Content creation method and device, electronic equipment and computer storage medium

Abstract

The embodiment of the invention provides a content creation method and device, electronic equipment and a computer storage medium. The content creation method comprises the following steps: displaying a creation interface through digital content creation software; receiving a creation triggering operation in the creation interface, calling the content creation model to execute a content creation task matched with the received creation triggering operation, and obtaining a content creation result; the content creation model is an artificial intelligence model used for executing a content creation task; and updating the creation interface based on the content creation result. According to the embodiment of the invention, the efficiency and quality of content creation can be improved.

Description

Technical Field

[0001] This application relates to the field of computer technology, and in particular to a content creation method, apparatus, electronic device, and computer storage medium. Background Technology

[0002] DCC (Digital Content Creation) software is a software tool that can be used to create and edit digital media content (such as 3D character models), and is widely used in animation production, game development, virtual reality and other fields.

[0003] Currently, the content creation functions provided by DCC software often fail to meet users' high creative needs. Therefore, users must use other software for creation and then manually import the results into DCC for further processing. This process involves users navigating to external platforms and importing / exporting data, making it complex and resulting in low efficiency and quality. Summary of the Invention

[0004] In view of this, embodiments of this application provide a content creation solution to at least partially solve the above-mentioned problems.

[0005] According to a first aspect of the embodiments of this application, a content creation method is provided, applied to an electronic device deployed with digital content creation software, the method comprising: The creation interface is displayed through the digital content creation software. The system receives a creation trigger operation in the creation interface, calls a content creation model to execute a content creation task that matches the received creation trigger operation, and obtains a content creation result; the content creation model is an artificial intelligence model used to execute the content creation task. Update the creation interface based on the content creation results.

[0006] According to a second aspect of the embodiments of this application, a content creation apparatus is provided, located in an electronic device on which digital content creation software is deployed, the apparatus comprising: The creation interface display module is used to display the creation interface through the digital content creation software. The intelligent creation module is used to receive creation trigger operations in the creation interface, call the content creation model to execute the content creation task that matches the received creation trigger operation, and obtain the content creation result; the content creation model is an artificial intelligence model used to execute the content creation task. The creation interface update module is used to update the creation interface based on the content creation results.

[0007] According to a third aspect of the present application, an electronic device is provided, comprising: a processor, a memory, a communication interface, and a communication bus, wherein the processor, the memory, and the communication interface communicate with each other via the communication bus; the memory is used to store at least one executable instruction, wherein the executable instruction causes the processor to perform an operation corresponding to the method described in the first aspect.

[0008] According to a fourth aspect of the embodiments of this application, a computer storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the method described in the first aspect.

[0009] According to a fifth aspect of the embodiments of this application, a computer program product is provided, which includes computer instructions that instruct a computing device to perform an operation corresponding to the method described in the first aspect.

[0010] According to the content creation scheme provided in the embodiments of this application, a creation interface is displayed through digital content creation software; a creation trigger operation is received in the creation interface, and a content creation model is invoked to execute a content creation task that matches the received creation trigger operation, thereby obtaining a content creation result; the content creation model is an artificial intelligence model used to execute content creation tasks; and the creation interface is updated based on the content creation result.

[0011] In this embodiment, content creation capabilities based on an artificial intelligence model are embedded in the digital content creation software. This allows the software to fully leverage the potential of the AI ​​model to simplify and accelerate the content creation process. When users create content using the software, they do not need to perform complex platform switching or data import / export operations. They can seamlessly switch to the AI-driven intelligent creation mode through a simple creation trigger, thereby quickly and efficiently obtaining high-quality creative results that meet their needs. Therefore, this embodiment improves the efficiency and quality of content creation. Attached Figure Description

[0012] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings.

[0013] Figure 1 This is a flowchart illustrating the steps of a content creation method according to an embodiment of this application; Figure 2 This is a schematic diagram of a content creation scenario according to an embodiment of this application; Figure 3This is a schematic diagram illustrating another content creation scenario according to an embodiment of this application; Figure 4 This is a schematic diagram illustrating another content creation scenario according to an embodiment of this application; Figures 5-7 This is a schematic diagram illustrating another content creation scenario according to an embodiment of this application; Figures 8-9 This is a schematic diagram illustrating another content creation scenario according to an embodiment of this application; Figure 10 This is a structural block diagram of a content creation device according to an embodiment of this application; Figure 11 This is a schematic diagram of the structure of an electronic device according to an embodiment of this application. Detailed Implementation

[0014] To enable those skilled in the art to better understand the technical solutions in the embodiments of this application, the technical solutions in the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art should fall within the protection scope of the embodiments of this application.

[0015] Reference Figure 1 , Figure 1 This is a flowchart illustrating the steps of a content creation method according to an embodiment of this application. The content creation method provided in this embodiment can be executed by any suitable electronic device deployed with digital content creation software. For example, the aforementioned electronic device can be a server, PC, etc., deployed with digital content creation software. The content creation method may include the following steps: Step 102: Display the creation interface using digital content creation software.

[0016] To illustrate, digital content creation software can refer to software tools used to create, edit, and manage digital content. Digital content refers to content stored in the form of digital signals, such as two-dimensional digital models, three-dimensional digital models, animations, etc.

[0017] When a user wants to create content using digital content creation software, they can perform a preset software startup operation (such as clicking the software icon). Afterwards, the electronic device executing the content creation method provided in this application embodiment can provide a creation interface to the user through the software, enabling better human-computer interaction during the subsequent content creation process.

[0018] Step 104: Receive the creation trigger operation in the creation interface, call the content creation model to execute the content creation task that matches the received creation trigger operation, and obtain the content creation result; the content creation model is an artificial intelligence model used to execute the content creation task.

[0019] Indicatively, a creation trigger operation can be predefined to trigger the content creation task. When the above creation trigger operation is received, the intelligent creation mode driven by the artificial intelligence model is entered: the content creation model is called to execute the content creation task that matches the received creation trigger operation, thereby obtaining the content creation result.

[0020] The content creation model can be an artificial intelligence model embedded or integrated into digital content creation software. This artificial intelligence model can be used to perform content creation tasks and output content creation results. In this embodiment, there are no limitations on the specific model type of the content creation model, or the specific values ​​of the model parameters. Any suitable artificial intelligence model can be used, such as a pre-trained neural network model.

[0021] In this embodiment, the specific content of the content creation task can be customized according to actual conditions. For example, it may include at least one of the following: digital character model creation task, digital character model editing task, digital character model stylization task, and digital character model texture generation task. The digital character model creation task is the task of creating digital character models (such as 2D animated characters, 3D animated characters, etc.); the digital character model editing task is the task of editing the overall or partial geometric features of an already created digital character model; the digital character model stylization task is the task of processing an already created digital character model to give it a specific visual style and artistic effect; and the digital character model texture generation task can refer to the task of creating textures on the surface of an already created digital character model.

