Three-dimensional media data generation method and system based on script

Abstract

The invention is suitable for the field of games, and particularly relates to a script-based three-dimensional media data generation method and system, and the method comprises the steps: carrying out the image collection of a script, and extracting a game text based on the collected image; performing networking query to obtain a basic building model, and constructing a game scene model according to the basic building model; when the game starts, generating a basic role model, obtaining a user demand, and modifying the role model to obtain a plot character model; and loading the corresponding model according to the development progress of the plot, generating a stage plot model, and generating video data based on the view angle of the plot character model in the stage plot model. According to the method, the corresponding character model and the building model are automatically constructed, so that the corresponding video content is generated according to the real-time instruction of the user and is loaded by the game equipment of the user, the game freedom degree is greatly improved, the immersion sense is higher, the reading amount of the user can be reduced, and the user experience is improved.

Description

Technical Field

[0001] The present invention belongs to the field of games, and in particular relates to a method and system for generating three-dimensional media data based on a script. Background Art

[0002] Script-killing is an entertainment activity that combines role-playing, reasoning and puzzle-solving, and social interaction. Participants play characters with different background stories in a virtual world built by the storyline. They collect clues, communicate and discuss, and use logical reasoning to jointly solve a core problem or case, such as finding the murderer hidden in the group, solving a complex puzzle, or completing a specific task. This form of game not only tests the player's acting and reasoning ability, but also emphasizes teamwork and personal strategy, providing an immersive experience that allows every participant to enjoy the surprise and fun brought by the development of the story.

[0003] The current script-killing games are mainly played as tabletop games. Different players will get different identities and can see different information in different stages of the game. They can then analyze the information they see and give their own inferences and conclusions in the final stage. In this process, players need to read a lot of text information, which makes it difficult to achieve an immersive experience and quickly integrate into the game atmosphere of the script-killing games. Summary of the invention

[0004] The purpose of the present invention is to provide a method for generating three-dimensional media data based on a script, aiming to solve the problem in the prior art that players need to read a large amount of text information, find it difficult to achieve an immersive experience, and cannot quickly integrate into the game atmosphere of the script-killing game.

[0005] The present invention is implemented as follows: a method for generating three-dimensional media data based on a script, the method comprising:

[0006] Capturing images of the script-killing script, and extracting game text based on the captured images, wherein the game text includes background text, character introduction text, plot development text, and clue text;

[0007] Performing online query based on background text and plot development text to obtain a basic building model, and constructing a game scene model based on the basic building model, wherein the game scene model at least includes a plot building model, a character residence model, and an infrastructure model;

[0008] At the beginning of the game, the character introduction text is imported into the preset AI modeling model to generate a basic character model, obtain user needs, modify the character model, and obtain a plot character model;

[0009] The corresponding model is loaded according to the progress of the plot development to generate a stage plot model, and video data is generated based on the perspective of the plot character model in the stage plot model. The video data is loaded and displayed through the game device.

[0010] Preferably, the step of performing online query based on background text and plot development text to obtain a basic building model and constructing a game scene model according to the basic building model comprises:

[0011] Importing background text and plot development text into a preset large language model, extracting building information recorded therein, wherein the building information includes plot building information, residence building information and background environment information;

[0012] Conduct online query based on building information and obtain corresponding building models from the network, including plot building models, character residence models and infrastructure models;

[0013] The plot building model and character residence model are loaded into the infrastructure model to obtain the game scene model. The maximum activity range of each character is determined according to the plot development text, and a global movement restricted area is set.

[0014] Preferably, at the start of the game, the steps of importing the character introduction text into a preset AI modeling model, generating a basic character model, obtaining user needs, modifying the character model, and obtaining a plot character model specifically include:

[0015] At the beginning of the game, the character introduction text is imported into the preset AI modeling model to generate a basic character model, and the number of basic character models is multiple;

[0016] Displaying the basic role model on the corresponding game device, receiving the user's selection instruction, and retaining the corresponding basic role model;

[0017] Receive user requirements input by users in real time, analyze the user requirements, determine the modification area specified by the user requirements, modify the corresponding modification area through the AI ​​modeling model, and obtain the plot character model.

