Editing method and device of three-dimensional model chartlet, electronic equipment, storage medium and program product

Abstract

The invention relates to a three-dimensional model mapping editing method and device, electronic equipment, a storage medium and a program product. The method comprises the steps of obtaining a to-be-edited target three-dimensional model map; obtaining target text description information used for editing the target three-dimensional model map; and editing the target three-dimensional model map according to the target text description information to obtain an edited three-dimensional model map. According to the method and the device, the three-dimensional model chartlet is automatically edited according to the input text description information, so that the editing is more flexible, and the editing effect is more diversified. By adopting the embodiment of the invention, the user can guide the editing of the three-dimensional model chartlet in a natural language description mode. The editing mode based on text description better conforms to the visual expression habit of human beings, the technical threshold is lowered, and more users can easily conduct personalized customization of the three-dimensional model chartlet.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to the technical field of computer, and particularly relates to a three-dimensional model map editing method, a three-dimensional model map editing device, an electronic device, a computer readable storage medium and a computer program product. BACKGROUND

[0002] At present, when a 3D (3 Dimensions) face map is edited, attribute editing is mainly based on a 2D (2 Dimensions) face image. For example, a 2D face image can be first input into a pre-trained face attribute editing model to obtain a face edited image; then target segmentation and portrait segmentation are respectively performed on the input 2D face image and the edited image to obtain a target domain mask image and a target portrait mask image, and a difference set of the two is calculated; finally, the images before and after editing are fused, and the fused image and the difference set are input into an image inpainting model together to output a target attribute edited face image.

[0003] When the related technology applies a 2D face image to a 3D face map scene, a 3D face map needs to be regenerated by using the edited image, which is very time-consuming and may change the mesh of the 3D face. SUMMARY

[0004] The present disclosure provides a three-dimensional model map editing technical solution.

[0005] According to an aspect of the present disclosure, a three-dimensional model map editing method is provided, comprising:

[0006] obtaining a target three-dimensional model map to be edited;

[0007] obtaining target text description information for editing the target three-dimensional model map;

[0008] editing the target three-dimensional model map according to the target text description information to obtain an edited three-dimensional model map.

[0009] In a possible implementation manner, the editing the target three-dimensional model map according to the target text description information to obtain an edited three-dimensional model map comprises:

[0010] generating an editable latent code corresponding to the target three-dimensional model map;

[0011] generating an edited latent code according to the target text description information and the editable latent code;

[0012] generating an edited three-dimensional model map according to the edited latent code.

[0013] In a possible implementation, the generating the editable latent code corresponding to the target three-dimensional model map comprises:

[0014] In response to the target text description information being used for overall editing of the target three-dimensional model map, the editable latent code corresponding to the target three-dimensional model map is generated.

[0015] In a possible implementation, the generating the edited latent code according to the target text description information and the editable latent code comprises:

[0016] training a mapping model according to the target text description information and the editable latent code;

[0017] In response to the mapping model being trained, inputting the target text description information and the editable latent code into the trained mapping model, and generating the edited latent code by using the trained mapping model.

[0018] In a possible implementation, the training the mapping model according to the target text description information and the editable latent code comprises:

[0019] inputting the target text description information and the editable latent code into the mapping model, and generating a predicted edited latent code by using the mapping model;

[0020] obtaining a predicted rendering image corresponding to the predicted edited latent code;

[0021] training the mapping model according to the predicted rendering image and the target text description information.

[0022] In a possible implementation, the obtaining the predicted rendering image corresponding to the predicted edited latent code comprises:

[0023] generating a predicted edited three-dimensional model map corresponding to the predicted edited latent code;

[0024] attaching the predicted edited three-dimensional model map to a mesh of a target three-dimensional model corresponding to the target three-dimensional model map, to obtain the predicted rendering image corresponding to the predicted edited latent code.

[0025] In a possible implementation, the training the mapping model according to the predicted rendering image and the target text description information comprises:

[0026] determining a value of a loss function corresponding to the mapping model according to similarity between the predicted rendering image and the target text description information;

[0027] According to the value of the loss function, the mapping model is trained.

[0028] In a possible implementation, the generating the edited potential code according to the target text description information and the editable potential code comprises:

[0029] A target mapping model is determined.

[0030] The target text description information and the editable potential code are input into the target mapping model, and the edited potential code is generated by the target mapping model.

[0031] In a possible implementation, the determining the target mapping model comprises:

[0032] A keyword of the target text description information is determined.

[0033] According to the keyword, a target mapping model is determined.

[0034] In a possible implementation, the generating the editable potential code corresponding to the target three-dimensional model map comprises:

[0035] The target three-dimensional model map corresponding to the potential code to be optimized is obtained.

[0036] A three-dimensional model map corresponding to the potential code to be optimized is generated.

[0037] According to the three-dimensional model map corresponding to the potential code to be optimized and the target three-dimensional model map, the potential code to be optimized is optimized until a preset stop optimization condition is met, and the editable potential code corresponding to the target three-dimensional model map is obtained.

[0038] In a possible implementation, the generating the editable potential code corresponding to the target three-dimensional model map comprises:

[0039] The target three-dimensional model map is input into a pre-trained inverse model, and the editable potential code corresponding to the target three-dimensional model map is generated by the inverse model.

[0040] In a possible implementation, the generating the edited three-dimensional model map according to the edited potential code comprises:

[0041] The edited potential code is input into a first preset generation model, and the edited three-dimensional model map is generated by the first preset generation model.

[0042] In a possible implementation, the first preset generation model is a generative adversarial model.