[0022] Furthermore, there are no restrictions on the specific content of the creation trigger operation. For example, it can be a trigger operation on a physical component of an electronic device, such as clicking a keyboard key, clicking the left or right mouse button, etc. Alternatively, content creation icons corresponding to the content creation task can be displayed in advance in the creation interface, and the creation trigger operation is the trigger operation on the aforementioned content creation icon, such as clicking the content creation icon, double-clicking the content creation icon, dragging the content creation icon, etc.

[0023] Step 106: Update the creation interface based on the content creation results.

[0024] Schematic illustration: After obtaining the content creation result through step 104, step 106 can be used to update the interface and display the content creation result to the user, allowing the user to determine whether the current content creation result meets the requirements. If the requirements are met, the content creation result is saved and the creation process ends; otherwise, if the current content creation result does not meet the user's requirements, the user can return to step 104 to create content again until the obtained content creation result meets the user's needs.

[0025] In this embodiment, content creation capabilities based on an artificial intelligence model are embedded in the digital content creation software. This allows the software to fully leverage the potential of the AI ​​model to simplify and accelerate the content creation process. When users create content using the software, they do not need to perform complex platform switching or data import / export operations. They can seamlessly switch to the AI-driven intelligent creation mode through a simple creation trigger, thereby quickly and efficiently obtaining high-quality creative results that meet their needs. Therefore, this embodiment improves the efficiency and quality of content creation.

[0026] Optionally, in some embodiments, multiple types of creation trigger operations can be pre-defined, with each creation trigger operation corresponding to a content creation task. Correspondingly, multiple types of content creation models can also be pre-configured, with each content creation model used to perform a content creation task.

[0027] The process described above, invoking the content creation model to execute a content creation task matching the received creation trigger operation and obtaining the content creation result, can include: Among various content creation models, identify the target content creation model that matches the received creation trigger operation; The target content creation model is invoked to execute the content creation task that matches the received creation trigger operation, and the content creation result is obtained.

[0028] Indicatively, digital content creation software typically allows for the completion of a series of creation tasks. The embodiments described above, considering the various creation tasks that may be involved in the content creation process, provide a set of AI-driven intelligent creation modes for each task, thereby achieving modular segmentation and scheduling of different creation tasks. In this way, users can execute corresponding creation trigger operations according to their needs to trigger the corresponding intelligent creation mode, thereby scheduling the corresponding content creation tasks and obtaining the corresponding content creation results.

[0029] Indicatively, a fully integrated content creation workflow can be constructed based on the various content creation tasks that digital content creation software may involve, thereby enabling the solution provided in this application embodiment to cover the entire content creation chain. For example, various content creation tasks may include: digital character model creation tasks, digital character model editing tasks, digital character model stylization tasks, and digital character model texture generation tasks. Users can uniformly manage and schedule the above-mentioned model creation, model editing, model stylization, and model texture generation tasks by executing the creation trigger operations corresponding to each task, thereby forming a closed-loop creation process.

[0030] The above embodiments of this application provide a variety of AI-driven intelligent creation modes. Users can choose the appropriate mode to create relevant content according to their actual needs. Therefore, while improving the quality and effect of content creation, it can further enhance the diversity and flexibility of content creation.

[0031] Optionally, in some embodiments, multiple types of content creation models can be pre-configured, with one content creation model used to perform a content creation task.

[0032] Correspondingly, the process of calling the content creation model to execute a content creation task matching the received creation trigger operation and obtaining a content creation result may also include: According to the pre-set task execution order, determine the target content creation task to be executed; call the target content creation model corresponding to the target content creation task to execute the target content creation task; in response to the completion of the target content creation task, return to the step of determining the target content creation task to be executed according to the pre-set task execution order, until the content creation result corresponding to the last content creation task is obtained; wherein, the last content creation task is the latest content creation task to be executed according to the task execution order.

[0033] To illustrate, the execution order of multiple different content creation tasks can be pre-set. When a user triggers a creation operation in the creation interface, an automatically flowing workflow can be built according to the above execution order. That is, after the current content creation task is completed, the process will automatically switch to the next content creation task according to the above execution order until all content creation tasks are completed.

[0034] In the above embodiments, the task execution order can be a pre-set fixed order or a custom order set by the user according to their own creative habits. Alternatively, users can choose from a variety of content creation tasks instead of following the workflow's execution order, and obtain the corresponding content creation result.

[0035] In the above embodiments of this application, by constructing an integrated content creation workflow, users can realize a complete closed-loop creation process from content creation to content editing and then to visual enhancement processing based on digital content creation software, thus further improving the user's creation experience.

[0036] Optionally, in some embodiments, the content creation method may further include: The input guidance information is displayed in the creation interface; the input guidance information is used to guide the input operations required for content creation.

[0037] Correspondingly, the process of receiving a creation trigger operation in the creation interface, invoking the content creation model to execute a content creation task matching the received creation trigger operation, and obtaining the content creation result can include: It receives input content creation requirements; in response to receiving a creation trigger operation in the creation interface, it calls the content creation model to execute a content creation task that matches the received creation trigger operation based on the content creation requirements, and obtains the content creation result.

[0038] To illustrate, different users typically have different creative requirements when creating content; similarly, even the same user typically has different creative requirements in different content creation scenarios.

[0039] Considering the above, in this embodiment of the application, before receiving the user's trigger operation in the creation interface, input guidance information can be displayed to the user through the creation interface to guide the user in inputting content creation requirements. Here, there is no limitation on the way the input guidance information is displayed. For example, it can be displayed in the form of text, or in the form of text plus voice broadcast, or in the form of images or animations.

[0040] Content creation requirements can be in text format, audio format, or other formats such as video.

[0041] After obtaining the content creation requirements input by the user, when the user executes the creation trigger operation, these requirements can be input into the content creation model. The model will then create content according to these requirements, resulting in a content creation outcome that better meets the user's customized needs. Therefore, the above embodiments can further enhance the user's creation experience.

[0042] Alternatively, in some embodiments, the above content creation requirements may be model creation requirements for creating digital character models.

[0043] Correspondingly, the process of responding to a creation trigger operation received in the creation interface, invoking the content creation model based on content creation requirement information, executing a content creation task matching the received creation trigger operation, and obtaining a content creation result may include: In response to receiving a model creation trigger operation in the creation interface, the first text-based image model is invoked to generate a digital character model that matches the model creation requirements.

[0044] Correspondingly, the process of updating the creation interface based on the content creation results can include: Display digital character models in the creation interface.