[0018] Preferably, the steps of loading the corresponding model according to the development progress of the plot, generating a stage plot model, and generating video data based on the perspective of the plot character model in the stage plot model specifically include:

[0019] Determine the activity range of each plot character in the current stage according to the plot development text, call up the game scene model, and set a local movement restricted area in the game scene model;

[0020] According to the positions of different plot characters, the plot character models are added to the game scene model to obtain the stage plot model, and the stage plot model is loaded;

[0021] Obtain user instructions to control the plot character model to move in the stage plot model, perform image rendering according to the real-time perspective of the plot character model, and generate video data.

[0022] Preferably, during the game, the game device collects images of the player, identifies the user's expression, and controls the plot character model to display the corresponding expression according to the user's expression.

[0023] Another object of the present invention is to provide a system for generating three-dimensional media data based on a script, the system comprising:

[0024] A plot extraction module is used to collect images of the script-killing script and extract the game text based on the collected images. The game text includes background text, character introduction text, plot development text and clue text;

[0025] A building model construction module, which is used to perform online query based on background text and plot development text to obtain a basic building model, and to construct a game scene model based on the basic building model, wherein the game scene model at least includes a plot building model, a character residence model, and an infrastructure model;

[0026] The character model building module is used to import the character introduction text into the preset AI modeling model at the beginning of the game, generate a basic character model, obtain user needs, modify the character model, and obtain a plot character model;

[0027] The model loading module is used to load the corresponding model according to the development progress of the plot, generate a stage plot model, and generate video data based on the perspective of the plot character model in the stage plot model. The video data is loaded and displayed through the game device.

[0028] Preferably, the building model construction module includes:

[0029] An information extraction unit, used for importing background text and plot development text into a preset large language model, and extracting building information recorded therein, wherein the building information includes plot building information, residence building information and background environment information;

[0030] The model acquisition unit is used to perform online query based on the building information and obtain the corresponding building models from the network, which are the plot building model, the character residence model and the infrastructure model;

[0031] The model loading unit is used to load the plot building model and the character residence model into the infrastructure model to obtain the game scene model, determine the maximum activity range of each character according to the plot development text, and set a global movement restricted area.

[0032] Preferably, the character model building module includes:

[0033] The character modeling unit is used to import the character introduction text into the preset AI modeling model at the beginning of the game to generate a basic character model, and the number of basic character models is multiple;

[0034] A basic model building unit, used to display the basic character model on the corresponding game device, receive a user's selection instruction, and retain the corresponding basic character model;

[0035] The model adjustment unit is used to receive user requirements input by the user in real time, analyze the user requirements, determine the modification area specified by the user requirements, modify the corresponding modification area through the AI ​​modeling model, and obtain a plot character model.

[0036] Preferably, the model loading module includes:

[0037] A restricted area setting unit is used to determine the activity range of each plot character in the current link according to the plot development text, retrieve the game scene model, and set a local movement restricted area in the game scene model;

[0038] A plot model loading unit, used to add plot character models to the game scene model according to the positions of different plot characters, obtain a stage plot model, and load the stage plot model;

[0039] The real-time rendering unit is used to obtain user instructions to control the plot character model to move in the stage plot model, perform image rendering according to the real-time perspective of the plot character model, and generate video data.

[0040] Preferably, during the game, the game device collects images of the player, identifies the user's expression, and controls the plot character model to display the corresponding expression according to the user's expression.

[0041] The present invention provides a script-based three-dimensional media data generation method, which automatically constructs corresponding character models and building models by extracting content from ordinary paper script-killing scripts, thereby generating corresponding video content according to the user's real-time instructions for loading on the user's gaming device, thereby greatly improving the degree of freedom of the game and providing a stronger sense of immersion, reducing the user's reading workload and improving the user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1A flowchart of a method for generating three-dimensional media data based on a script provided by an embodiment of the present invention;

[0043] Figure 2 A flowchart of the steps of performing online query based on background text and plot development text, obtaining a basic building model, and constructing a game scene model based on the basic building model provided by an embodiment of the present invention;

