution subject, etc. The execution subject can directly obtain it from the server or cloud, or the first preference information can also be obtained by the execution subject based on the content understanding and analysis of the historical video content watched by the target user. The specific process will be described in detail later. It should be noted that the user-related information mentioned in this embodiment, such as the acquisition and use of the first preference information, is known and agreed to by the user.

[0030] Step 202: Filter the target story templates matching the first preference information in the pre-built story template knowledge base.

[0031] This step is intended to allow the above-mentioned execution subject to select a target story template that meets the preferences of the target user in the pre-constructed story template knowledge base according to the first preference information of the target user. Among them, the story template knowledge base pre-stores a plurality of candidate story templates, and the story template refers to a summary text used to extract the content of materials such as videos and movies, which contains elements such as characters, time, place, and plot, for example, it can be a script, outline, story outline, etc. Correspondingly, the story template can also be divided according to type, plot characteristics, character settings, etc. Therefore, the execution subject can select a matching story template from the story template knowledge base according to the first preference information of the target user.

[0032] Step 203: Generate a story text corresponding to the target user based on the target story template and the first preference information.

[0033] The purpose of this step is that after the above-mentioned execution subject obtains the target story template that matches the first preference information, it combines the first preference information and generates a story text corresponding to the target user within the framework of the target story template, so that the story text can better meet the preferences of the target user.

[0034] The story text generation method provided by the disclosed embodiment obtains the first preference information of the historical videos watched by the target user, and selects the target story template that meets the target user's preference in the story template knowledge base according to the first preference information of the target user, and generates a story text corresponding to the target user under the framework of the target story template in combination with the first preference information, so that the story text can better meet the target user's preference. In this way, it is possible to generate a story text that is more in line with the user's preference on the basis of the existing story template and in combination with the user's preference, without being limited to the deconstruction and re-splicing of the existing story text, on the one hand, it can achieve more adaptive content expansion, and on the other hand, it can also be more in line with the user's preferences and meet the user's personalized needs, thereby improving the accuracy of the generated story text and further improving the user experience.

[0035] In some optional implementations of the disclosed embodiments, the above step 201, obtaining the first preference information of the target user for the historical video content, can be obtained by the execution subject through content understanding and analysis based on the historical video content watched by the target user. First, the execution subject obtains the interaction record of the target user, collects the specific interaction operations of the target user in watching the historical videos, including but not limited to likes, favorites, comments, and video viewing positions (start and end points), etc., so as to analyze and obtain the second preference information of the target user. Then, extract the video attribute information of the historical videos watched by the target user, wherein the video attribute information refers to the information used to characterize various aspects of the video content, such as: the theme classification of the video, the theme classification of the videos watched by the target user (such as romance, science fiction, action, etc.); the plot structure of the video, extract the plot features preferred by the target user (such as plot reversal, high-burning scenes, touching moments, etc.); character analysis, analyze the personality characteristics and behavior patterns of the characters in the video, and summarize the specific character settings preferred by the target user (such as strong protagonists, villains, etc.); interest point extraction, based on the viewing behavior of the target user, mark the highlight scenes in the video that can trigger high interaction, and generate interest point tags (such as hilarious clips, conflict scenes, etc.).

[0036] After extracting the video attribute information of the historical video content, the execution subject can generate the first preference information of the target user based on the video attribute information and the second preference information. In some optional implementations of the disclosed embodiments, the video attribute information may include at least one of the theme classification information, plot feature information, character setting information and scene information, and can be combined with at least one of the theme classification information, plot feature information, character setting information and scene information for screening to obtain the first preference information of the target user. For example, if the target user's interactive operations on love video content are likes and favorites, and the interactive operations on horror videos are close and skip, then the first preference information of the target user is love video content, and the first preference information of the target user can be obtained; for another example, if the target user's interactive operations on superhero characters are likes and flowers, and the interactive operations on villain characters are stepping on and bad reviews, then the first preference information of the target user is positive hero characters, and the first preference information of the target user can be obtained. For another example, it is known that the target user has a large collection of videos about sea scenes and deep diving, and likes the most videos of a senior diving master XXX, then the first preference information of the target user is the deep diving video of the senior diving master XXX.

[0037] In practical applications, the historical video content may be a complete movie of a long duration, or a short video of a short duration. Regardless of the type of video, the video may not only contain a climax plot or a reversal plot or other plots that are more likely to attract the audience's attention. Therefore, the target user's interactive operation on the historical video content may appear multiple times in a video content, and the interactive tendency is also different. Therefore, in this embodiment, based on the target user's interactive operation on the historical video content, the process of generating the target user's second preference information mainly includes: dividing the historical video content into multiple video clips, obtaining the target user's interactive operation on each video clip; and generating the second preference information based on the type of interactive operation. The historical video content is divided into multiple video clips, mainly to be able to divide the multiple climax points, reversal points or other plots, characters, and places that may appear in the video, and to analyze and obtain the target user's preferences for different plots, characters, places, etc. from a more detailed perspective, so as to more accurately identify the target user's second preference information.

[0038] The specific second preference information of the target user can be divided from the dimension of positive or negative. For example, if the target user's interactive operations such as likes, collections, and flowers are obtained for the video, it usually indicates that the target user's interactive operation on the video content is a positive behavior; for the video comment content, such as "good-looking, classic, wonderful, worth watching, fascinating", etc., it usually indicates that the target user's behavior on the video content is a positive behavior. Relatively speaking, if the target user's interactive operations such as swiping, closing, and stepping on the video are obtained, it usually indicates that the target user's interactive operation on the video content is a negative behavior; for the video comment content, such as "not good-looking, bad movie, waste of time", etc., it usually indicates that the target user's interactive operation on the video content is a negative behavior. It should be understood that the user's second preference information can also be divided from other dimensions, such as summarizing the number and frequency of different types of videos watched by the target user to reflect the target user's preference for different types of videos, and this embodiment is not limited to this.