[0045] The above process is illustrated as the execution of a digital persona model creation task. Input guidance information can be displayed in the content creation interface to guide the user in inputting the model creation requirements.

[0046] See Figure 2 , Figure 2 A creation interface provided for digital content creation software. Exemplarily, this creation interface may include two parts: a secondary interface and a main interface. The secondary interface displays input guidance information such as "Upload an image or enter text to create a digital character model," and provides input boxes. Furthermore, input guidance information can be displayed again in the input boxes. For ease of user understanding, this input guidance information may differ from the aforementioned information, such as... Figure 2 The prompt in the text is: "Please enter the prompt word for the 3D task generation, for example: a science fiction-themed cyberpunk-style woman."

[0047] After the user enters their model creation requirements in the input box, a model creation trigger action can be executed, for example: clicking... Figure 2The arrow icon is shown in the auxiliary interface. Next, the digital content creation software will call the text-based graph model (for ease of distinction, the content creation model used to generate the data character model is called the first text-based graph model; this first text-based graph model can be a neural network model used to generate a 3D character model, etc.) to generate a data character model that meets the user's input model creation requirements, and then display this data character model on the main interface.

[0048] Schematic, the process of calling the first text-based image model to generate a digital character model that matches the model creation requirements can include: using a natural language understanding model to perform semantic parsing and keyword completion on the user-input model creation requirements (i.e., automatically filling in or predicting key information that conforms to semantic logic based on the incomplete information input by the user to improve the expression of the user's intent); then using an information extraction model to extract information from the completed information to obtain modeling auxiliary information such as character posture information and clothing style; and then inputting the above modeling auxiliary information into the first text-based image model to generate a digital character model.

[0049] For example, the natural language understanding model described above can be a recurrent neural network model, a convolutional neural network model, a Transformer model, etc. The training process of the natural language understanding model may include: obtaining the text to be parsed described in natural language as a training sample; inputting the training sample into the natural language understanding model; performing semantic parsing and keyword completion through the natural language understanding model to output predicted text; calculating the loss value based on the predicted text and the training labels (the complete text corresponding to the text to be parsed); adjusting the model parameters according to the loss value to obtain the trained natural language understanding model.

[0050] The information extraction model described above can also be a recurrent neural network model, a convolutional neural network model, a Transformer model, etc. The training process of the information extraction model may include: obtaining the text to be extracted described in natural language as training samples, inputting the training samples into the information extraction model, and outputting predicted information through the information extraction model; calculating the loss value based on the predicted information and the training labels (the modeling auxiliary information actually contained in the text to be extracted), adjusting the model parameters according to the loss value, and obtaining the trained information extraction model.

[0051] The first text-based image model can include an image rendering sub-model and a diffusion sub-model. The image rendering sub-model can be a differentiable renderer, such as NeRF or DMTet; the diffusion sub-model can employ a Stable Diffusion model, a Point-E model, etc. During the inference phase, the process of generating a digital character model using the first text-based image model can be as follows: Obtain a preset initial neural radiation field; input the initial neural radiation field and modeling auxiliary information into the image rendering sub-model, and render an initial 2D image; the diffusion sub-model performs noise image prediction based on the initial 2D image, calculates gradients based on the predicted noise images, and sends the calculation results back to the image rendering sub-model to update the initial neural radiation field; the image rendering sub-model renders an updated 2D image based on the updated neural radiation field; and 3D reconstruction is performed based on the updated 2D image to obtain the digital character model.

[0052] During the training phase, it is usually sufficient to adjust the parameters of the diffusion sub-model in the first raw image model, without needing to adjust the parameters of the image rendering sub-model. The training process of the diffusion sub-model may include: inputting the two-dimensional sample image and the image description text into the diffusion sub-model to obtain the predicted noisy image; calculating the loss value based on the predicted noisy image and the training label (the actual noisy image corresponding to the two-dimensional sample image), and adjusting the parameters of the diffusion sub-model according to the loss value.

[0053] The above process, after obtaining the user's input requirements for model creation, uses natural language understanding models and information extraction models to perform semantic parsing and keyword completion operations. This process can effectively reduce the user's expression threshold for model creation requirements, while improving the accuracy of the final generated digital character model.

[0054] Optionally, in some embodiments, the creation interface may include: a digital character model, and model part icons corresponding to each model part in the digital character model.

[0055] The process of displaying input guidance information in the creation interface can include: In response to the selection operation of the target model part icon in the creation interface, input guidance information is displayed in the creation interface; the content creation requirements are the part redrawing requirements for regenerating the target model part corresponding to the target model part icon.

[0056] In response to receiving a creation trigger operation in the creation interface, the content creation model is invoked to execute a content creation task matching the received creation trigger operation based on the content creation requirement information, and the content creation result is obtained, including: In response to receiving a part redrawing trigger operation in the creation interface, the second text image model is invoked to generate the target model part to be redrawn based on the part redrawing requirements.

[0057] Based on the content creation results, update the creation interface, including: In the creation interface, the target model parts in the digital character model are replaced by redrawing the target model parts.

[0058] To illustrate, the above embodiments of this application describe the execution process of a digital character model editing task. Before the digital character model editing task is executed, the digital character model and model part icons corresponding to each model part in the digital character model can be displayed in the content creation interface.

[0059] The digital character model displayed in the content creation interface can be a model obtained through the aforementioned digital character model creation task, or a model created by other platforms and imported into the digital content creation software of this application embodiment. Each model part icon displayed in the content creation interface corresponds to a model part of the aforementioned digital character model.

[0060] The creation interface displays the digital character model and icons for each model part. This allows users to select the target model part to be edited and redrawn from multiple options. Then, input guidance is displayed in the content creation interface, guiding the user to input their redrawing requirements for the target model part. The second-generation image model then regenerates the model part based on these requirements, resulting in the redrawn target model part. Alternatively, users can also regenerate the entire digital character model as a whole. The specific regeneration process is similar to the regeneration process for target model parts described above, and will not be repeated here.

[0061] Through the above process, users can selectively edit the model: they can edit and redraw the entire digital character model, or they can selectively edit and redraw a specific part of the model. This improves the convenience of model editing.

[0062] To illustrate, after obtaining a digital character model, the individual model parts within the model can be extracted as follows: The digital character model is input into a pre-trained semantic segmentation model, which then breaks it down into different model parts. For example, taking a 3D animated character model as an example, a 3D semantic segmentation model can segment the character into parts such as head, hair, legs, and feet. Furthermore, for easier differentiation, a 3D masking mechanism can be used, employing different mask values ​​to represent different model parts.