[0044] Figure 3 A flowchart of the steps of importing a character introduction text into a preset AI modeling model at the beginning of a game, generating a basic character model, obtaining user needs, modifying the character model, and obtaining a plot character model provided by an embodiment of the present invention;

[0045] Figure 4 A flowchart of the steps of loading a corresponding model according to the development progress of a plot, generating a stage plot model, and generating video data based on the perspective of a plot character model in the stage plot model provided by an embodiment of the present invention;

[0046] Figure 5 An architectural diagram of a script-based three-dimensional media data generation system provided in an embodiment of the present invention;

[0047] Figure 6 An architectural diagram of a building model construction module provided by an embodiment of the present invention;

[0048] Figure 7 An architectural diagram of a character model building module provided by an embodiment of the present invention;

[0049] Figure 8 An architectural diagram of a model loading module provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0050] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0051] like Figure 1 FIG. 1 is a flowchart of a method for generating three-dimensional media data based on a script according to an embodiment of the present invention, wherein the method comprises:

[0052] S100, collecting images of the script-killing game script, and extracting the game text based on the collected images, wherein the game text includes background text, character introduction text, plot development text and clue text.

[0053] In this step, images of the script-killing script are collected. Users can provide the script-killing script by themselves, and image collection is performed through the user's game device (such as a mobile phone or tablet computer), so as to collect multiple groups of script images. By performing text extraction on the script images, the text content in each group of script images is extracted, so as to determine the game text contained in each script image, and the text content is analyzed by a large language model to determine the type of text content corresponding to each script image, and it is divided into background text, character introduction text, plot development text and clue text. The background text is used to explain the background conditions of the current script, such as the location of the event, the content of the event, the time of the event and other information. The character introduction text is used to explain the relevant information of the plot characters, such as personality, experience, character relationships, etc. The plot development text is used to record the order of plot development, such as the events that occurred on the first day, the actions taken by each plot character, etc. The clue text is the clue that the plot characters can choose in each round of the game, which is used to promote the development of the plot.

[0054] S200, performing online query based on background text and plot development text, obtaining a basic building model, and constructing a game scene model according to the basic building model, wherein the game scene model at least includes a plot building model, a character residence model, and an infrastructure model.

[0055] In this step, an online query is performed based on the background text and the plot development text, and all building information is extracted from the background text and the plot development text, such as the building type where the event occurs. The building type can be a hotel, an office building, a residential building, etc., and an online query is performed based on the extracted building information to obtain 3D models of various types of buildings from the network, such as a hotel 3D model, an office building 3D model, a residential building 3D model, etc. Since the creation software of the above 3D models may be inconsistent, the format of the 3D model is unified by performing format conversion, so as to determine all 3D models involved in the script according to the background text and the plot development text, and According to the location where the event occurs in the script, query and obtain the infrastructure model. If the event occurs in the city, obtain the corresponding city model. Import all selected 3D models into the city model to obtain the game scene model. According to the use of the building in the plot, it is divided into plot building model, character residence model and infrastructure model. The plot building model is the place where the plot occurs, such as the crime scene. The character residence model is the residence of each character in the script. If the residence of the plot character is not explained, a residence model is randomly matched according to the relevant information of the plot character. The infrastructure model is a 3D model containing urban elements such as roads and vegetation.

[0056] S300, at the beginning of the game, imports the character introduction text into the preset AI modeling model, generates a basic role model, obtains user needs, modifies the role model, and obtains a plot character model.

[0057] In this step, at the beginning of the game, the character introduction text is imported into a preset AI modeling model, which may be Meshy AI and Tripo AI, etc. By importing the character introduction text into it, the corresponding basic role model is automatically generated using the above AI modeling model. In order to make the image of the plot character more in line with the user's preferences, after displaying the basic role model, the user inputs user needs according to his or her preferences, and the AI ​​modeling model modifies the basic role model according to the user's needs, thereby obtaining a plot character model that meets the user's preferences.

[0058] S400, loading the corresponding model according to the development progress of the plot, generating a stage plot model, generating video data based on the perspective of the plot character model in the stage plot model, and the video data is loaded and displayed through the game device.