[0043] In a possible implementation, the editing the target three-dimensional model map according to the target text description information comprises:

[0044] obtaining position information of a target editing area of the target three-dimensional model map;

[0045] inputting the target three-dimensional model map, the position information of the target editing area, and the target text description information into a second preset generation model, and generating an edited three-dimensional model map by using the second preset generation model.

[0046] In a possible implementation, the obtaining the position information of the target editing area of the target three-dimensional model map comprises:

[0047] determining the position information of the target editing area of the target three-dimensional model map according to an area painted by a user on the target three-dimensional model map.

[0048] In a possible implementation, the second preset generation model is a diffusion process-based image generation model.

[0049] In a possible implementation, the target three-dimensional model map is a three-dimensional face map.

[0050] According to an aspect of the present disclosure, a three-dimensional model map editing device is provided, which comprises:

[0051] a first obtaining module configured to obtain a target three-dimensional model map to be edited;

[0052] a second obtaining module configured to obtain target text description information used for editing the target three-dimensional model map;

[0053] an editing module configured to edit the target three-dimensional model map according to the target text description information, and obtain an edited three-dimensional model map.

[0054] In a possible implementation, the editing module is configured to:

[0055] generate an editable latent code corresponding to the target three-dimensional model map;

[0056] generate an edited latent code according to the target text description information and the editable latent code;

[0057] generate an edited three-dimensional model map according to the edited latent code.

[0058] In a possible implementation, the editing module is configured to:

[0059] In response to the target text description information being used for overall editing of the target three-dimensional model map, an editable latent code corresponding to the target three-dimensional model map is generated.

[0060] In a possible implementation, the editing module is configured to:

[0061] training a mapping model according to the target text description information and the editable latent code;

[0062] In response to the mapping model being trained, inputting the target text description information and the editable latent code into the trained mapping model, and generating an edited latent code by the trained mapping model.

[0063] In a possible implementation, the editing module is configured to:

[0064] inputting the target text description information and the editable latent code into the mapping model, and generating a predicted edited latent code by the mapping model;

[0065] obtaining a predicted rendering image corresponding to the predicted edited latent code;

[0066] training the mapping model according to the predicted rendering image and the target text description information.

[0067] In a possible implementation, the editing module is configured to:

[0068] generating a predicted edited three-dimensional model map corresponding to the predicted edited latent code;

[0069] attaching the predicted edited three-dimensional model map to a mesh of a target three-dimensional model corresponding to the target three-dimensional model map, to obtain a predicted rendering image corresponding to the predicted edited latent code.

[0070] In a possible implementation, the editing module is configured to:

[0071] determining a value of a loss function of the mapping model according to similarity between the predicted rendering image and the target text description information;

[0072] training the mapping model according to the value of the loss function.

[0073] In a possible implementation, the editing module is configured to:

[0074] determining a target mapping model;

[0075] input the target text description information and the editable potential code into the target mapping model, and generate an edited potential code through the target mapping model.

[0076] In a possible implementation, the editing module is configured to:

[0077] determine a keyword of the target text description information;

[0078] determine a target mapping model according to the keyword.

[0079] In a possible implementation, the editing module is configured to:

[0080] obtain a to-be-optimized potential code corresponding to the target three-dimensional model map;

[0081] generate a three-dimensional model map corresponding to the to-be-optimized potential code;

[0082] optimize the to-be-optimized potential code according to the three-dimensional model map corresponding to the to-be-optimized potential code and the target three-dimensional model map until a preset stop optimizing condition is met, to obtain an editable potential code corresponding to the target three-dimensional model map.

[0083] In a possible implementation, the editing module is configured to:

[0084] input the target three-dimensional model map into a pre-trained inverse model, and generate an editable potential code corresponding to the target three-dimensional model map through the inverse model.

[0085] In a possible implementation, the editing module is configured to:

[0086] input the edited potential code into a first preset generation model, and generate an edited three-dimensional model map through the first preset generation model.

[0087] In a possible implementation, the first preset generation model is a generative adversarial model.

[0088] In a possible implementation, the editing module is configured to:

[0089] obtain position information of a target editing area of the target three-dimensional model map;

[0090] input the target three-dimensional model map, the position information of the target editing area, and the target text description information into a second preset generation model, and generate an edited three-dimensional model map through the second preset generation model.

[0091] In a possible implementation, the editing module is configured to:

[0092] According to a region painted by the user for the target three-dimensional model map, position information of a target editing region of the target three-dimensional model map is determined.

[0093] In a possible implementation, the second preset generation model is an image generation model based on a diffusion process.

[0094] In a possible implementation, the target three-dimensional model map is a three-dimensional face map.

[0095] According to an aspect of the present disclosure, an electronic device is provided, comprising: one or more processors; a memory for storing executable instructions; wherein the one or more processors are configured to invoke the executable instructions stored in the memory to perform the above method.

[0096] According to an aspect of the present disclosure, a computer-readable storage medium is provided, having stored thereon computer program instructions, which, when executed by a processor, implement the above method.

[0097] According to an aspect of the present disclosure, a computer program product is provided, comprising computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, when the computer-readable code is run in an electronic device, a processor in the electronic device performs the above method.

[0098] In the embodiments of the present disclosure, by obtaining a target three-dimensional model map to be edited, obtaining target text description information for editing the target three-dimensional model map, and editing the target three-dimensional model map according to the target text description information to obtain an edited three-dimensional model map, the three-dimensional model map can be directly edited based on the text description information without changing the mesh of the three-dimensional model, thereby improving the flexibility of three-dimensional model map editing. In related technologies, the use of a trained attribute editing model is not flexible enough and the editing function is very limited. In the embodiments of the present disclosure, by automatically editing the three-dimensional model map according to the input text description information, the editing can be more flexible and the editing effect can be more diverse. By adopting the embodiments of the present disclosure, the user can guide the editing of the three-dimensional model map in a natural language description manner. This text description-based editing method is more in line with human intuitive expression habits, reduces the technical threshold, and enables more users to easily customize the three-dimensional model map.