[0063] For example, the semantic segmentation model described above can be a PointNet++ model, an S-LRM model, a BPNet model, a Point-Voxel CNN model, etc. The training process of the semantic segmentation model may include: obtaining a digital character model to be segmented as a training sample, inputting the training sample into the semantic segmentation model, and outputting the predicted semantic labels of each vertex of the digital character model through the semantic segmentation model; calculating the loss value based on the predicted semantic labels and the training labels (the actual semantic labels of each vertex of the digital character model), adjusting the model parameters according to the loss value, and obtaining the trained semantic segmentation model.

[0064] In the above embodiments of this application, the triggering method for the digital character model editing task is not limited. It can be triggered independently by the user performing the creation trigger operation corresponding to the task, or it can be triggered automatically after the digital character model creation task is completed, by first automatically segmenting the created digital character model into parts, and then automatically starting the digital color correction model editing task based on the pre-set task execution order.

[0065] Correspondingly, the process of calling the second raw image model to generate the target model part for redrawing based on the part redrawing requirements can include: determining the target position information and associated position information according to the mask values ​​corresponding to each vertex in the digital character model; wherein, the target position information is the regional position information of the target model part, and the associated position information is the regional position information of the associated model parts that have a relationship with the target model part; inputting the target position information, associated position information and part redrawing requirements into the second raw image model, and generating the target model part for redrawing through the second raw image model.

[0066] Similar to the first text-based image model, the second text-based image model can also include an image rendering sub-model and a diffusion sub-model. The image rendering sub-model can be a differentiable renderer, such as NeRF or DMTet; the diffusion sub-model can use a Stable Diffusion model, a Point-E model, etc. During the inference phase, a preset initial neural radiation field is obtained; the initial neural radiation field and the redrawing requirements are input into the image rendering sub-model, which renders an initial 2D image; the diffusion sub-model performs noise image prediction based on the initial 2D image, calculates gradients based on the predicted noise images, and sends the calculation results back to the image rendering sub-model to update the initial neural radiation field; the image rendering sub-model renders an updated 2D image based on the updated neural radiation field; and 3D reconstruction is performed based on the target location information, associated location information, and the updated 2D image to obtain the redrawn target model part.

[0067] During the training phase, only the parameters of the diffusion sub-model need to be adjusted; there is no need to adjust the parameters of the image rendering sub-model. The training process of the diffusion sub-model can be referred to the training process of the diffusion sub-model in the first raw image model above, and will not be repeated here.

[0068] The above-described model redrawing process determines the target location information of the target model part based on a masking mechanism, enabling the second text-based image model to redraw the target model part based on this target location information. Therefore, it can be guaranteed that the final redrawn target model part is completely consistent with the original target model part in terms of space occupancy; that is, it ensures that the redrawn target model part has a consistent spatial form with the original target model part, avoiding interference or gaps between the redrawn target model part and other surrounding model parts.

[0069] In addition to inputting the target location information into the second raw image model, the redrawing process also inputs the location information of related model parts that are associated with the target model part. Related model parts are typically those whose spatial form is related to the target model part; for example, when the target model part is the eye, the related model parts could be the eyebrow and the nose. Because the redrawing process also considers the location information of related model parts, the final redrawn target model part and related model parts are more visually harmonious, further improving the quality of content creation.

[0070] In addition, in some other embodiments of this application, a part asset library can be provided, which can store model parts of different shapes. When a user wants to redraw a part of a digital character model, they can search in the part asset library to obtain the target model part to be redrawn. For example, for the eyebrow model part, various eyebrow parts of different shapes can be stored in the part asset library. When a user wants to redraw the eyebrows of a digital character model, they can select a suitable eyebrow from the part asset library as the eyebrow part to be redrawn. By pre-building a part asset library for partial part redrawing, there is no need to perform model generation operations in real time, thus the redrawing efficiency is higher.

[0071] Optionally, in some embodiments, in the creation interface, the icons of each model part are displayed in the form of a hierarchical icon tree; when the first model part belongs to the second model part, the first model part icon corresponding to the first model part is a sub-icon belonging to the second model part icon corresponding to the second model part.

[0072] Indicative, for reference Figure 3In the creation interface, the icons for each model part are displayed in a hierarchical icon tree format. For example, the "head" model part includes multiple parts such as the "left eye," "right eye," and "nose." Therefore, the icon corresponding to the "head" model part can be used as the parent icon (parent node), while the icons corresponding to the "left eye," "right eye," and "nose" can be used as child icons (child nodes). Similarly, other parts can form a similar hierarchical relationship.

[0073] In the above embodiments, the icons of each model part are hierarchically organized according to the hierarchical relationship between different model parts, and thus presented to the user in the form of a hierarchical icon tree. When the user wants to redraw a model part, they can select the parent icon based on this icon tree to achieve simultaneous redrawing of multiple child parts belonging to the same parent part, further improving the efficiency of content creation.

[0074] Optionally, in some embodiments, the creation interface may include: a digital character model; and content creation requirements are age requirements for the digital character model.

[0075] Correspondingly, the process of responding to a creation trigger operation received in the creation interface, invoking the content creation model to execute a content creation task matching the received creation trigger operation based on the content creation requirements, and obtaining the content creation result, may include: In response to receiving an age modification trigger in the creation interface, the facial generation model is invoked to generate the modified facial features based on the age requirement.

[0076] Based on the content creation results, updating the creation interface may include: In the creation interface, the modified facial features replace the original facial features in the digital character model.

[0077] To illustrate, in some creative scenarios, it may be necessary to create digital character models that visually match a certain age group. To address this requirement, this application provides a specific model editing task execution process.

[0078] In this embodiment, the triggering method for the aforementioned facial modification task is not limited. It can be triggered independently by the user performing a creative triggering operation corresponding to the task, or it can be automatically triggered after the completion of the previous task in a workflow that is automatically executed according to a preset task execution order. For example, after the digital character model creation task is completed, the created digital character model can be automatically segmented into model parts, and then the aforementioned facial modification task can be automatically started.

[0079] The creation interface displays input guidance information to guide users in entering the age requirements for the digital character model. For example: Figure 4 The input box displays a prompt "Enter your age". Alternatively, images can be used to guide the user, such as... Figure 4 The icon in the image shows an age modifier. This icon contains multiple levels, each corresponding to a different age range. Users can select a level, and the age range corresponding to that level will be used as the age requirement entered by the user.

[0080] The process of calling the facial generation model to generate modified facial features based on age requirements can be implemented as follows: determine the current facial features based on the mask values ​​corresponding to each vertex in the digital character model; input the current facial features and age requirements into the third-party image model, and generate the modified facial features through the third-party image model.