[0039] After obtaining the target user's interactive operation, the target user's second preference information can be determined according to the type of the interactive operation. For example, through positive behavior, it can be positively and directly determined that the target user's behavior belongs to the second preference information; through negative behavior, it can be reversely determined that the target user's behavior does not belong to the second preference information, so it can be excluded; or, it can be determined that the target user's behavior belongs to the second preference information through positive behavior, and combined with negative behavior to further verify whether the behavior belongs to the second preference information.

[0040] Through the above process, the historical video content is divided, and for each divided video, the target user's preferences for different plots, characters, locations, etc. can be analyzed from a more detailed perspective to obtain the target user's second preference information, thereby achieving more accurate identification of the target user's second preference information.

[0041] In some optional implementations of the disclosed embodiments, the execution subject extracts the video attribute information of the historical video content, generates the first preference information according to the video attribute information and the second preference information, and can also obtain the first preference information based on the video attribute information and the second preference information through a preset user preference model. Among them, the user preference model is a multimodal interest model of user preferences that is relatively mature in the relevant field, and is a model that uses multiple types of data (i.e., multimodal data) to understand and model user interests. Among them, multimodal data refers to data containing multiple media information, such as text, audio, video, images, etc. In this embodiment, the model is mainly trained based on the text features, visual features and audio features of the historical video content. These data can come from different data sources, such as social media, audio and video files, etc. Exemplarily, the text feature can be extracted by analyzing the video title, introduction, barrage and comments through natural language processing technology, and extracting the core keywords and emotional tendencies of the text content. Image features are generated by combining computer vision technology to analyze scene switching, character expressions and actions in the video. Voice features are obtained by analyzing the voice content in the video through voice recognition technology and extracting key information (such as classic lines and background sound effects).

[0042] In some optional implementations of the disclosed embodiments, the pre-constructed story template knowledge base described in the above step 202 may be constructed by the execution subject through the following process: First, extract the character setting information, plot development mode information and rhythm characteristic information of different types of story samples. Among them, the story sample refers to the story text used as sample data, and the story text may include various forms, such as video scripts, outlines, short stories, etc. Different types of story samples are obtained, and their structural characteristics are extracted. The structural characteristics may include, for example: character setting (protagonist-supporting role-antagonist), plot development mode (prelude-climax-reversal) and rhythm characteristics (compact-soothing-ups and downs), etc. It can be understood that the character setting, plot development mode and rhythm characteristics of different types of story samples may have similarities or differences. Then, different types of candidate story templates are established based on the character setting information, plot development mode information and rhythm characteristic information. For example, for a martial arts story template, the character setting usually includes a righteous character and a villain, the plot development mode usually includes a foreshadowing part, a climax part, and an ending part, and the rhythm characteristics are usually compact. Based on these structural characteristics, a martial arts story template can be established. Similarly, different types of candidate story templates can be established for different types of stories (romance, science fiction, suspense, etc.). Then, the story template knowledge base can be constructed through the established different types of candidate story templates.

[0043] For the story template knowledge base described in the above embodiment, the execution subject can filter the required target story template from it in a variety of ways. As one implementation method, the execution subject can be an intelligent agent with a large language model embedded in it, through which the target story template that matches the first preference information can be filtered from the story template knowledge base. For the intelligent agent, a specific prompt text (prompt) can be provided by an administrator to indicate the specific content of the first preference information to be obtained by the intelligent agent, for example, "extract the script full of suspense and reversal that the target user likes". After the execution subject obtains the prompt text, the corresponding content information can be extracted according to the prompt text, and the target story template that matches the content information is obtained in the story template knowledge base. As another optional implementation method, the execution subject can also be a server, processor, etc. that can call a natural language processing model (NLP), which can perform semantic understanding according to the user's first preference information, and match the story templates in the story template knowledge base, so as to obtain the target story template that matches the first preference information.

[0044] After obtaining the target story template, the execution subject executes the above step 203, and generates a story text corresponding to the target user according to the target story template and the first preference information. In this embodiment, it means that a story text matching the first preference information of the target user can be generated. Specifically, the first preference information can include at least one of type feature information, emotional feature information and rhythm feature information. The process of the execution subject generating the story text is that at least one of the type feature information, emotional feature information and rhythm feature information in the first preference information can be correspondingly embedded into the target story template according to the target story template, thereby generating a story text. Exemplarily, the first preference information of the target user includes: martial arts (type feature), clear love and hate (emotional feature), compact plot (plot feature), and correspondingly, the execution subject can generate a story text according to the obtained martial arts target story template, based on the character setting, emotional setting, etc., combined with the first preference information of the target user. It can be understood that the user's first preference information can also include only one of the type feature information, emotional feature information, and plot feature information, or two of them, and this embodiment is not limited to this. Furthermore, the story text generated by the execution entity is not limited to generating only one story text, but more than one story text can be generated according to the user's first preference information for the target user to make further selections.

[0045] In the above embodiments, some embodiments are proposed for extracting preferences for the historical video content watched by the target user, so as to generate a story text that better matches the user's preferences. In actual applications, the user's preferences may also fluctuate to a certain extent according to the different situations and circumstances at the time. Considering this situation, in this embodiment, the execution subject can further obtain the story elements selected by the target user in real time to adjust the generated story text. The story element can refer to the included theme classification information, plot feature information, character setting information and scene information. The story element selected by the target user may be different from the information preferred by the target user. Therefore, if the execution subject obtains that the target user has selected a specific story element, it is determined that the story element is the content that the target user is currently more concerned about. Therefore, the story element selected by the target user can be inserted into the story text to update the story text. The insertion process can be to add the story element selected by the target user as an additional element to the target story template to update the story text, or to replace the original first preference information with the story element selected by the target user to update the story text based on the target story template. This embodiment is not limited to this. Through the above process, the content can be more in line with the current preferences of the target users and better match the preferences of the target users, which improves the accuracy of the generated story text and further enhances the user experience.