[0099] It should be understood that the foregoing general description and the following detailed description are only exemplary and explanatory, rather than limiting the present disclosure.

[0100] Other features and aspects of the present disclosure will become apparent from a detailed description of exemplary embodiments with reference to the following drawings. BRIEF DESCRIPTION OF DRAWINGS

[0101] The accompanying drawings, which are incorporated herein and constitute part of the specification, illustrate embodiments consistent with the present disclosure and, together with the description, further serve to explain the principles of the present disclosure.

[0102] Figure 1 A flowchart of an editing method of a three-dimensional model map provided by an embodiment of the present disclosure is shown.

[0103] Figure 2 A schematic diagram of an editing flow of a three-dimensional model map provided by an embodiment of the present disclosure is shown.

[0104] Figure 3 A schematic diagram of overall editing in an editing method of a three-dimensional model map provided by an embodiment of the present disclosure is shown.

[0105] Figure 4 A schematic diagram of local editing in an editing method of a three-dimensional model map provided by an embodiment of the present disclosure is shown.

[0106] Figure 5 A block diagram of an editing device of a three-dimensional model map provided by an embodiment of the present disclosure is shown.

[0107] Figure 6 A block diagram of an electronic device 1900 provided by an embodiment of the present disclosure is shown. DETAILED DESCRIPTION

[0108] Various exemplary embodiments, features and aspects of the present disclosure will be explained in detail below with reference to the drawings. The same reference numbers in the drawings denote the same or similar elements. Although various aspects of the embodiments are illustrated in the drawings, the drawings are not necessarily drawn to scale unless specifically indicated.

[0109] The term "exemplary" is used herein to mean "serving as an example, instance, or illustration." Any implementation described herein as "exemplary" is not necessarily to be construed as preferred or advantageous over other implementations.

[0110] The term "and / or" used herein only means an association relationship of the associated objects, and means that there can be three relationships, for example, A and / or B can mean that A exists alone, A and B exist together, and B exists alone. In addition, the term "at least one" herein means any one of a plurality or any combination of at least two of a plurality, for example, including at least one of A, B, and C can mean including any one or more elements selected from the set consisting of A, B, and C.

[0111] In addition, for a better understanding of the present disclosure, numerous specific details are given in the following detailed description. It will be understood by those skilled in the art that the present disclosure can be practiced without certain specific details. In some instances, well-known methods, apparatuses, elements and circuits have not been described in detail in order to emphasize the principles of the present disclosure.

[0112] The embodiment of the present disclosure provides a three-dimensional model map editing method. The method comprises the following steps: obtaining a target three-dimensional model map to be edited; obtaining target text description information used for editing the target three-dimensional model map; and editing the target three-dimensional model map according to the target text description information to obtain an edited three-dimensional model map. Thus, the three-dimensional model map can be directly edited based on the text description information without changing the mesh of the three-dimensional model, thereby improving the flexibility of the three-dimensional model map editing. In the related art, the attribute editing model trained is not flexible enough and the editing function is very limited. In the embodiment of the present disclosure, the three-dimensional model map is automatically edited according to the input text description information, which makes the editing more flexible and the editing effect more diverse. By using the embodiment of the present disclosure, the user can guide the editing of the three-dimensional model map in the form of natural language description. This text description-based editing method is more in line with the intuitive expression habits of human beings, reduces the technical threshold, and enables more users to easily customize the three-dimensional model map.

[0113] The mesh can represent the structure of the surface of the three-dimensional model. The three-dimensional model is spliced by polygons, and the polygon is actually spliced by multiple triangles. Therefore, the surface of a three-dimensional model is actually composed of multiple triangular faces connected to each other. In three-dimensional space, the set of points and edges constituting these triangles can be referred to as a mesh.

[0114] The three-dimensional model map editing method provided by the embodiment of the present disclosure will be described in detail below in combination with the drawings.

[0115] Figure 1A flowchart of the method for editing a three-dimensional model map provided by an embodiment of the present disclosure is shown. In a possible implementation, the execution subject of the method for editing a three-dimensional model map can be an editing device for a three-dimensional model map, for example, the method for editing a three-dimensional model map can be executed by a terminal device or a server or other electronic device. The terminal device can be a user equipment (UE), a mobile device, a user terminal, a terminal, a cellular phone, a cordless phone, a personal digital assistant (PDA), a handheld device, a computing device, a vehicle-mounted device, or a wearable device, etc. In some possible implementations, the method for editing a three-dimensional model map can be implemented by a processor invoking computer-readable instructions stored in a memory. As shown in Figure 1 The method for editing a three-dimensional model map includes steps S11 to S13.

[0116] In step S11, a target three-dimensional model map to be edited is obtained.

[0117] In step S12, target text description information for editing the target three-dimensional model map is obtained.

[0118] In step S13, the target three-dimensional model map is edited according to the target text description information, to obtain an edited three-dimensional model map.

[0119] In an embodiment of the present disclosure, the target three-dimensional model map can represent any three-dimensional model map requested by a user to edit.

[0120] In a possible implementation, the target three-dimensional model map is a three-dimensional face map. In this implementation, the three-dimensional model can be a three-dimensional face, and the three-dimensional model map can be a three-dimensional face map.