[0059] In this step, the corresponding model is loaded according to the progress of the plot development, the entire game scene model is loaded, and each plot character model is simultaneously imported into the game scene model. Since different plot characters are allowed to move to different positions at different times, such as evidence A is in building M, and only characters a and b have obtained evidence A, then building M is a restricted area for movement of other characters. According to the relationship between characters and the distribution relationship between information in this stage, a restricted area for movement is set in the game scene model to obtain a stage plot model. In different game stages, the stage plot model is different. This is because as the plot progresses, the information obtained by the characters gradually increases, and the restricted area for movement gradually decreases. The stage plot model is loaded, and the user can input control instructions through the game device to control the movement of the plot character model in the stage plot model. According to the position of the plot character model, video data is generated by rendering and displayed on the game device.

[0060] like Figure 2 As shown, as a preferred embodiment of the present invention, the steps of performing online query based on background text and plot development text to obtain a basic building model and constructing a game scene model according to the basic building model include:

[0061] S201, importing background text and plot development text into a preset large language model, and extracting building information recorded therein, wherein the building information includes plot building information, residential building information and background environment information.

[0062] In this step, the background text and plot development text are imported into the preset large language model, and processed by the large language model to extract the building information. In order to improve the accuracy of modeling, the description text of the building is refined by the large language model to obtain more specific building information.

[0063] S202, perform online query based on the building information, and obtain corresponding building models from the network, which are respectively a plot building model, a character residence model, and an infrastructure model.

[0064] In this step, an online query is performed based on the building information, a query is performed on the network based on the building information, or a local database is queried to obtain the corresponding building model, and the plot building model, character residence model and infrastructure model are obtained respectively according to the plot building information, residence building information and background environment information.

[0065] S203, loading the plot building model and the character residence model into the infrastructure model to obtain the game scene model, determining the maximum activity range of each character according to the plot development text, and setting a global movement restricted area.

[0066] In this step, the plot building model and the character residence model are loaded into the infrastructure model. In order to unify the models, the format is first unified, and then the display ratio of the model is adjusted to ensure the consistency of the ratio of each model. The game scene model is obtained by superimposing the models. The content of the entire script is analyzed through the large language model to determine the areas that each plot character is prohibited from entering during the entire game. For example, if all plot characters have not been to building C, building C is set as a global mobile restricted area, and the global mobile restricted area is further filled, retaining only its appearance, and the internal data is not loaded to reduce the amount of data loaded.

[0067] like Figure 3 As shown, as a preferred embodiment of the present invention, at the beginning of the game, the character introduction text is imported into the preset AI modeling model, a basic role model is generated, user needs are obtained, the role model is modified, and the steps of obtaining the plot character model specifically include:

[0068] S301, when the game starts, import the character introduction text into the preset AI modeling model to generate a basic character model, and the number of basic character models is multiple.

[0069] In this step, at the beginning of the game, the character introduction text is imported into the preset AI modeling model. Before that, the character introduction text is also imported into the big prediction model, and the big language model is used to refine the character introduction text, thereby helping the AI ​​modeling model to generate a more accurate basic role model. In order to meet the needs of users, multiple basic role models can be generated at the same time for users to choose from.

[0070] S302, displaying the basic character model on the corresponding game device, receiving a user's selection instruction, and retaining the corresponding basic character model.

[0071] In this step, the basic character model is displayed on the corresponding game device. By displaying the basic character model, the user preliminarily selects it according to his or her own preferences, thereby obtaining a basic character model.

[0072] S303, receiving user requirements input by the user in real time, analyzing the user requirements, determining the modification area specified by the user requirements, modifying the corresponding modification area through the AI ​​modeling model, and obtaining a plot character model.

[0073] In this step, user needs are received in real time. Users input their needs through text or voice based on their understanding of the plot characters. The AI ​​modeling model modifies the basic character model based on the user needs. In this process, the user needs are analyzed to determine the areas that need to be modified. For example, if the nose is to be raised a little, the nose is selected as the modification area, and other areas are not modified to obtain the plot character model.