[0046] Please refer to Figure 3 , Figure 3 This is a flowchart of another story text generation method provided by an embodiment of the present disclosure, wherein process 300 includes the following steps:

[0047] Step 301: Obtain the first preference information of the target user for historical video content.

[0048] Step 302: Filter a target story template that matches the first preference information in a pre-built story template knowledge base; wherein the story template knowledge base pre-stores a plurality of candidate story templates.

[0049] Step 303: Generate a story text corresponding to the target user based on the target story template and the first preference information.

[0050] The above steps 301-303 are similar to Figure 2 Steps 201-203 shown are consistent, and for the same contents, please refer to the corresponding parts of the previous embodiment, which will not be repeated here.

[0051] Step 304: Generate a target video based on the story text.

[0052] In this embodiment, the story text generated by the story text generation method described in any of the above embodiments can be further used as a script or outline. Based on this, the execution subject can generate a target video according to the story text, and can push the target video to the target user, and further generate video content including images, voices, special effects, etc. from the text content. Exemplarily, the execution subject can use a relatively mature artificial intelligence algorithm (such as Stable Diffusion or GAN) in the relevant field to generate the target video. Through this process, it is possible to provide users with richer sharing content; and the video content is constructed based on the generated story text, which is combined with the user's preferences to generate a story text that is more in line with the user's preferences, and is not limited to the deconstruction and re-splicing of the existing story text. On this basis, the video content generated can achieve more adaptive content expansion on the one hand, and on the other hand, it can also be more in line with the user's preferences and meet the user's personalized needs, thereby improving the accuracy of the generated story text and further improving the user experience. It should be noted that the story text described in this embodiment and any of the embodiments below may be the story text generated based on the target story template and the user's first preference information as described in any of the above embodiments, or may be the story text updated based on the user's further selection, and this embodiment is not limited to this.

[0053] Furthermore, the generated video content can also be updated and adjusted according to the feedback of the target user. In some optional implementations of the disclosed embodiments, the execution subject can further obtain the feedback information of the target user for the target video, and then update the first preference information based on the feedback information to obtain the updated third preference information; and then update the story text and the target video based on the updated third preference information. Among them, the feedback information includes but is not limited to: the target user's interactive operations on the pushed target video, such as: likes, favorites, comments, shares, etc. When the execution subject obtains the feedback information, it indicates that the target user has a preference for the target video. Therefore, the execution subject can further update the first preference information of the target user according to the feedback information, and update the corresponding story text and target video according to the updated third preference information, so as to generate video content that is more in line with the preferences of the target user, better meet the needs of the user, improve the accuracy of the generated story text, and further improve the user experience.

[0054] In practical applications, the target video generated by the execution subject in the process of executing the above embodiment may include multiple characters, plots and climax points, etc., all of which can be included in the story text based on the target video as one or more plot node information. Based on this, each plot node information in the story text may also be used as a separate story point to form independent video content. In some implementation methods of the present embodiment, the execution subject can extract at least one plot node from the story text, and according to the plot node information, the target video is segmented, so that the target video is divided into multiple video clips. Through this process, the duration of the target video can be adjusted according to actual needs (which may be user needs or the needs of the playback platform, etc.), and a video content that may be longer in time is divided into multiple video contents with shorter time, and in order to ensure the continuity of the content between each video clip, a small section of the adjacent video clip content (for example, 1-2 seconds) can also be included at the beginning and end of each video, so that the content of two adjacent videos can be better connected. Through the above process, since the target video is divided based on the plot node information, each divided video segment can contain at least one plot node, thereby ensuring the integrity of the video segment content, and also avoiding the problem of incoherent content of some video segments that may be caused by mechanically dividing the video only by time periods.

[0055] In actual applications, there may be certain limitations in the user's reception of video content, and the user needs to face the screen of the display device specifically to view the corresponding content. Based on this, in this embodiment, a further implementation method is proposed, which can generate the voice commentary content of the target video based on the story text. Exemplarily, the execution subject can use a relatively mature artificial intelligence algorithm in the relevant field (such as TTS (Text-to-Speech) technology) to generate the voice commentary content. Through this process, the user is not limited to obtaining the corresponding information by watching the video content displayed on the screen, but can directly obtain the description content for the target video by listening to the voice playback, which expands the ways and means for users to receive information, and can also help users better understand the background and atmosphere of the video content.

[0056] In order to further improve the diversity of the generated video content, the execution entity can also obtain the common features of the reference video and modify the target video content based on the common features. Among them, the reference video can be a preset number of videos determined by ranking according to preset indicators. For example, the features of some videos (such as the top three, top five, etc.) in the popularity list and other rankings ranked according to the number of user views. The common features can, for example, be to analyze the common features such as the playback effects, dubbing characteristics, and music style of such videos, and modify the style, special effects, etc. of the generated target video based on such common features, thereby improving the user's interest, improving the accuracy of the generated story text, and further improving the user experience.