[0081] The aforementioned third-generation facial image model can include a spatial alignment module and an age-gradient diffusion sub-model. The spatial alignment module primarily uses 3DMM (3D Morphable Models) technology to spatially align the current facial features, obtaining the facial code corresponding to the current age. This module does not have trainable parameters. The age-gradient diffusion sub-model adjusts the facial code corresponding to the current age to a target facial code that meets the age requirement, thus obtaining the modified facial features based on the target facial code. For example, the age-gradient diffusion sub-model can be an AgeDiffuse-5 model, a TADM-3D model, etc. During the inference phase, the current facial features are input into the spatial alignment module to obtain the facial code corresponding to the current age; the facial code corresponding to the current age and the user-inputted age requirement are input into the age-gradient diffusion sub-model, which outputs the target facial code that meets the age requirement; the target facial code is then decoded to obtain the modified facial features.

[0082] During the training phase, only parameter adjustments are needed for the age-gradient diffusion sub-model, without requiring parameters for the spatial alignment module. The training process for the age-gradient diffusion sub-model can include: acquiring facial latent codes and adding noise to them to obtain processed latent codes; inputting the processed latent codes and the target age into the age-gradient diffusion sub-model to obtain predicted latent codes; calculating the loss value based on the predicted latent codes and training labels (the real noise added during the noise addition process), and adjusting the parameters of the age-gradient diffusion sub-model according to the loss value.

[0083] Through the above embodiments, users can perform simple operations in the creation interface to dynamically generate facial features that meet user requirements based on the fusion architecture of 3D Morphable Models and age-gradient diffusion sub-models. Examples include skin wrinkle distribution, bone contour changes, and fat loss simulation that match the user's age. In other words, through simple operations in the creation interface, users can parameterize the human body model and achieve age-appropriate facial feature reconstruction that conforms to physiological laws, further enhancing the user's creative experience.

[0084] Optionally, in some embodiments, the creation interface includes: a digital character model; and texture icons corresponding to various parts of the model in the digital character model. The process of displaying input guidance information in the creation interface can include: In response to the selection of a texture icon for a target part in the creation interface, input guidance information is displayed in the creation interface; the content creation requirements are the texture requirements information used to generate the texture map of the target model part corresponding to the texture icon of the target part.

[0085] The process of responding to a creation trigger operation received in the creation interface, invoking the content creation model to execute a content creation task matching the received creation trigger operation based on the content creation requirements, and obtaining the content creation result, may include: In response to receiving a texture generation trigger operation in the creation interface, the texture generation model is invoked to generate the target texture map based on the texture requirements.

[0086] Based on the content creation results, update the creation interface, including: In the creation interface, add the target texture map to the surface of the target model part in the digital character model.

[0087] Schematic illustration: The above embodiments of this application describe the execution process of a digital character model texture generation task. In these embodiments, the triggering method for the digital character model texture generation task is not limited. It can be triggered independently by the user performing a creation trigger operation corresponding to the task, or it can be automatically triggered after the completion of the previous task in a workflow that is automatically executed according to a preset task execution order. For example, after the digital character model creation task is completed, the created digital character model can be automatically segmented into parts, and then the digital character model texture generation task can be automatically started.

[0088] Before executing the digital character model texture generation task, the digital character model and the corresponding part texture icons for each model part in the digital character model can be displayed in the content creation interface.

[0089] Similar to the digital character model editing task, the digital character model displayed in the content creation interface of this application embodiment can be a model obtained through the above-mentioned digital character model creation task, or a model created by other platforms and imported into the digital content creation software of this application embodiment.

[0090] The texture icons displayed in the content creation interface each correspond to a model part of the aforementioned digital character model. See also... Figure 5 As shown, it includes four texture icons for different body parts: the texture icons for the upper body of the digital character model, and so on. Figure 5 The middle part is labeled "top"; the corresponding texture icon is for the lower body of the digital character model. Figure 5 The middle part is shown as "lower garment"; the corresponding texture icon is for the upper limbs of a digital character model. Figure 5 The icon in the middle is labeled "sleeves"; it corresponds to the texture icon for the foot area of ​​a digital character model. Figure 5 The character in the middle is "shoe".

[0091] The creation interface displays a digital character model to the user, along with texture icons for each model part. Users can select the target model part from multiple model parts to generate texture maps, based on their needs. Figure 5 For example, when a user selects the "Lower Clothes" icon, the creation interface can further display, such as... Figure 6 The input guidance information shown is: "Upload material reference" or "Enter material", etc.

[0092] For example, refer to Figure 7 When a user uploads an image containing reference materials, the digital content creation software calls a texture generation model and inputs the image containing the reference materials into the texture generation model to generate a target texture map. Then, based on the mask values ​​corresponding to each vertex in the digital character model, the position information of the lower body part can be determined, and the generated target texture map can be added to the surface of the lower body part of the digital character model.

[0093] For example, the texture generation model described above can be a UV-aware diffusion model, such as Texture / Paint3D, Meta 3D TextureGen, etc. During inference, the target texture requirements and the UV information of the target model parts can be input into the texture generation model, and the texture generation model outputs a target texture map (such as a high-quality PBR material map) that is geometrically aligned with the target model parts. The training process of the texture generation model can include: inputting the texture requirements and the UV information of the preset model parts into the texture generation model, and outputting a predicted texture map through the texture generation model; calculating the loss value based on the predicted texture map and training labels (texture maps of the preset model parts that meet the texture requirements); adjusting the model parameters according to the loss value to obtain the trained texture generation model.

[0094] The process of calculating the loss value based on the predicted texture map and training labels can include: calculating the color consistency loss value based on the predicted texture map and training labels; the color consistency loss value represents the difference in color features between the predicted texture map and the training labels; calculating the texture consistency loss value based on the predicted texture map and training labels; the texture consistency loss value can usually be represented by the distance between the Gram matrix of the predicted texture map and the Gram matrix of the training labels (such as the Frobenius distance); calculating the illumination response loss value based on the predicted texture map and training labels; the illumination response loss value represents the illumination response loss between the predicted texture map and the training labels, representing the difference in emission amplitude of corresponding material points of the predicted texture map and the training labels in the same ambient illumination field; and fusing the above color consistency loss value, texture consistency loss value, and illumination response loss value to obtain the final loss value.

[0095] Through the above process, users can selectively perform texture mapping operations on models: they can apply texture mapping to the entire digital character model or target specific parts of the model. This improves the convenience of texture mapping.

[0096] Optionally, in some embodiments, the creation interface includes: a digital character model; and content creation requirements are style requirement information for the digital character model.

[0097] Correspondingly, the process of responding to a creation trigger operation received in the creation interface, invoking the content creation model to execute a content creation task matching the received creation trigger operation based on the content creation requirements, and obtaining the content creation result, may include: In response to receiving a style-triggered operation in the creation interface, the style generation model is invoked to generate a stylized character model that matches the style requirements.

[0098] The process of updating the creation interface based on the content creation results may include: The stylized character models are displayed in the creation interface.