[0121] In this implementation, by obtaining a target 3D face map to be edited, obtaining target text description information for editing the target 3D face map, and editing the target 3D face map according to the target text description information to obtain an edited 3D face map, the 3D face map can be directly edited based on the text description information without changing the mesh of the 3D face, thereby improving the flexibility of editing the 3D face map. In related technologies, using a trained face attribute editing model is not flexible enough, and the editing function is very limited. In this implementation, the 3D face map is automatically edited according to the input text description information, which can make the editing more flexible and the editing effect more diverse.

[0122] Of course, the embodiments of the present disclosure can also be applied to editing other types of three-dimensional model maps. For example, the three-dimensional model can also be a three-dimensional car, a three-dimensional household appliance, a three-dimensional toy, a three-dimensional building, and the like. Correspondingly, the three-dimensional model map can be a three-dimensional car map, a three-dimensional household appliance map, a three-dimensional toy map, a three-dimensional building map, and the like.

[0123] In the embodiments of the present disclosure, the target text description information can represent text description information used for editing the target three-dimensional model map. The target text description information can be any information input by the user.

[0124] In the embodiments of the present disclosure, after obtaining the target three-dimensional model map to be edited and obtaining the target text description information used for editing the target three-dimensional model map, the target three-dimensional model map can be edited according to the target text description information to obtain an edited three-dimensional model map.

[0125] For example, the target three-dimensional model map is a 3D face map, and the target text description information input by the user is "a cartoon face". According to the target text description information, the target three-dimensional model map can be converted into a cartoon style, thereby obtaining an edited three-dimensional model map.

[0126] For another example, the target three-dimensional model map is a 3D face map, and the target text description information input by the user is "a avatar face". According to the target text description information, the target three-dimensional model map can be converted into the style of a user's avatar, thereby obtaining an edited three-dimensional model map.

[0127] For another example, the target three-dimensional model map is a 3D face map, and the target text description information input by the user is "a tattoo of flower". According to the target text description information, a tattoo of a flower can be added to the target three-dimensional model map, thereby obtaining an edited three-dimensional model map.

[0128] For another example, the target three-dimensional model map is a 3D face map, and the target text description information input by the user is "a tattoo of love". According to the target text description information, a tattoo of a love heart shape can be added to the target three-dimensional model map, thereby obtaining an edited three-dimensional model map.

[0129] In a possible implementation, an editing type corresponding to the target text description information can be determined. The editing type can be global editing or local editing.

[0130] As an example of the implementation, if the user selects the target editing area, it can be determined that the editing type corresponding to the target text description information is local editing; if the user does not select the target editing area, it can be determined that the editing type corresponding to the target text description information is overall editing.

[0131] As another example of the implementation, the editing type corresponding to the target text description information can be determined according to the editing type selected by the user. In this example, an overall editing control and a local editing control can be displayed in the interactive interface; in response to the overall editing control being triggered, it can be determined that the editing type corresponding to the target text description information is overall editing; in response to the local editing control being triggered, it can be determined that the editing type corresponding to the target text description information is local editing.

[0132] As another example of the implementation, the target text description information can be analyzed to determine the editing type corresponding to the target text description information. For example, an editing type prediction model can be pre-trained. The target text description information can be input into the editing type prediction model, and the editing type corresponding to the target text description information can be output by the editing type prediction model.

[0133] In a possible implementation, the target three-dimensional model map is edited according to the target text description information to obtain an edited three-dimensional model map, including: generating an editable latent code corresponding to the target three-dimensional model map; generating an edited latent code according to the target text description information and the editable latent code; and generating an edited three-dimensional model map according to the edited latent code.

[0134] In this implementation, after obtaining the target three-dimensional model map to be edited, an editable latent code corresponding to the target three-dimensional model map can be generated. By editing the editable latent code, the editing of the target three-dimensional model map can be realized.

[0135] In the implementation manner, the editable latent code corresponding to the target three-dimensional model map is generated, the edited latent code is generated according to the target text description information and the editable latent code, and the edited three-dimensional model map is generated according to the edited latent code, so that the three-dimensional model map can be edited as a whole, that is, the entire three-dimensional model map is edited. In addition, the implementation manner provides a flexible and efficient way to edit the three-dimensional model map. The latent code, as a form of internal representation of the three-dimensional model, can capture the inherent rules and characteristics of the three-dimensional model map. By editing the latent code, fine adjustment of the appearance of the three-dimensional model can be realized. This editing manner simplifies the operation process and improves the accuracy and efficiency of editing.

[0136] In a possible implementation manner, the latent code can be a latent code in W+ space. Of course, in other possible implementation manners, the latent code can also be a latent code in W space, a latent code in Z space, and the like, which are not limited herein.

[0137] In a possible implementation manner, the generating of the editable latent code corresponding to the target three-dimensional model map comprises: in response to the target text description information being used for overall editing of the target three-dimensional model map, generating the editable latent code corresponding to the target three-dimensional model map. In this implementation manner, in the case that the target text description information is used for overall editing of the target three-dimensional model map, the editable latent code corresponding to the target three-dimensional model map can be generated, the edited latent code can be generated according to the target text description information and the editable latent code, and the edited three-dimensional model map can be generated according to the edited latent code, so that the three-dimensional model map can be edited as a whole.

[0138] In a possible implementation manner, the generating of the editable latent code corresponding to the target three-dimensional model map comprises: obtaining a latent code to be optimized corresponding to the target three-dimensional model map; generating a three-dimensional model map corresponding to the latent code to be optimized; optimizing the latent code to be optimized according to the three-dimensional model map corresponding to the latent code to be optimized and the target three-dimensional model map until a preset stop optimization condition is met, to obtain the editable latent code corresponding to the target three-dimensional model map.

[0139] In the implementation manner, firstly, the latent code to be optimized can be generated in a random initialization manner.