[0057] Further references Figure 4 As an implementation of the methods shown in the above figures, the present disclosure provides an embodiment of a story text generation device, and the device embodiment is Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0058] like Figure 4 As shown, the story text generation device 400 of this embodiment may include: a preference information acquisition module 401, a story template acquisition module 402, and a story text generation module 403. Among them, the preference information acquisition module 401 is configured to obtain the first preference information of the target user for the historical video content. The story template acquisition module 402 is configured to filter the target story template that matches the first preference information in a pre-constructed story template knowledge base, wherein a plurality of candidate story templates are pre-stored in the story template knowledge base. The story text generation module 403 is configured to generate a story text corresponding to the target user based on the target story template and the first preference information.

[0059] In the present embodiment, the specific processing of the preference information acquisition module 401, the story template acquisition module 402, and the story text generation module 403 and the technical effects thereof can be referred to in the respective Figure 2 The relevant descriptions of steps 201 - 203 in the corresponding embodiment are not repeated here.

[0060] In some optional implementations of this embodiment, the preference information acquisition module 401 includes: a second preference information generation submodule, configured to generate the second preference information of the target user based on the target user's interactive operation on the historical video content; a video attribute information extraction submodule, configured to extract the video attribute information of the historical video content; and a preference information generation submodule, configured to generate the first preference information according to the video attribute information and the second preference information. Among them, the second preference information generation submodule is configured to obtain the target user's interactive record, collect the target user's specific interactive operations on the historical video, including but not limited to likes, collections, comments, and video viewing positions (start and end points), so as to analyze and obtain the second preference information of the target user. The video attribute information extraction submodule is configured to extract the video attribute information of the historical video watched by the target user, wherein the video attribute information refers to information used to characterize various aspects of the video content, for example, it may include: the theme classification of the video, the plot structure of the video, character analysis, interest point extraction, etc. The preference information generation submodule can generate the first preference information of the target user according to the second preference information. In some optional implementations of the embodiments of the present disclosure, the video attribute information may include at least one of theme classification information, plot feature information, character setting information and scene information, and may be combined with at least one of the theme classification information, plot feature information, character setting information and scene information to obtain the first preference information of the target user.

[0061] In some optional implementations of this embodiment, the second preference information generation submodule includes: a video division unit configured to divide the historical video content into multiple video clips, an interactive operation acquisition unit configured to acquire the target user's interactive operations on each video clip; and a second preference information generation unit configured to generate second preference information based on the type of interactive operation. The historical video content is divided into multiple video clips, mainly to be able to divide multiple climax points, reversal points or other plots, characters, and locations that may appear in the video, and to analyze and acquire the target user's preferences for different plots, characters, locations, etc. from a more detailed perspective, so as to more accurately identify the target user's second preference information.

[0062] The specific second preference information of the target user can be divided from the dimension of positive or negative. For example, if the target user's interactive operations such as likes, collections, and flowers are obtained for the video, it usually indicates that the target user's interactive operation on the video content is a positive behavior; for the video comment content, such as "good-looking, classic, wonderful, worth watching, fascinating", etc., it usually indicates that the target user's behavior on the video content is a positive behavior. Relatively speaking, if the target user's interactive operations such as swiping, closing, and stepping on the video are obtained, it usually indicates that the target user's interactive operation on the video content is a negative behavior; for the video comment content, such as "not good-looking, bad movie, waste of time", etc., it usually indicates that the target user's interactive operation on the video content is a negative behavior. It should be understood that the user's second preference information can also be divided from other dimensions, such as summarizing the number and frequency of different types of videos watched by the target user to reflect the target user's preference for different types of videos, and this embodiment is not limited to this.

[0063] After obtaining the target user's interactive operation, the target user's second preference information can be determined according to the type of the interactive operation. For example, through positive behavior, it can be positively and directly determined that the target user's behavior belongs to the second preference information; through negative behavior, it can be reversely determined that the target user's behavior does not belong to the second preference information, so it can be excluded; or, it can be determined that the target user's behavior belongs to the second preference information through positive behavior, and combined with negative behavior to further verify whether the behavior belongs to the second preference information.

[0064] Through the above process, the historical video content is divided, and for each divided video, the target user's preferences for different plots, characters, locations, etc. can be analyzed from a more detailed perspective to obtain the target user's second preference information, thereby achieving more accurate identification of the target user's second preference information.

[0065] In some optional implementations of this embodiment, the preference information generation submodule is further configured to extract at least one of the theme classification information, plot feature information, character setting information and scene information of the historical video content, and based on the second preference information, filter at least one of the theme classification information, plot feature information, character setting information and scene information to obtain the first preference information.

[0066] In some optional implementations of this embodiment, the preference information generation submodule is further configured to: obtain the first preference information based on the video attribute information and the second preference information through a preset user preference model; wherein the preset user preference model is obtained by training based on the text features, visual features and audio features of the historical video content. Among them, the user preference model is a multimodal interest model of user preferences that is relatively mature in the relevant field, and is a model that uses multiple types of data (i.e., multimodal data) to understand and model user interests. Among them, multimodal data refers to data containing multiple media information, such as text, audio, video, images, etc. In this embodiment, the model is mainly obtained by training based on the text features, visual features and audio features of the historical video content. These data can come from different data sources, such as social media, audio and video files, etc. Exemplarily, the text feature can be extracted by analyzing the video title, introduction, barrage and comments through natural language processing technology. The image feature is to analyze the scene switching, character expressions and actions in the video in combination with computer vision technology to generate image feature labels. Voice features are obtained by analyzing the voice content in the video through voice recognition technology and extracting key information (such as classic lines and background sound effects).

[0067] In some optional implementations of the embodiments of the present disclosure, the story text generation device also includes: a story sample extraction module, configured to extract character setting information, plot development pattern information and rhythm characteristic information of different types of story samples; a story template establishment module, configured to establish different types of candidate story templates based on the character setting information, plot development pattern information and rhythm characteristic information; a story template knowledge base construction module, configured to construct a story template knowledge base based on different types of candidate story templates.