[0074] like Figure 4 As shown, as a preferred embodiment of the present invention, the steps of loading the corresponding model according to the development progress of the plot, generating a stage plot model, and generating video data based on the perspective of the plot character model in the stage plot model specifically include:

[0075] S401, determining the activity range of each plot character in the current link according to the plot development text, calling the game scene model, and setting a local movement restricted area in the game scene model.

[0076] In this step, the activity range of each character in the current link is determined according to the plot development text. By setting the activity range for each character, the union of the activity ranges of all characters is the movable area, and the other parts are the local restricted area for movement in the current link.

[0077] S402, adding the plot character models to the game scene model according to the positions of different plot characters, obtaining the stage plot model, and loading the stage plot model.

[0078] In this step, the plot character models are loaded according to the locations of different plot characters, so that the plot characters corresponding to each user are added to the game scene model. In different time periods, the activity areas of the plot characters are different, and the stage plot model in the current time period is loaded.

[0079] S403, obtaining a user instruction to control the plot character model to move in the stage plot model, performing image rendering according to the real-time viewing angle of the plot character model, and generating video data.

[0080] In this step, user instructions are obtained to control the plot character model to move in the stage plot model. The user generates control instructions by inputting movement instructions, and the real-time position of the plot character model in the stage plot model is continuously updated according to the control instructions, so as to determine the field of view according to the position of the plot character model, and obtain video data through real-time rendering for display on the user's game device; during the game, the player's image is collected through the game device, the user's expression is identified, and the plot character model is controlled to display the corresponding expression according to the user's expression. Therefore, when the characters in the plot are talking, the player's expression can be indirectly seen, so that the player can determine the authenticity of the speech, thereby enhancing the fun and freedom of the game.

[0081] like Figure 5 As shown, a system for generating three-dimensional media data based on a script is provided in an embodiment of the present invention, and the system includes:

[0082] The plot extraction module 100 is used to collect images of the script-killing script and extract the game text based on the collected images. The game text includes background text, character introduction text, plot development text and clue text.

[0083] In this system, the plot extraction module 100 performs image acquisition on the script-killing script. The user can provide the script-killing script by himself, and the image acquisition is performed through the user's game device (such as a mobile phone or a tablet computer), so as to acquire multiple groups of script images. By performing text extraction on the script images, the text content in each group of script images is extracted, so as to determine the game text contained in each script image. The text content is analyzed by a large language model to determine the type of text content corresponding to each script image, and it is divided into background text, character introduction text, plot development text and clue text. The background text is used to explain the background conditions of the current script, such as the location of the event, the content of the event, the time of the event and other information. The character introduction text is used to explain the relevant information of the plot characters, such as personality, experience, character relationships, etc. The plot development text is used to record the development sequence of the plot, such as the events that occurred on the first day, the actions taken by each plot character, etc. The clue text is the clue that the plot characters can choose in each round of the game to promote the development of the plot.

[0084] The building model construction module 200 is used to perform online query based on background text and plot development text to obtain a basic building model, and to construct a game scene model based on the basic building model. The game scene model at least includes a plot building model, a character residence model, and an infrastructure model.

[0085] In this system, the building model construction module 200 performs online query based on the background text and the plot development text, extracts all the building information from the background text and the plot development text, such as the building type where the event occurs, which can be a hotel, an office building, a residential building, etc., performs online query based on the extracted building information, and obtains three-dimensional models of various types of buildings from the network, such as a hotel three-dimensional model, an office building three-dimensional model, a residential building three-dimensional model, etc. Since the creation software of the above three-dimensional models may be inconsistent, the format of the three-dimensional model is unified by performing format conversion, so as to determine all the three-dimensional models involved in the script according to the background text and the plot development text. 3D model, and query and obtain the infrastructure model according to the location where the event occurs in the script. For example, if the event occurs in the city, the corresponding city model is obtained. All selected 3D models are imported into the city model to obtain the game scene model. According to the use of the building in the plot, it is divided into plot building model, character residence model and infrastructure model. The plot building model is the place where the plot occurs, such as the crime scene. The character residence model is the residence of each character in the script. If the residence of the plot character is not explained, a residence model is randomly matched according to the relevant information of the plot character. The infrastructure model is a 3D model containing urban elements such as roads and vegetation.