[0099] The above process is illustrative of the stylization task for a digital character model. Input guidance information can be displayed in the content creation interface to guide the user in inputting style requirements. For example, see... Figure 8 The input guidance information can be "Upload style reference" or "Enter style description"; it can also be a style icon corresponding to different styles.

[0100] For example, refer to Figure 9 When a user inputs the style requirement of "hand-drawn style", the digital content creation software calls the style generation model corresponding to "hand-drawn style" to generate a stylized texture map; the stylized texture map is then added to the surface of the digital character model to obtain a stylized character model.

[0101] The process of obtaining the style generation model includes: determining the target stylized low-rank adaptation model that matches the style requirements of the user input from a number of pre-defined stylized low-rank adaptation (LoRA) models; and fusing the weight information of the target stylized low-rank adaptation model into the main style generation model (for example, injecting LoRA weights into the attention layer of the main style generation model) to obtain the style generation model.

[0102] For example, the main style generation model can be Stable Diffusion, Imagen, etc. The stylized low-order adaptation model can refer to a low-rank matrix pair inserted into the main style generation model (such as an attention layer or a feedforward layer). The weight information of the stylized low-order adaptation model is the value information of each element in the aforementioned low-rank matrix pair. The low-rank matrix pair corresponds one-to-one with a style. The process of obtaining the low-rank matrix pair can include: inserting a trainable matrix pair (A, B) into the main style generation model (such as an attention layer or a feedforward layer); training the model based on training samples with a preset style, keeping the parameters of the main style generation model unchanged during training, and adjusting the values ​​of the elements in the trainable matrix pair (A, B) to obtain the low-rank matrix pair corresponding to the aforementioned preset style, that is: the stylized low-order adaptation model corresponding to the aforementioned preset style.

[0103] The main style generation model can include an encoder and a normalization layer (such as AdaIN). After obtaining the style requirement information, the encoder can extract style features and the normalization layer can normalize these features. Then, based on the UV geometry information of the digital character model and the normalized style features, a stylized texture map that meets the style requirements is generated. Finally, the generated stylized texture map is added to the surface of the digital character model to obtain the stylized character model.

[0104] The above embodiments embed model stylization capabilities based on style generation models into digital content creation software, enabling the software to fully leverage the potential of artificial intelligence models to simplify and accelerate the model stylization process. When users create content using the software, they do not need to perform complex platform switching or data import / export operations. They can seamlessly switch to the AI-driven model stylization mode through a simple creation trigger, thereby quickly and efficiently achieving cross-domain style mapping from 2D style images to 3D character models.

[0105] Furthermore, stylization of any local model part in a digital character model can be achieved through the following steps: After obtaining the style requirement information, style features are extracted using an encoder and normalized using a normalization layer. Then, based on the UV geometry information of the local model part and the normalized style features, a stylized texture map that meets the style requirements is generated. The generated stylized texture map is then added to the surface of the local model part to obtain the stylized local model part. This process, by extracting style features through an encoder and combining them with a local fusion mechanism for digital character models, can accurately transfer a specified style to a designated model part of the digital character model, further enhancing the content creation experience.

[0106] See Figure 10 , Figure 10 This is a structural block diagram of a content creation apparatus according to an embodiment of this application. The content creation apparatus provided in this application includes: The creation interface display module 1002 is used to display the creation interface through digital content creation software. The intelligent creation module 1004 is used to receive creation trigger operations in the creation interface, call the content creation model to execute the content creation task that matches the received creation trigger operation, and obtain the content creation result; the content creation model is an artificial intelligence model used to execute content creation tasks. The creation interface update module 1006 is used to update the creation interface based on the content creation results.

[0107] Optionally, in some embodiments, multiple types of creation triggering operations are pre-set, with each creation triggering operation corresponding to a content creation task; multiple types of content creation models are pre-configured, with each content creation model used to perform a content creation task. The intelligent creation module 1004, when executing the step of calling the content creation model to perform a content creation task that matches the received creation trigger operation and obtaining the content creation result, is specifically used for: Among various content creation models, identify the target content creation model that matches the received creation trigger operation; The target content creation model is invoked to execute the content creation task that matches the received creation trigger operation, and the content creation result is obtained.

[0108] Optionally, in some embodiments, multiple types of content creation models are pre-configured, with one content creation model used to perform a content creation task.

[0109] The intelligent creation module 1004, when executing the step of calling the content creation model to perform a content creation task that matches the received creation trigger operation and obtaining the content creation result, is specifically used for: Determine the current target content creation task to be executed according to the pre-set task execution order; Call the target content creation model corresponding to the target content creation task and execute the target content creation task; In response to the completion of the target content creation task, return to the steps of the target content creation task to be executed according to the pre-set task execution order, until the content creation result corresponding to the last content creation task is obtained; Among them, the last content creation task is the latest content creation task to be executed, determined according to the task execution order.

[0110] Optionally, in some embodiments, the content creation task includes at least one of the following: digital character model creation task, digital character model editing task, digital character model stylization task, and digital character model texture generation task.

[0111] Optionally, in some embodiments, the intelligent authoring module 1004 is specifically used for: The input guidance information is displayed in the creation interface; the input guidance information is used to guide the input operations required for content creation. Receive and submit content creation requirements; In response to receiving a creation trigger operation in the creation interface, the content creation model is invoked to execute a content creation task that matches the received creation trigger operation based on the content creation requirements, and the content creation result is obtained.

[0112] Optionally, in some embodiments, the content creation requirements are model creation requirements for creating digital character models; The intelligent creation module 1004, in the step of responding to a creation trigger operation received in the creation interface, calling the content creation model based on the content creation requirement information, executing a content creation task matching the received creation trigger operation, and obtaining the content creation result, is specifically used for: In response to receiving a model creation trigger operation in the creation interface, the first text-based image model is invoked to generate a digital character model that matches the model creation requirements. Creation interface update module 1006, specifically used for: Display digital character models in the creation interface.

[0113] Optionally, in some embodiments, the creation interface includes: a digital character model, and model part icons corresponding to each model part in the digital character model; The intelligent creation module 1004, when executing the step of displaying input guidance information in the creation interface, is specifically used for: In response to the selection operation of the target model part icon in the creation interface, input guidance information is displayed in the creation interface; the content creation requirement is the part redrawing requirement for regenerating the target model part corresponding to the target model part icon; The intelligent creation module 1004, in the step of responding to a creation trigger operation received in the creation interface, calling the content creation model based on the content creation requirement information, executing a content creation task matching the received creation trigger operation, and obtaining the content creation result, is specifically used for: In response to receiving a part redrawing trigger operation in the creation interface, the second text image model is invoked to generate the target model part to be redrawn based on the part redrawing requirements. Creation interface update module 1006, specifically used for: In the creation interface, the target model parts in the digital character model are replaced by redrawing the target model parts.