[0068] In some optional implementations of the disclosed embodiments, the story template acquisition module 402 includes: a prompt word acquisition submodule, configured to acquire a prompt word text, the prompt word text is used to indicate the specific content of the first preference information to be acquired; a content extraction submodule, configured to extract the content information in the prompt word text; and a story template acquisition submodule, configured to acquire a target story template that matches the content information in the story template knowledge base. As one implementation, the execution subject can be an intelligent agent with a large language model embedded in it, through which the target story template that matches the first preference information can be screened from the story template knowledge base. For the intelligent agent, the administrator can provide a specific prompt word text (prompt) to indicate the specific content of the first preference information to be acquired by the intelligent agent, for example, "extract the script full of suspense and reversal that the target user likes". After the execution subject acquires the prompt word text, the corresponding content information can be extracted according to the prompt word text, and the target story template that matches the content information can be acquired in the story template knowledge base. As another optional implementation, the execution entity may also be a server, processor, etc. that can call a natural language processing model (NLP), which can perform semantic understanding based on the user's first preference information and match story templates in a story template knowledge base to obtain a target story template that matches the first preference information.

[0069] In some optional implementations of the disclosed embodiment, the story text generation module 403 is further configured to: match at least one of the type feature information, emotional feature information and rhythm feature information with the elements in the target story template to generate a story text. Specifically, the first preference information may include at least one of the type feature information, emotional feature information and rhythm feature information. The process of the execution subject generating the story text is that at least one of the type feature information, emotional feature information and rhythm feature information in the first preference information can be correspondingly embedded into the target story template according to the target story template, thereby generating a story text. It is understandable that the user's first preference information may also include only one of the type feature information, emotional feature information, and plot feature information, or two of them, and the present embodiment is not limited to this. In addition, the story text generated by the execution subject is not limited to generating only one story text, but can generate more than one story text according to the user's first preference information for the target user to make further selections.

[0070] In some optional implementations of the disclosed embodiments, the story text generation device further includes: a story text update module, which is configured to insert the story element into the story text in response to obtaining the story element selected by the target user to obtain an updated story text. In practical applications, the user's preferences may also fluctuate to a certain extent according to the different scenarios and circumstances at the time. Considering this situation, in this embodiment, the story text update module can further obtain the story element selected by the target user in real time to adjust the generated story text. The story element may refer to the included theme classification information, plot feature information, character setting information and scene information. The story element selected by the target user may be different from the information preferred by the target user. Therefore, if the story text update module obtains that the target user has selected a specific story element, it is determined that the story element is the content that the target user is currently more concerned about. Therefore, the story element selected by the target user can be inserted into the story text to update the story text. The insertion process may be to add the story element selected by the target user as an additional element to the target story template to update the story text, or to replace the original first preference information with the story element selected by the target user to update the story text based on the target story template, and this embodiment is not limited thereto. Through the above process, the current preferred content of the target user can be more closely matched with the preferences of the target user, which can further enhance the user's interest, improve the accuracy of the generated story text, and further enhance the user experience.

[0071] Further references Figure 5 As an implementation of the methods shown in the above figures, the present disclosure provides another embodiment of a story text generation device, which is similar to Figure 3 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0072] like Figure 5 As shown, the story text generation device 500 of this embodiment may include: a preference information acquisition module 501, a story template acquisition module 502, a story text generation module 503 and a video generation module 504. Among them, the preference information acquisition module 501 is configured to obtain the first preference information of the target user for the historical video content. The story template acquisition module 502 is configured to screen the target story template that matches the first preference information in a pre-constructed story template knowledge base, wherein a plurality of candidate story templates are pre-stored in the story template knowledge base. The story text generation module 503 is configured to generate a story text corresponding to the target user based on the target story template and the first preference information. The video generation module 504 is configured to generate a target video based on the story text.

[0073] In this embodiment, the specific processing and technical effects of the preference information acquisition module 501, the story template acquisition module 502, the story text generation module 503 and the video generation module 504 in the story text generation device 500 can be referred to respectively. Figure 3 The relevant descriptions of steps 301 - 304 in the corresponding embodiment are not repeated here.

[0074] In this embodiment, the story text described in any of the above embodiments can be further used as a script or outline. Based on this, the video generation module 504 can generate a target video according to the story text, and can push the target video to the target user, and further generate the text content into video content including images, voices, special effects, etc. Exemplarily, the video generation module 504 can use a relatively mature artificial intelligence algorithm (such as Stable Diffusion or GAN) in the relevant field to generate the target video. Through this process, it is possible to provide users with richer sharing content; and the video content is constructed based on the generated story text, which is combined with the user's preferences to generate a story text that is more in line with the user's preferences, and is not limited to the deconstruction and re-splicing of the existing story text. On this basis, the video content generated can achieve more adaptive content expansion on the one hand, and on the other hand, it can also be more in line with the user's preferences and meet the user's personalized needs, thereby improving the accuracy of the generated story text and further improving the user experience. It should be noted that the story text described in this embodiment and any of the embodiments below may be the story text generated based on the target story template and the user's first preference information as described in any of the above embodiments, or may be the story text updated based on the user's further selection, and this embodiment is not limited to this.

[0075] In some optional implementations of the disclosed embodiments, the story text generation device 500 includes: a feedback information acquisition module, configured to obtain feedback information of the target user for the target video. A preference information update module, configured to update the first preference information based on the feedback information to obtain the updated third preference information. The story and video update module, configured to update the story text and the target video based on the updated third preference information. Wherein, the feedback information includes but is not limited to: the target user's interactive operations on the pushed target video, such as: like, favorite, comment, share, etc. When the feedback information acquisition module obtains the feedback information, it indicates that the target user has a preference for the target video. Therefore, the preference information update module can further update the first preference information of the target user according to the feedback information, and the story and video update module can update the corresponding story text and target video according to the updated third preference information, thereby realizing the generation of video content that is more in line with the preferences of the target user, more in line with the needs of the user, improving the accuracy of the generated story text, and further improving the user experience.