[0086] The character model construction module 300 is used to import the character introduction text into the preset AI modeling model at the beginning of the game, generate a basic role model, obtain user needs, modify the role model, and obtain a plot character model.

[0087] In this system, at the beginning of the game, the character model construction module 300 imports the character introduction text into a preset AI modeling model, which may be Meshy AI and Tripo AI, etc. By importing the character introduction text into it, the corresponding basic role model is automatically generated using the above AI modeling model. In order to make the image of the plot character more in line with the user's preferences, after displaying the basic role model, the user inputs user requirements according to his or her preferences, and the AI ​​modeling model modifies the basic role model according to the user's requirements, thereby obtaining a plot character model that meets the user's preferences.

[0088] The model loading module 400 is used to load the corresponding model according to the development progress of the plot, generate a stage plot model, and generate video data based on the perspective of the plot character model in the stage plot model. The video data is loaded and displayed through the game device.

[0089] In this system, the model loading module 400 loads the corresponding model according to the progress of the plot development, loads the entire game scene model, and simultaneously imports each plot character model into the game scene model. Since different plot characters are allowed to move to different positions at different times, such as evidence A is in building M, and only characters a and b have obtained evidence A, then building M is a restricted area for movement of other characters. According to the relationship between characters and the distribution relationship between information in this stage, a restricted area for movement is set in the game scene model to obtain a stage plot model. In different game stages, the stage plot model is different. This is because as the plot progresses, the information obtained by the characters gradually increases, and the restricted area for movement gradually decreases. The stage plot model is loaded, and the user can input control instructions through the game device to control the movement of the plot character model in the stage plot model. According to the position of the plot character model, video data is generated by rendering and displayed on the game device.

[0090] like Figure 6 As shown, as a preferred embodiment of the present invention, the building model construction module 200 includes:

[0091] The information extraction unit 201 is used to import the background text and the plot development text into a preset large language model, and extract the building information recorded therein, wherein the building information includes the plot building information, the residence building information and the background environment information.

[0092] In this module, the information extraction unit 201 imports the background text and the plot development text into the preset large language model, processes them through the large language model, and thus extracts the building information. In order to improve the accuracy of modeling, the description text of the building is refined through the large language model to obtain more specific building information.

[0093] The model acquisition unit 202 is used to perform online query according to the building information and acquire the corresponding building models from the network, which are the plot building model, the character residence model and the infrastructure model.

[0094] In this module, the model acquisition unit 202 performs online query based on the building information, queries on the network based on the building information, or queries the local database to obtain the corresponding building model, and obtains the plot building model, character residence model and infrastructure model according to the plot building information, residence building information and background environment information.

[0095] The model loading unit 203 is used to load the plot building model and the character residence model into the infrastructure model to obtain the game scene model, determine the maximum activity range of each character according to the plot development text, and set a global movement restricted area.

[0096] In this module, the model loading unit 203 loads the plot building model and the character residence model in the infrastructure model. In order to unify the models, the format is first unified, and then the display ratio of the model is adjusted to ensure that the ratio of each model is consistent. By superimposing the models, the game scene model is obtained, and the content of the entire script is analyzed through the large language model to determine the areas that each plot character is prohibited from entering during the entire game. If all plot characters have not been to building C, building C is set as a global mobile restricted area, and the global mobile restricted area is further filled, retaining only its appearance, and the internal data is not loaded to reduce the amount of data loaded.

[0097] like Figure 7 As shown, as a preferred embodiment of the present invention, the character model construction module 300 includes:

[0098] The role modeling unit 301 is used to import the character introduction text into the preset AI modeling model at the beginning of the game to generate a basic role model, and the number of basic role models is multiple.

[0099] In this module, the role modeling unit 301 imports the character introduction text into the preset AI modeling model at the beginning of the game. Before that, the character introduction text is also imported into the big prediction model, and the big language model is used to refine the character introduction text, thereby helping the AI ​​modeling model to generate a more accurate basic role model. In order to meet the needs of users, multiple basic role models can be generated at the same time for users to choose from.