[0114] Optionally, in some embodiments, in the creation interface, the icons of each model part are displayed in the form of a hierarchical icon tree; when the first model part belongs to the second model part, the first model part icon corresponding to the first model part is a sub-icon belonging to the second model part icon corresponding to the second model part.

[0115] Optionally, in some embodiments, the creation interface includes: a digital character model; and the content creation requirements are age requirements for the digital character model. The intelligent creation module 1004, in the step of responding to a creation trigger operation received in the creation interface, calling the content creation model based on the content creation requirement information, executing a content creation task matching the received creation trigger operation, and obtaining the content creation result, is specifically used for: In response to receiving an age modification trigger operation in the creation interface, the facial generation model is invoked to generate modified facial features based on the age requirement; Creation interface update module 1006, specifically used for: In the creation interface, the modified facial features replace the original facial features in the digital character model.

[0116] Optionally, in some embodiments, the creation interface includes: a digital character model; and texture icons corresponding to various parts of the model in the digital character model. The intelligent creation module 1004, when executing the step of displaying input guidance information in the creation interface, is specifically used for: In response to the selection of a texture icon for a target part in the creation interface, input guidance information is displayed in the creation interface; the content creation requirements are the texture requirements information used to generate the texture map of the target model part corresponding to the texture icon of the target part. The intelligent creation module 1004, in the step of responding to a creation trigger operation received in the creation interface, calling the content creation model based on the content creation requirement information, executing a content creation task matching the received creation trigger operation, and obtaining the content creation result, is specifically used for: In response to receiving a texture generation trigger operation in the creation interface, the texture generation model is invoked to generate the target texture map based on the texture requirements; Creation interface update module 1006, specifically used for: In the creation interface, add the target texture map to the surface of the target model part in the digital character model.

[0117] Optionally, in some embodiments, the creation interface includes: a digital character model; and content creation requirements are style requirement information for the digital character model. The intelligent creation module 1004, in the step of responding to a creation trigger operation received in the creation interface, calling the content creation model based on the content creation requirement information, executing a content creation task matching the received creation trigger operation, and obtaining the content creation result, is specifically used for: In response to receiving a style-triggered operation in the creation interface, the style generation model is invoked to generate a stylized character model that matches the style requirements. Creation interface update module 1006, specifically used for: The stylized character models are displayed in the creation interface.

[0118] The content creation apparatus of this embodiment is used to implement the corresponding content creation methods in the foregoing embodiments and has the beneficial effects of the corresponding method embodiments, which will not be repeated here. Furthermore, the functional implementation of each module in the content creation apparatus of this embodiment can be referred to the description of the corresponding part in the foregoing method embodiments, which will also not be repeated here.

[0119] Reference Figure 11 This document illustrates a schematic diagram of an electronic device according to an embodiment of this application. The specific embodiments of this application do not limit the specific implementation of the electronic device.

[0120] like Figure 11 As shown, the electronic device may include: a processor 1102, a communications interface 1104, a memory 1106, and a communications bus 1108.

[0121] in: The processor 1102, communication interface 1104, and memory 1106 communicate with each other via communication bus 1108.

[0122] Communication interface 1104 is used for communication with other electronic devices.

[0123] The processor 1102 is used to execute program 1110, specifically the relevant steps in the above method embodiments.

[0124] Schematic, program 1110 may include program code that includes computer operation instructions.

[0125] The processor 1102 may be a CPU, an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement the embodiments of this application. The smart device may include one or more processors of the same type, such as one or more CPUs; or it may include processors of different types, such as one or more CPUs and one or more ASICs.

[0126] Memory 1106 is used to store program 1110. Memory 1106 may include high-speed RAM memory, and may also include non-volatile memory, such as at least one disk storage device.

[0127] Program 1110 may include multiple computer instructions. Specifically, program 1110 may use multiple computer instructions to cause processor 1102 to perform the operation corresponding to any of the methods described in the foregoing multiple method embodiments.

[0128] The specific implementation of each step in program 1110 can be found in the corresponding descriptions of the steps and units in the above method embodiments, and has corresponding beneficial effects, which will not be repeated here. Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working process of the devices and modules described above can be referred to the corresponding process descriptions in the foregoing method embodiments, and will not be repeated here.

[0129] This application also provides a computer storage medium storing a computer program thereon, which, when executed by a processor, implements the method described in any of the foregoing method embodiments. The computer storage medium includes, but is not limited to, compact disc read-only memory (CD-ROM), random access memory (RAM), floppy disk, hard disk, or magneto-optical disk.

[0130] This application also provides a computer program product, including computer instructions that instruct a computing device to perform an operation corresponding to any of the methods in the above-described multiple method embodiments.

[0131] Furthermore, it should be noted that the user-related information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to sample data used for training the model, data used for analysis, stored data, displayed data, etc.) involved in the embodiments of this application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use and processing of related data must comply with the relevant laws, regulations and standards of the relevant countries and regions, and corresponding operation entry points are provided for users to choose to authorize or refuse.

[0132] It should be noted that, depending on the implementation needs, the various components / steps described in the embodiments of this application can be broken down into more components / steps, or two or more components / steps or parts of the operation of components / steps can be combined into new components / steps to achieve the purpose of the embodiments of this application.

[0133] The methods described in the embodiments of this application can be implemented in hardware, firmware, or as software or computer code that can be stored in a recording medium (such as a CD-ROM, RAM, floppy disk, hard disk, or magneto-optical disk), or as computer code downloaded over a network that is originally stored in a remote recording medium or a non-transitory machine-readable medium and will be stored in a local recording medium. Thus, the methods described herein can be stored on a recording medium using a general-purpose computer, a dedicated processor, or programmable or dedicated hardware (such as an Application Specific Integrated Circuit (ASIC) or a Field Programmable Gate Array (FPGA)). It is understood that the computer, processor, microprocessor controller, or programmable hardware includes storage components (e.g., Random Access Memory (RAM), Read-Only Memory (ROM), Flash Memory, etc.) capable of storing or receiving software or computer code, which, when accessed and executed by the computer, processor, or hardware, implements the methods described herein. Furthermore, when a general-purpose computer accesses code used to implement the methods shown herein, the execution of the code transforms the general-purpose computer into a dedicated computer for executing the methods shown herein.

[0134] Those skilled in the art will recognize that the units and method steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the embodiments of this application.

[0135] The above embodiments are only used to illustrate the embodiments of this application, and are not intended to limit the embodiments of this application. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of the embodiments of this application. Therefore, all equivalent technical solutions also fall within the scope of the embodiments of this application, and the patent protection scope of the embodiments of this application should be defined by the claims.