[0076] In some optional implementations of the disclosed embodiments, the story text generation device 500 includes: a plot extraction module, configured to extract at least one plot node information in the story text; a video splitting module, configured to decompose the target video based on the plot node information to obtain multiple video clips. Through this process, the duration of the target video can be adjusted according to actual needs (which may be user needs or the needs of the playback platform, etc.), and a video content that may be longer in time can be divided into multiple video contents with shorter time. In addition, in order to ensure the continuity of the content between each video clip, a small section of the adjacent video clip content (for example, 1-2 seconds) can also be included at the beginning and end of each video, so that the content of the two adjacent videos can be better connected. Through the above process, since the target video is divided based on the plot node, the divided video clips can contain at least one plot node, thereby ensuring the integrity of the video clips, and can also avoid the problem of incoherent content of some video clips that may be caused by mechanically dividing the video only by time periods.

[0077] In some optional implementations of the disclosed embodiments, the story text generation device 500 includes: a speech generation module, which is configured to generate speech narration content of the target video based on the story text. In practical applications, there may be certain limitations in the user's reception of video content, and the user needs to face the screen of the display device specifically to view the corresponding content. Based on this, in this embodiment, a further implementation method is proposed, which can generate the speech narration content of the target video based on the story text. Exemplarily, the execution subject can use a relatively mature artificial intelligence algorithm (such as TTS (Text-to-Speech) technology) in the relevant field to generate the speech narration content. Through this process, the user is not limited to obtaining the corresponding information by watching the video content displayed on the screen, but can directly obtain the description content for the target video by listening to the voice playback, which expands the ways and means for the user to receive information, and can also help the user better understand the background and atmosphere of the video content.

[0078] In some optional implementations of the disclosed embodiments, the story text generation device 500 includes: a video analysis module, configured to obtain common features of reference videos, the reference videos including: a preset number of videos determined by ranking according to preset indicators; a video modification module, configured to modify the target video based on the common features. Among them, the reference video can be a preset number of videos determined by ranking according to preset indicators. For example, the features of some videos (such as the top three, top five, etc.) in the popularity list and other rankings ranked according to the number of user views. The common features can be, for example, analyzing the common features such as the playback special effects, dubbing characteristics, and music style of such videos, and modifying the style, special effects, etc. of the generated target video based on such common features, thereby further improving the user's interest, improving the accuracy of the generated story text, and further improving the user experience.

[0079] This embodiment exists as a device embodiment corresponding to the above method embodiment. The story text generation device provided by this embodiment obtains the first preference information of the historical videos watched by the target user, and selects the target story template that meets the target user's preferences in the story template knowledge base according to the first preference information of the target user. In combination with the first preference information, a story text corresponding to the target user is generated within the framework of the target story template, so that the story text can better meet the target user's preferences. In this way, on the basis of the existing story templates, combined with the user's preferences, a story text that is more in line with the user's preferences can be generated, without being limited to the deconstruction and re-splicing of the existing story text. On the one hand, it can achieve more adaptive content expansion, and on the other hand, it can also be more in line with the user's preferences and meet the user's personalized needs, thereby improving the accuracy of the generated story text and further improving the user experience.

[0080] According to an embodiment of the present disclosure, the present disclosure also provides an electronic device, which includes: 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 are executed by the at least one processor so that the story text generation method described in any of the above embodiments can be implemented when the at least one processor executes.

[0081] According to an embodiment of the present disclosure, the present disclosure further provides a readable storage medium, which stores computer instructions, and the computer instructions are used to enable a computer to implement the story text generation method described in any of the above embodiments when executed.

[0082] According to an embodiment of the present disclosure, the present disclosure further provides a computer program product, which, when executed by a processor, can implement the story text generation method described in any of the above embodiments.

[0083] Figure 6 A schematic block diagram of an example electronic device 600 that can be used to implement an embodiment of the present disclosure is shown. The electronic device is intended to represent various forms of digital computers, such as laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as personal digital processing, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present disclosure described and / or required herein.

[0084] like Figure 6 As shown, the device 600 includes a computing unit 601, which can perform various appropriate actions and processes according to a computer program stored in a read-only memory (ROM) 602 or a computer program loaded from a storage unit 608 into a random access memory (RAM) 603. In the RAM 603, various programs and data required for the operation of the device 600 can also be stored. The computing unit 601, the ROM 602, and the RAM 603 are connected to each other via a bus 604. An input / output (I / O) interface 605 is also connected to the bus 604.

[0085] A number of components in the device 600 are connected to the I / O interface 605, including: an input unit 606, such as a keyboard, a mouse, etc.; an output unit 607, such as various types of displays, speakers, etc.; a storage unit 608, such as a disk, an optical disk, etc.; and a communication unit 609, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 609 allows the device 600 to exchange information / data with other devices through a computer network such as the Internet and / or various telecommunication networks.

[0086] The computing unit 601 may be a variety of general and / or special processing components with processing and computing capabilities. Some examples of the computing unit 601 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various computing units running machine learning model algorithms, digital signal processors (DSPs), and any appropriate processors, controllers, microcontrollers, etc. The computing unit 601 performs the various methods and processes described above, such as the story text generation method. For example, in some embodiments, the story text generation method may be implemented as a computer software program, which is tangibly contained in a machine-readable medium, such as a storage unit 608. In some embodiments, part or all of the computer program may be loaded and / or installed on the device 600 via ROM 602 and / or communication unit 609. When the computer program is loaded into RAM 603 and executed by the computing unit 601, one or more steps of the story text generation method described above may be performed. Alternatively, in other embodiments, the computing unit 601 may be configured to perform the story text generation method by any other appropriate means (e.g., by means of firmware).