[0140] In this implementation, the three-dimensional model map corresponding to the to-be-optimized latent code can be generated by generating a model, and the to-be-optimized latent code can be optimized according to the difference information between the three-dimensional model map corresponding to the to-be-optimized latent code and the target three-dimensional model map, to obtain a new to-be-optimized latent code. This step can be repeated until a preset stop optimization condition is met. The preset stop optimization condition can be that the similarity between the three-dimensional model map corresponding to the latest to-be-optimized latent code and the target three-dimensional model map is greater than or equal to a preset similarity, or the number of times of optimizing the to-be-optimized latent code reaches a preset number, and the like.

[0141] In this implementation, the latest to-be-optimized latent code can be used as the editable latent code corresponding to the target three-dimensional model map in response to the preset stop optimization condition being met.

[0142] In this implementation, the to-be-optimized latent code is generated by random initialization, which provides a starting point for finding the editable latent code corresponding to the target three-dimensional model map. Through the subsequent optimization process, the optimized latent code can gradually approach the characteristics of the target three-dimensional model map. By continuously comparing the difference between the generated three-dimensional model map and the target three-dimensional model map, the latent code can be accurately adjusted to optimize the generation result. By optimizing the to-be-optimized latent code according to the difference information, the latent code can gradually approach the internal representation of the target three-dimensional model map. Therefore, the editable latent code generated by this implementation can more accurately represent the target three-dimensional model map.

[0143] In another possible implementation, the generating the editable latent code corresponding to the target three-dimensional model map includes: inputting the target three-dimensional model map into a pre-trained inverse model (Inverse Model), and generating the editable latent code corresponding to the target three-dimensional model map by the inverse model.

[0144] The inverse model can be used to map an input picture to an editable latent code. Moreover, the inverse model can be pre-trained using a large amount of data.

[0145] In this implementation, the target three-dimensional model map can be input into the inverse model, and the editable latent code corresponding to the target three-dimensional model map can be generated by the inverse model.

[0146] In this implementation, the pre-trained inverse model directly generates an editable latent code corresponding to the target three-dimensional model map, which can improve the efficiency and accuracy of generating editable latent codes. Compared with traditional optimization processes, this method can directly obtain a latent code highly consistent with the target three-dimensional model map without multiple iterations and adjustments, greatly saving computing resources and time costs. In addition, the inverse model has learned a large number of model map features and internal laws during the training process, so it can accurately capture the unique attributes of the target three-dimensional model map. This makes the generated editable latent code not only accurately reflect the appearance of the target three-dimensional model map, but also maintain its internal structure and texture information, ensuring the quality and consistency of the generated results.

[0147] In the embodiments of the present disclosure, in the case that the target text description information is used for overall editing of the target three-dimensional model map, the target text description information and the editable latent code can be input into a mapping model, and an edited latent code is output by the mapping model. The edited latent code has the related attributes described in the target text description information.

[0148] In a possible implementation, the generating the edited latent code according to the target text description information and the editable latent code comprises: training a mapping model according to the target text description information and the editable latent code; and in response to the training of the mapping model being completed, inputting the target text description information and the editable latent code into the trained mapping model, and generating the edited latent code by the trained mapping model.

[0149] In this implementation, the mapping model can be trained using the target text description information and the editable latent code corresponding to the target three-dimensional model map. That is, in this implementation, different mapping models can be optimized for different target text description information and target three-dimensional model maps.

[0150] In this implementation, the mapping model is trained by combining the target text description information and the editable latent code, which can deeply integrate natural language description and three-dimensional model internal features. In this way, the mapping model can learn how to accurately reflect the intent and requirements in the text description on the latent code, thereby realizing accurate editing of the three-dimensional model map. Secondly, different mapping models are optimized for different target text description information and target three-dimensional model maps, and this personalized processing method can ensure that each mapping model can be accurately adjusted for specific editing requirements. This greatly improves the flexibility and accuracy of editing, so that users can perform customized editing operations according to their own needs.

[0151] In a possible implementation, the training of the mapping model according to the target text description information and the editable latent code comprises: inputting the target text description information and the editable latent code into the mapping model, generating a predicted edited latent code through the mapping model; obtaining a predicted rendering image corresponding to the predicted edited latent code; and training the mapping model according to the predicted rendering image and the target text description information.

[0152] In this implementation, the predicted edited latent code can represent an output result obtained by the mapping model for the target text description information and the editable latent code in the process of training the mapping model. The predicted rendering image can represent a rendering image corresponding to the predicted edited latent code.

[0153] In this implementation, by obtaining the predicted rendering image corresponding to the predicted edited latent code and comparing it with the target text description information, the prediction accuracy of the mapping model can be evaluated. This evaluation mechanism helps to discover possible problems in the conversion process of the mapping model in a timely manner, thereby guiding the training and optimization of the mapping model in the future. By continuously adjusting the parameters of the mapping model, the prediction accuracy of the mapping model can be gradually improved, so that it can better meet the editing needs of users.

[0154] In a possible implementation, the obtaining of the predicted rendering image corresponding to the predicted edited latent code comprises: generating a predicted edited three-dimensional model texture corresponding to the predicted edited latent code; and pasting the predicted edited three-dimensional model texture on a mesh of a target three-dimensional model corresponding to the target three-dimensional model texture to obtain the predicted rendering image corresponding to the predicted edited latent code.

[0155] In this implementation, the predicted edited three-dimensional model texture can represent a three-dimensional model texture corresponding to the predicted edited latent code.

[0156] In this implementation, the predicted edited latent code generated by the mapping model can be converted into a three-dimensional model texture, i.e., a predicted edited three-dimensional model texture, through a decoding or reconstruction process. By applying this predicted edited three-dimensional model texture to the mesh structure of the target three-dimensional model and rendering the entire three-dimensional model, a predicted rendering image corresponding to the predicted edited latent code can be obtained.