[0100] The basic model building unit 302 is used to display the basic role model on the corresponding game device, receive the user's selection instruction, and retain the corresponding basic role model.

[0101] In this module, the basic model construction unit 302 displays the basic character model on the corresponding game device. By displaying the basic character model, the user preliminarily selects it according to his or her own preferences, thereby obtaining a basic character model.

[0102] The model adjustment unit 303 is used to receive user requirements input by the user in real time, analyze the user requirements, determine the modification area specified by the user requirements, modify the corresponding modification area through the AI ​​modeling model, and obtain a plot character model.

[0103] In this module, the model adjustment unit 303 receives user requirements input by the user in real time. The user inputs the user requirements through text or voice according to his or her understanding of the plot character. The AI ​​modeling model modifies the basic character model according to the user requirements. In this process, the user requirements are analyzed to determine the areas that need to be modified. For example, if the nose is to be raised a little, the nose is selected as the modification area, and other areas are not modified to obtain the plot character model.

[0104] like Figure 8 As shown, as a preferred embodiment of the present invention, the model loading module 400 includes:

[0105] The restricted area setting unit 401 is used to determine the activity range of each plot character in the current link according to the plot development text, call the game scene model, and set a local movement restricted area in the game scene model.

[0106] In this module, the restricted area setting unit 401 determines the activity range of each plot character in the current link according to the plot development text. By setting the activity range for each plot character, the union of the activity ranges of all plot characters is the movable area, and the other parts are the local movement restricted area of ​​the current link.

[0107] The plot model loading unit 402 is used to add plot character models to the game scene model according to the positions of different plot characters, obtain a stage plot model, and load the stage plot model.

[0108] In this module, the plot model loading unit 402 loads the plot character models according to the locations of different plot characters, thereby adding the plot characters corresponding to each user to the game scene model. In different time periods, the activity areas of the plot characters are different, and the stage plot model in the current time period is loaded.

[0109] The real-time rendering unit 403 is used to obtain user instructions to control the plot character model to move in the stage plot model, perform image rendering according to the real-time perspective of the plot character model, and generate video data.

[0110] In this module, the real-time rendering unit 403 obtains user instructions to control the plot character model to move in the stage plot model. The user generates a control instruction by inputting a movement instruction, and continuously updates the real-time position of the plot character model in the stage plot model according to the control instruction, thereby determining the field of view according to the position of the plot character model, and obtaining video data through real-time rendering for display on the user's game device; during the game, the game device collects images of the player, identifies the user's expression, and controls the plot character model to display the corresponding expression according to the user's expression. Therefore, when the characters in the plot are talking, the player's expression can be indirectly seen, so that the player can determine the authenticity of the speech, thereby enhancing the fun and freedom of the game.

[0111] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for generating three-dimensional media data based on a script, characterized in that: The method comprises: Capturing images of the script-killing script, and extracting game text based on the captured images, wherein the game text includes background text, character introduction text, plot development text, and clue text; Performing online query based on background text and plot development text to obtain a basic building model, and constructing a game scene model based on the basic building model, wherein the game scene model at least includes a plot building model, a character residence model, and an infrastructure model; At the beginning of the game, the character introduction text is imported into the preset AI modeling model to generate a basic character model, obtain user needs, modify the character model, and obtain a plot character model; The corresponding model is loaded according to the progress of the plot development to generate a stage plot model, and video data is generated based on the perspective of the plot character model in the stage plot model. The video data is loaded and displayed through the game device.

2. The method for generating three-dimensional media data based on a script according to claim 1, characterized in that: The steps of performing online query based on background text and plot development text to obtain a basic building model and constructing a game scene model according to the basic building model include: Importing background text and plot development text into a preset large language model, extracting building information recorded therein, wherein the building information includes plot building information, residence building information and background environment information; Conduct online query based on building information and obtain corresponding building models from the network, including plot building models, character residence models and infrastructure models; The plot building model and character residence model are loaded into the infrastructure model to obtain the game scene model. The maximum activity range of each character is determined according to the plot development text, and a global movement restricted area is set.