[0087] Various implementations of the systems and techniques described above herein can be implemented in digital electronic circuit systems, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), load programmable logic devices (CPLDs), computer hardware, firmware, software, and / or combinations thereof. These various implementations can include: being implemented in one or more computer programs that can be executed and / or interpreted on a programmable system including at least one programmable processor, which can be a special purpose or general purpose programmable processor that can receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device.

[0088] The program code for implementing the method of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general-purpose computer, a special-purpose computer, or other programmable data processing device, so that the program code, when executed by the processor or controller, enables the functions / operations specified in the flow chart and / or block diagram to be implemented. The program code may be executed entirely on the machine, partially on the machine, partially on the machine and partially on a remote machine as a stand-alone software package, or entirely on a remote machine or server.

[0089] In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0090] To provide interaction with a user, the systems and techniques described herein can be implemented on a computer having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and pointing device (e.g., a mouse or trackball) through which the user can provide input to the computer. Other types of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, voice input, or tactile input).

[0091] The systems and techniques described herein may be implemented in a computing system that includes back-end components (e.g., as a data server), or a computing system that includes middleware components (e.g., an application server), or a computing system that includes front-end components (e.g., a user computer with a graphical user interface or a web browser through which a user can interact with implementations of the systems and techniques described herein), or a computing system that includes any combination of such back-end components, middleware components, or front-end components. The components of the system may be interconnected by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: a local area network (LAN), a wide area network (WAN), and the Internet.

[0092] A computer system may include a client and a server. The client and the server are generally remote from each other and usually interact through a communication network. The relationship between the client and the server is generated by computer programs running on the corresponding computers and having a client-server relationship with each other. The server may be a cloud server, also known as a cloud computing server or cloud host, which is a host product in the cloud computing service system to solve the defects of difficult management and weak business scalability in traditional physical hosts and virtual private servers (VPS) services.

[0093] According to the technical solution of the embodiment of the present disclosure, the first preference information of the historical videos watched by the target user is obtained, and according to the first preference information of the target user, the target story template that meets the preferences of the target user is screened in the story template knowledge base, and combined with the first preference information, a story text corresponding to the target user is generated within the framework of the target story template, so that the story text can better meet the preferences of the target user. In this way, on the basis of the existing story templates, combined with the user's preferences, a story text that better meets the user's preferences can be generated, without being limited to the deconstruction and re-splicing of the existing story texts. On the one hand, more adaptive content expansion can be achieved, and on the other hand, it can also better fit the user's preferences and meet the user's personalized needs, thereby improving the accuracy of the generated story text and further improving the user experience.

[0094] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps recorded in this disclosure can be executed in parallel, sequentially or in different orders, as long as the desired results of the technical solutions disclosed in this disclosure can be achieved, and this document does not limit this.

[0095] The above specific implementations do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.

Claims

1. A method for generating a story text, comprising: Obtain the target user's first preference information for historical video content; Screening a target story template that matches the first preference information in a pre-built story template knowledge base; wherein the story template knowledge base pre-stores a plurality of candidate story templates; A story text corresponding to the target user is generated based on the target story template and the first preference information.

2. The method according to claim 1, wherein: The obtaining of the first preference information of the target user for the historical video content includes: generating second preference information of the target user based on the interactive operation of the target user on the historical video content; Extracting video attribute information of the historical video content; The first preference information is generated according to the video attribute information and the second preference information.

3. The method according to claim 2, wherein: The generating the second preference information of the target user based on the interactive operation of the target user on the historical video content includes: Dividing the historical video content into a plurality of video clips; Acquiring the interactive operation of the target user on each of the video clips; The second preference information is generated based on the type of the interactive operation.

4. The method according to claim 2, wherein: The video attribute information includes: at least one of: theme classification information, plot feature information, character setting information and scene information; The generating the first preference information according to the video attribute information and the second preference information includes: The video attribute information is filtered based on the second preference information to obtain the first preference information.

5. The method according to claim 2, wherein: The generating the first preference information according to the video attribute information and the second preference information includes: The first preference information is obtained based on the video attribute information and the second preference information through a preset user preference model; wherein the preset user preference model is trained based on the text features, visual features and audio features of the historical video content.

6. The method according to claim 1, further comprising: Extract character setting information, plot development pattern information and rhythm characteristics information of different types of story samples; Based on the character setting information, the plot development mode information and the rhythm characteristic information, obtaining different types of candidate story templates; The story template knowledge base is constructed based on the candidate story templates.

7. The method according to claim 1, wherein: The step of selecting a target story template matching the first preference information from a pre-built story template knowledge base includes: Acquire a prompt word text, where the prompt word text is used to indicate specific content of the first preference information to be acquired; Extracting content information from the prompt word text; The target story template matching the content information is obtained from the story template knowledge base.

8. The method according to claim 1, wherein: The first preference information includes at least one of type feature information, emotion feature information and rhythm feature information. The generating a story text corresponding to the target user based on the target story template and the first preference information includes: The first preference information is matched with elements in the target story template to generate the story text.

9. The method according to any one of claims 1 to 8, further comprising: A target video is generated based on the story text.

10. The method according to claim 9, further comprising: Obtaining feedback information of the target user regarding the target video; updating the first preference information based on the feedback information to obtain updated third preference information; The story text and the target video are updated based on the third preference information.