Claims

1. A content creation method, applied to an electronic device equipped with digital content creation software, the method comprising: The creation interface is displayed through the digital content creation software. The system receives a creation trigger operation in the creation interface, calls a content creation model to execute a content creation task that matches the received creation trigger operation, and obtains a content creation result; the content creation model is an artificial intelligence model used to execute the content creation task. Update the creation interface based on the content creation results.

2. The method according to claim 1, wherein, Multiple types of creation trigger operations are pre-defined, with each creation trigger operation corresponding to a content creation task; multiple types of content creation models are pre-configured, with each content creation model used to execute a content creation task. The invocation of the content creation model to execute a content creation task matching the received creation trigger operation, and obtaining the content creation result, includes: Among various content creation models, determine the target content creation model that matches the received creation trigger operation; The target content creation model is invoked to execute a content creation task that matches the received creation trigger operation, thereby obtaining the content creation result.

3. The method according to claim 1, wherein, Multiple types of content creation models are pre-configured, with one content creation model used to perform one type of content creation task; The invocation of the content creation model to execute a content creation task matching the received creation trigger operation, and obtaining the content creation result, includes: Determine the current target content creation task to be executed according to the pre-set task execution order; Invoke the target content creation model corresponding to the target content creation task and execute the target content creation task; In response to the completion of the target content creation task, the process returns to the step of determining the target content creation task to be executed according to the pre-set task execution order, until the content creation result corresponding to the last content creation task is obtained; The last content creation task is the latest content creation task to be executed, as determined by the task execution order.

4. The method according to claim 2 or 3, wherein, The content creation task includes at least one of the following: digital character model creation task, digital character model editing task, digital character model stylization task, and digital character model texture generation task.

5. The method according to any one of claims 1-3, wherein, Before receiving the creation trigger operation in the creation interface, the method further includes: The creation interface displays input guidance information; this information is used to guide the input operations required for content creation. The process of receiving a creation trigger operation in the creation interface, invoking the content creation model to execute a content creation task matching the received creation trigger operation, and obtaining a content creation result includes: Receive and submit content creation requirements; In response to receiving a creation trigger operation in the creation interface, the content creation model is invoked to execute a content creation task that matches the received creation trigger operation based on the content creation requirements, and a content creation result is obtained.

6. The method according to claim 5, wherein, The content creation requirements are the model creation requirements used to create digital character models; The response to receiving a creation trigger operation in the creation interface involves invoking a content creation model based on the content creation requirement information to execute a content creation task matching the received creation trigger operation, thereby obtaining a content creation result, including: In response to receiving a model creation trigger operation in the creation interface, the first text-based image model is invoked to generate a digital character model that matches the model creation requirements based on the model creation requirements; The step of updating the creation interface based on the content creation results includes: The digital character model is displayed in the creation interface.

7. The method according to claim 5, wherein, The creation interface includes: a digital character model, and model part icons corresponding to each model part in the digital character model; The display of input guidance information in the creation interface includes: In response to the selection operation of the target model part icon in the creation interface, input guidance information is displayed in the creation interface; the content creation requirements are the part redrawing requirements for regenerating the target model part corresponding to the target model part icon; The response to receiving a creation trigger operation in the creation interface involves invoking a content creation model based on the content creation requirement information to execute a content creation task matching the received creation trigger operation, thereby obtaining a content creation result, including: In response to receiving a part redrawing trigger operation in the creation interface, the second text image model is invoked to generate the redrawing target model part based on the part redrawing requirements; The step of updating the creation interface based on the content creation results includes: In the creation interface, the target model part in the digital character model is replaced by the redrawn target model part.

8. The method according to claim 7, wherein, In the creation interface, the icons of each model part are displayed in the form of a hierarchical icon tree; when the first model part belongs to the second model part, the first model part icon corresponding to the first model part is a sub-icon belonging to the second model part icon corresponding to the second model part.

9. The method according to claim 5, wherein, The creation interface includes: a digital character model; the content creation requirements are age requirements for the digital character model; The response to receiving a creation trigger operation in the creation interface involves invoking a content creation model to execute a content creation task matching the received creation trigger operation based on the content creation requirements, thereby obtaining a content creation result, including: In response to receiving an age modification trigger operation in the creation interface, the facial generation model is invoked to generate modified facial features based on the age requirement; The step of updating the creation interface based on the content creation results includes: In the creation interface, the modified facial features are used to replace the original facial features in the digital character model.

10. The method according to claim 5, wherein, The creation interface includes: a digital character model; and texture icons corresponding to each part of the model in the digital character model. The display of input guidance information in the creation interface includes: In response to the selection operation of the target part texture icon in the creation interface, input guidance information is displayed in the creation interface; the content creation requirements are texture requirement information for generating the texture map of the target model part corresponding to the target part texture icon; The response to receiving a creation trigger operation in the creation interface involves invoking a content creation model to execute a content creation task matching the received creation trigger operation based on the content creation requirements, thereby obtaining a content creation result, including: In response to receiving a texture generation trigger operation in the creation interface, the texture generation model is invoked to generate a target texture map based on the texture requirements; The step of updating the creation interface based on the content creation results includes: In the creation interface, the target texture map is added to the surface of the target model part in the digital character model.

11. The method according to claim 5, wherein, The creation interface includes: a digital character model; the content creation requirements are style requirements for the digital character model; The response to receiving a creation trigger operation in the creation interface involves invoking a content creation model to execute a content creation task matching the received creation trigger operation based on the content creation requirements, thereby obtaining a content creation result, including: In response to receiving a style trigger operation in the creation interface, the style generation model is invoked to generate a stylized character model that matches the style requirements. The step of updating the creation interface based on the content creation results includes: The stylized character model is displayed in the creation interface.

12. A content creation device located on an electronic device on which digital content creation software is deployed, the device comprising: The creation interface display module is used to display the creation interface through the digital content creation software. The intelligent creation module is used to receive creation trigger operations in the creation interface, call the content creation model to execute the content creation task that matches the received creation trigger operation, and obtain the content creation result; the content creation model is an artificial intelligence model used to execute the content creation task. The creation interface update module is used to update the creation interface based on the content creation results.

13. An electronic device, comprising: The processor, memory, communication interface, and communication bus are provided, wherein the processor, memory, and communication interface communicate with each other via the communication bus. The memory is used to store at least one executable instruction that causes the processor to perform the operation corresponding to the method as described in any one of claims 1-11.

14. A computer storage medium having a computer program stored thereon, which, when executed by a processor, implements the method as described in any one of claims 1-11.

15. A computer program product comprising computer instructions that instruct a computing device to perform an operation corresponding to any one of the methods described in claims 1-11.

Images