3. The method for generating three-dimensional media data based on a script according to claim 1, characterized in that: The steps of importing the character introduction text into the preset AI modeling model at the beginning of the game, generating a basic character model, obtaining user needs, modifying the character model, and obtaining a plot character model specifically include: At the beginning of the game, the character introduction text is imported into the preset AI modeling model to generate a basic character model, and the number of basic character models is multiple; Displaying the basic role model on the corresponding game device, receiving the user's selection instruction, and retaining the corresponding basic role model; Receive user requirements input by users in real time, analyze the user requirements, determine the modification area specified by the user requirements, modify the corresponding modification area through the AI ​​modeling model, and obtain the plot character model.

4. The method for generating three-dimensional media data based on a script according to claim 1, characterized in that: The steps of loading the corresponding model according to the development progress of the plot, generating a stage plot model, and generating video data based on the perspective of the plot character model in the stage plot model specifically include: Determine the activity range of each plot character in the current stage according to the plot development text, call up the game scene model, and set a local movement restricted area in the game scene model; According to the positions of different plot characters, the plot character models are added to the game scene model to obtain the stage plot model, and the stage plot model is loaded; Obtain user instructions to control the plot character model to move in the stage plot model, perform image rendering according to the real-time perspective of the plot character model, and generate video data.

5. The method for generating three-dimensional media data based on a script according to claim 1, characterized in that: During the game, the game device collects images of the player, recognizes the user's expression, and controls the plot character model to display the corresponding expression based on the user's expression.

6. A script-based three-dimensional media data generation system, characterized in that: The system comprises: A plot extraction module is used to collect images of the script-killing script and extract the game text based on the collected images. The game text includes background text, character introduction text, plot development text and clue text; A building model construction module, which is used to perform online query based on background text and plot development text to obtain a basic building model, and to construct a game scene model based on the basic building model, wherein the game scene model at least includes a plot building model, a character residence model, and an infrastructure model; The character model building module is used to import the character introduction text into the preset AI modeling model at the beginning of the game, generate a basic character model, obtain user needs, modify the character model, and obtain a plot character model; The model loading module is used to load the corresponding model according to the development progress of the plot, generate a stage plot model, and generate video data based on the perspective of the plot character model in the stage plot model. The video data is loaded and displayed through the game device.

7. The script-based three-dimensional media data generation system according to claim 6, characterized in that: The building model construction module includes: An information extraction unit, used for importing background text and plot development text into a preset large language model, and extracting building information recorded therein, wherein the building information includes plot building information, residence building information and background environment information; The model acquisition unit is used to perform online query based on the building information and obtain the corresponding building models from the network, which are the plot building model, the character residence model and the infrastructure model; The model loading unit is used to load the plot building model and the character residence model into the infrastructure model to obtain the game scene model, determine the maximum activity range of each character according to the plot development text, and set a global movement restricted area.

8. The script-based three-dimensional media data generation system according to claim 6, characterized in that: The character model building module includes: The character modeling unit is used to import the character introduction text into the preset AI modeling model at the beginning of the game to generate a basic character model, and the number of basic character models is multiple; A basic model building unit, used to display the basic character model on the corresponding game device, receive a user's selection instruction, and retain the corresponding basic character model; The model adjustment unit is used to receive user requirements input by the user in real time, analyze the user requirements, determine the modification area specified by the user requirements, modify the corresponding modification area through the AI ​​modeling model, and obtain a plot character model.

9. The script-based three-dimensional media data generation system according to claim 6, characterized in that: The model loading module includes: A restricted area setting unit is used to determine the activity range of each plot character in the current link according to the plot development text, retrieve the game scene model, and set a local movement restricted area in the game scene model; A plot model loading unit, used to add plot character models to the game scene model according to the positions of different plot characters, obtain a stage plot model, and load the stage plot model; The real-time rendering unit is used to obtain user instructions to control the plot character model to move in the stage plot model, perform image rendering according to the real-time perspective of the plot character model, and generate video data.

10. The script-based three-dimensional media data generation system according to claim 6, characterized in that: During the game, the game device collects images of the player, recognizes the user's expression, and controls the plot character model to display the corresponding expression based on the user's expression.

Images