[0157] In the implementation, the predicted edited latent code is corresponded to a predicted edited three-dimensional model map, and the predicted edited three-dimensional model map is pasted on a mesh of a target three-dimensional model corresponding to the target three-dimensional model map, to obtain a predicted rendering image corresponding to the predicted edited latent code, so that the prediction accuracy of the mapping model can be verified, thereby helping to improve the reliability and stability of the editing process.

[0158] In a possible implementation, the training of the mapping model according to the predicted rendering image and the target text description information includes: determining a value of a loss function corresponding to the mapping model according to a similarity between the predicted rendering image and the target text description information; and training the mapping model according to the value of the loss function. In the implementation, the mapping model can be trained in a back propagation manner.

[0159] In another possible implementation, the generating of the edited latent code according to the target text description information and the editable latent code includes: determining a target mapping model; and inputting the target text description information and the editable latent code into the target mapping model, to generate the edited latent code by using the target mapping model.

[0160] In the implementation, a plurality of mapping models can be trained using a large amount of training data and text description information.

[0161] In the implementation, the edited latent code is generated by determining a target mapping model and inputting the target text description information and the editable latent code into the target mapping model, so that the speed of obtaining the edited latent code can be improved.

[0162] In a possible implementation, the determining of the target mapping model includes: determining a keyword of the target text description information; and determining a target mapping model according to the keyword.

[0163] In the implementation, when the mapping model needs to be used, a keyword of the target text description information can be determined first, and then a target mapping model corresponding to the keyword can be searched. In a case where the target mapping model corresponding to the keyword is searched, the target text description information and the editable latent code can be processed by using the target mapping model to obtain the edited latent code. In a case where the target mapping model corresponding to the keyword is not searched, a new mapping model can be trained according to the editable latent code corresponding to the target text description information and the target three-dimensional model map.

[0164] In a possible implementation, the generating the edited three-dimensional model map according to the edited latent code comprises: inputting the edited latent code into a first preset generation model, and generating the edited three-dimensional model map by the first preset generation model.

[0165] In this implementation, the edited latent code can be processed by the first preset generation model to obtain the edited three-dimensional model map.

[0166] As an example of this implementation, the first preset generation model is a generative adversarial model. For example, the first preset generation model can be a styleGAN-based generation model, and the like, without limitation.

[0167] In this example, by using a generative adversarial model as the first preset generation model, inputting the edited latent code into the first preset generation model, and generating the edited three-dimensional model map by the first preset generation model, the properties of the three-dimensional model can be accurately controlled by the first preset generation model, thereby helping to generate a high-quality and realistic three-dimensional model map.

[0168] In a possible implementation, the editing the target three-dimensional model map according to the target text description information to obtain an edited three-dimensional model map comprises: obtaining position information of a target editing region of the target three-dimensional model map; inputting the target three-dimensional model map, the position information of the target editing region, and the target text description information into a second preset generation model, and generating the edited three-dimensional model map by the second preset generation model.

[0169] In this implementation, the target editing region can represent a local region of the target three-dimensional model map that needs to be edited. The target editing region can be determined by user interaction, automatic detection, or preset parameters, and the like. The position information of the target editing region can include coordinates, size, and the like of the target editing region.

[0170] The second preset generation model can be a pre-trained model that can edit the target three-dimensional model map locally according to the input text description information and the position information of the target editing region. By calculation and processing of the second preset generation model, the edited three-dimensional model map can be obtained. The second preset generation model can adjust the texture, color, pattern, and the like of the target editing region according to the input text description information, so as to meet the editing requirements of the user. At the same time, since the position information of the target editing region is provided to the second preset generation model, the second preset generation model can accurately locate the region that needs to be edited, and achieve the effect of local editing.

[0171] As an example of this implementation, the obtaining of the position information of the target editing region of the target three-dimensional model map includes: determining the position information of the target editing region of the target three-dimensional model map according to an area painted by the user for the target three-dimensional model map.

[0172] In this example, the user can manually paint the editing region. After the user finishes painting, the area painted by the user can be determined as the target editing region.

[0173] In one example, the target three-dimensional model map after the target editing region is painted and the target text description information can be input into a second preset generation model, and an edited three-dimensional model map is generated by the second preset generation model, wherein the edited three-dimensional model map edits the target editing region.

[0174] In this example, by allowing the user to directly paint the target three-dimensional model map to specify the editing region, this method provides an intuitive and easy-to-understand user interaction mode. The user does not need to set complex parameters or operate the interface, but only needs to simply paint the area that needs to be edited to quickly locate and determine the position information of the editing region. In addition, the user's painting method makes the range definition of the editing region very flexible. The user can accurately paint an editing region of any shape and size according to needs, thereby meeting different editing needs. This flexibility greatly improves the accuracy and individuality of editing. Compared with other editing methods that require accurate input of position information, the user's painting method reduces the difficulty of operation. The user does not need to have professional graphic editing skills or accurate hand-eye coordination ability, but only needs to paint according to his own intention. This simplified operation method enables more users to easily perform local editing of three-dimensional model maps. Since the user can directly paint the editing region on the model map, the selection and positioning of the editing region can be quickly completed. This greatly shortens the time required for the editing process and improves the editing efficiency. The user can complete complex editing tasks in a shorter time, thereby speeding up the project progress or improving the work efficiency.

[0175] As an example of this implementation, the second preset generation model is a diffusion process-based image generation model. For example, the second preset generation model can use a generation model based on a stable diffusion (stablediffusion) generation model, etc., without limitation here.