11. The method according to claim 9, further comprising: Extracting at least one plot node information from the story text; The target video is decomposed based on the plot node information to obtain multiple video clips.

12. The method according to claim 9, further comprising: Generate voice narration content of the target video based on the story text.

13. The method according to claim 1, further comprising: In response to acquiring the story element selected by the target user, the story element is inserted into the story text to obtain an updated story text.

14. The method according to claim 13, further comprising: A target video is generated based on the updated story text.

15. The method according to claim 14, further comprising: Obtaining feedback information of the target user regarding the target video; updating the first preference information based on the feedback information to obtain updated third preference information; The updated story text and the target video are updated based on the third preference information.

16. The method according to claim 14, further comprising: Extracting at least one plot node information from the updated story text; The target video is decomposed based on the plot node information to obtain multiple video clips.

17. The method according to claim 14, further comprising: Generate voice narration content for the target video based on the updated story text.

18. The method according to claim 9 or 14, further comprising: Acquire common features of reference videos, wherein the reference videos include: a preset number of videos ranked according to preset indicators; The target video is modified based on the common features.

19. A story text generating device, comprising: A preference information acquisition module, configured to acquire first preference information of a target user for historical video content; A story template acquisition module is configured to screen a target story template that matches the first preference information from a pre-built story template knowledge base; wherein the story template knowledge base pre-stores a plurality of candidate story templates; The story text generation module is configured to generate a story text corresponding to the target user based on the target story template and the first preference information.

20. The device according to claim 19, wherein The preference information acquisition module includes: A second preference information generating submodule is configured to generate second preference information of the target user based on the interactive operation of the target user on the historical video content; A video attribute information extraction submodule, configured to extract video attribute information of the historical video content; The preference information generating submodule is configured to generate the first preference information according to the video attribute information and the second preference information.

21. The device according to claim 20, wherein: The second preference information generating submodule includes: a video division unit, configured to divide the historical video content into a plurality of video segments; an interactive operation acquisition unit, configured to acquire the interactive operation of the target user on each of the video clips; The second preference information generating unit is configured to generate the second preference information based on the type of the interactive operation.

22. The device according to claim 20, wherein: The video attribute information includes: at least one of: theme classification information, plot feature information, character setting information and scene information; The preference information generation submodule is further configured to filter the video attribute information based on the second preference information to obtain the first preference information.

23. The device according to claim 20, wherein: The preference information generation submodule is further configured to: The first preference information is obtained based on the video attribute information and the second preference information through a preset user preference model; wherein the preset user preference model is trained based on the text features, visual features and audio features of the historical video content.

24. The apparatus according to claim 19, further comprising: A story sample extraction module is configured to extract character setting information, plot development mode information and rhythm characteristic information of different types of story samples; A story template creation module is configured to obtain different types of candidate story templates based on the character setting information, plot development mode information and rhythm feature information; The story template knowledge base construction module is configured to construct the story template knowledge base based on the candidate story templates.

25. The device according to claim 19, wherein The story template acquisition module includes: A prompt word acquisition submodule is configured to acquire a prompt word text, wherein the prompt word text is used to indicate specific content of the first preference information to be acquired; A content extraction submodule, configured to extract content information from the prompt word text; The story template acquisition submodule is configured to acquire the target story template matching the content information from the story template knowledge base.

26. The device according to claim 19, wherein The first preference information includes at least one of type feature information, emotion feature information and rhythm feature information. The story text generation module is further configured to: The first preference information is matched with elements in the target story template to generate the story text.

27. The apparatus according to any one of claims 19 to 26, further comprising: The first video generation module is configured to generate a target video based on the story text.

28. The apparatus according to claim 27, further comprising: A first feedback information acquisition module is configured to acquire feedback information of the target user for the target video; A first preference information updating module, configured to update the first preference information based on the feedback information to obtain updated third preference information; The first story and video updating module is configured to update the story text and the target video based on the third preference information.

29. The apparatus according to claim 27, further comprising: A first plot extraction module is configured to extract at least one plot node information in the story text; The first video splitting module is configured to decompose the target video based on the plot node information to obtain multiple video clips.

30. The apparatus of claim 27, further comprising: The first voice generation module is configured to generate voice commentary content of the target video based on the story text.

31. The apparatus of claim 19, further comprising: The story text updating module is configured to, in response to acquiring the story element selected by the target user, insert the story element into the story text to obtain an updated story text.

32. The apparatus according to claim 31, further comprising: The second video generation module is configured to generate a target video based on the updated story text.

33. The apparatus according to claim 32, further comprising: A second feedback information acquisition module is configured to acquire feedback information of the target user for the target video; a second preference information updating module, configured to update the first preference information based on the feedback information to obtain updated third preference information; The second story and video updating module is configured to update the updated story text and the target video based on the third preference information.

34. The apparatus of claim 32, further comprising: A second plot extraction module is configured to extract at least one plot node information in the updated story text; The second video splitting module is configured to decompose the target video based on the plot node information to obtain multiple video clips.

35. The apparatus of claim 32, further comprising: The second voice generation module is configured to generate voice commentary content of the target video based on the updated story text.

36. The apparatus according to claim 27 or 32, further comprising: The video analysis module is configured to obtain common features of reference videos, wherein the reference videos include: a preset number of videos ranked according to preset indicators; The video modification module is configured to modify the target video based on the common features.

37. An electronic device comprising: 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, and the instructions are executed by the at least one processor to enable the at least one processor to execute the story text generation method according to any one of claims 1 to 18.

38. A non-transitory computer-readable storage medium storing computer instructions, wherein the computer instructions are used to cause the computer to execute the story text generation method according to any one of claims 1 to 18.

39. A computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the steps of the story text generation method according to any one of claims 1 to 18 are implemented.

Images