[0176] In this example, by adopting the image generation model based on the diffusion process as the second preset generation model, inputting the target three-dimensional model map, the position information of the target editing region, and the target text description information into the second preset generation model, and generating the edited three-dimensional model map through the second preset generation model, it is helpful to generate high-quality and diversified three-dimensional model maps.

[0177] The three-dimensional model map editing method provided by the embodiments of the present disclosure can be applied to the technical field of digital human production, without limitation. By adopting the three-dimensional model map editing method provided by the embodiments of the present disclosure, a 3D face map editing system driven by text description and the like can be realized.

[0178] The three-dimensional model map editing method provided by the embodiments of the present disclosure will be described below through a specific application scenario. In this application scenario, the three-dimensional model map can be a 3D face map. Figure 2 A schematic diagram of the three-dimensional model map editing process provided by the embodiments of the present disclosure is shown. The overall editing process and the local editing process of the 3D face map will be described below.

[0179] I. Overall editing (corresponding to Figure 2 the first editing method in

[0180] Figure 3 A schematic diagram of the overall editing in the three-dimensional model map editing method provided by the embodiments of the present disclosure is shown.

[0181] In combination with Figure 2 and Figure 3 , first, the 3D face map and the target text description information (i.e. Figure 2 and Figure 3 “text description”) can be obtained. Figure 3 Two examples of overall editing of the 3D face map are shown, and the target text description information is “a cartoon face” and “a avatar face”, respectively.

[0182] The 3D face map can be input into a pre-trained inverse model, and the editable latent code corresponding to the 3D face map can be generated through the inverse model.

[0183] The target text description information and the editable latent code can be input into a mapping model, and the edited latent code can be output through the mapping model.

[0184] The edited latent code can be input into a generation model (for example, a generation model based on styleGAN), and the edited 3D face map can be obtained. Figure 3The edited 3D face map corresponding to the target text description information "a cartoon face" is shown, and the edited 3D face map corresponding to the target text description information "a avatar face" is shown.

[0185] II. Local editing (corresponding to Figure 2 editing mode two) in

[0186] Figure 4 A schematic diagram of local editing in the editing method of the three-dimensional model map provided by the embodiments of the present disclosure is shown.

[0187] In combination with Figure 2 and Figure 4 , first, the 3D face map and the target text description information (i.e. Figure 2 and Figure 4 "Text description") in Figure 4 are acquired. Two examples of local editing of the 3D face map are shown, and the target text description information is "a tattoo of flower" and "a tattoo of love", respectively.

[0188] The user can manually paint the editing area. After the user finishes painting, the area painted by the user can be determined as the target editing area.

[0189] Figure 4 The painted 3D face map and the target text description information can be input into the diffusion model, and the edited 3D face map is output by the diffusion model. The edited 3D face map corresponding to the target text information "a tattoo of flower" is shown, and the edited 3D face map corresponding to the target text description information "a tattoo of love" is shown.

[0190] By adopting this application scenario, the editing of the 3D face map can be realized based on the text description information input by the user in the case of fixed grid, so as to realize the editing of the 3D image. Moreover, by adopting this application scenario, the overall editing and the local editing can be realized.

[0191] It can be understood that the above-mentioned various method embodiments mentioned in the present disclosure can be combined with each other to form combined embodiments without violating the principle logic. Limited by the length, the present disclosure will not be repeated. Those skilled in the art can understand that in the above-mentioned method of the specific embodiment, the specific execution order of each step should be determined according to its function and possible internal logic.

[0192] In addition, the disclosure also provides a three-dimensional model map editing device, an electronic device, a computer readable storage medium and a computer program product, which can be used to implement any of the three-dimensional model map editing methods provided by the disclosure. The corresponding technical solutions and technical effects can be referred to the corresponding description in the method part, and will not be repeated here.

[0193] Figure 5 A block diagram of the three-dimensional model map editing device provided by the embodiments of the disclosure is shown. As shown in Figure 5 The three-dimensional model map editing device comprises:

[0194] A first obtaining module 51 is configured to obtain a target three-dimensional model map to be edited.

[0195] A second obtaining module 52 is configured to obtain target text description information used for editing the target three-dimensional model map.

[0196] An editing module 53 is configured to edit the target three-dimensional model map according to the target text description information to obtain an edited three-dimensional model map.

[0197] In a possible implementation, the editing module 53 is configured to:

[0198] generate an editable latent code corresponding to the target three-dimensional model map;

[0199] generate an edited latent code according to the target text description information and the editable latent code;

[0200] generate an edited three-dimensional model map according to the edited latent code.

[0201] In a possible implementation, the editing module 53 is configured to:

[0202] generate an editable latent code corresponding to the target three-dimensional model map in response to the target text description information being used for overall editing of the target three-dimensional model map.

[0203] In a possible implementation, the editing module 53 is configured to:

[0204] train a mapping model according to the target text description information and the editable latent code;

[0205] input the target text description information and the editable latent code into the trained mapping model to generate an edited latent code through the trained mapping model in response to the training of the mapping model being completed.

[0206] In a possible implementation, the editing module 53 is configured to:

[0207] inputting the target text description information and the editable latent code into the mapping model, and generating a predicted edited latent code through the mapping model;

[0208] obtaining a predicted rendering image corresponding to the predicted edited latent code;

[0209] training the mapping model according to the predicted rendering image and the target text description information.

[0210] In a possible implementation, the editing module 53 is configured to:

[0211] generate a predicted edited three-dimensional model texture corresponding to the predicted edited latent code;

[0212] paste the predicted edited three-dimensional model texture on a mesh of a target three-dimensional model corresponding to the target three-dimensional model texture, to obtain a predicted rendering image corresponding to the predicted edited latent code.

[0213] In a possible implementation, the editing module 53 is configured to:

[0214] determine a value of a loss function corresponding to the mapping model according to a similarity between the predicted rendering image and the target text description information;

[0215] train the mapping model according to the value of the loss function.

[0216] In a possible implementation, the editing module 53 is configured to:

[0217] determine a target mapping model;

[0218] input the target text description information and the editable latent code into the target mapping model, and generate an edited latent code through the target mapping model.

[0219] In a possible implementation, the editing module 53 is configured to:

[0220] determine a keyword of the target text description information;

[0221] determine a target mapping model according to the keyword.

[0222] In a possible implementation, the editing module 53 is configured to:

[0223] obtain a latent code to be optimized corresponding to the target three-dimensional model texture;

[0224] generate a three-dimensional model texture corresponding to the latent code to be optimized;

[0225] According to the three-dimensional model map corresponding to the potential code to be optimized and the target three-dimensional model map, the potential code to be optimized is optimized until a preset stop optimization condition is met, to obtain an editable potential code corresponding to the target three-dimensional model map.

[0226] In a possible implementation, the editing module 53 is configured to:

[0227] input the target three-dimensional model map into a pre-trained inverse model, and generate, by the inverse model, an editable potential code corresponding to the target three-dimensional model map.

[0228] In a possible implementation, the editing module 53 is configured to:

[0229] input the edited potential code into a first preset generation model, and generate, by the first preset generation model, an edited three-dimensional model map.

[0230] In a possible implementation, the first preset generation model is a generative adversarial model.

[0231] In a possible implementation, the editing module 53 is configured to:

[0232] obtain position information of a target editing region of the target three-dimensional model map;

[0233] input the target three-dimensional model map, the position information of the target editing region, and the target text description information into a second preset generation model, and generate, by the second preset generation model, an edited three-dimensional model map.

[0234] In a possible implementation, the editing module 53 is configured to:

[0235] determine, according to a region painted by a user for the target three-dimensional model map, position information of a target editing region of the target three-dimensional model map.

[0236] In a possible implementation, the second preset generation model is a diffusion process-based image generation model.

[0237] In a possible implementation, the target three-dimensional model map is a three-dimensional face map.

[0238] In some embodiments, the apparatus provided by the embodiments of the present disclosure has functions or includes modules that can be used to perform the methods described in the above method embodiments, and the specific implementation and technical effects can refer to the description of the above method embodiments. For brevity, they will not be repeated here.

[0239] The embodiment of the present disclosure further provides a computer readable storage medium, which stores computer program instructions, and the computer program instructions are executed by a processor to implement the method.

[0240] The embodiment of the present disclosure further provides a computer program, which comprises computer readable codes, and when the computer readable codes are run in an electronic device, a processor in the electronic device executes the method.

[0241] The embodiment of the present disclosure further provides a computer program product, which comprises computer readable codes, or a non-volatile computer readable storage medium carrying the computer readable codes, and when the computer readable codes are run in an electronic device, a processor in the electronic device executes the method.

[0242] The embodiment of the present disclosure further provides an electronic device, which comprises one or more processors, and a memory for storing executable instructions, wherein the one or more processors are configured to invoke the executable instructions stored in the memory to execute the method.

[0243] The electronic device can be provided as a terminal, a server or other forms of devices.

[0244] Figure 6 A block diagram of an electronic device 1900 provided by the embodiment of the present disclosure is shown. For example, the electronic device 1900 can be provided as a server or a terminal. Referring to Figure 6 , the electronic device 1900 comprises a processing component 1922, which further comprises one or more processors, and a memory resource represented by a memory 1932 for storing instructions executable by the processing component 1922, such as an application program. The application program stored in the memory 1932 can comprise one or more than one module each corresponding to a set of instructions. In addition, the processing component 1922 is configured to execute the instructions to execute the method.

[0245] The electronic device 1900 can further comprise a power supply component 1926 configured to perform power management of the electronic device 1900, a wired or wireless network interface 1950 configured to connect the electronic device 1900 to a network, and an input / output interface 1958 (I / O interface). The electronic device 1900 can operate based on an operating system stored in the memory 1932, such as a Microsoft server operating system (Windows Server TM ), a graphical user interface operating system (MacOS X TM ) developed by Apple, a multi-user multi-process computer operating system (Unix TM), a free and open-source Unix-like operating system (Linux TM ), an open-source Unix-like operating system (FreeBSD TM ), or the like.

[0246] In an example embodiment, there is also provided a non-transitory computer- readable storage medium, such as the memory 1932 including computer program instructions, which can be executed by the processing component 1922 of the electronic device 1900 to implement the above-described method.

[0247] The present disclosure can be a system, a method, and / or a computer program product. The computer program product can include a computer readable storage medium (or media) having computer readable program instructions thereon for causing a processor to carry out aspects of the present disclosure.

[0248] The computer readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. The computer readable storage medium can be, for example, but is not limited to, an electronic storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. A non-exhaustive list of more specific examples of the computer readable storage medium includes the following: a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disk (DVD), a memory stick, a floppy disk, a mechanically encoded device such as punch-cards or punched tape, a magnetically encoded device such as magnetic strip cards, an optically encoded device such as a compact disc (CD) or DVD, and / or any suitable combination of the foregoing. A computer readable storage medium, as used herein, is not to be construed as being transitory signals per se, such as radio waves or other freely propagating electromagnetic waves, electromagnetic waves propagating through a waveguide or other transmission media (e.g., light pulses passing through a fiber-optic cable), or electrical signals transmitted through a wire.

[0249] Computer readable program instructions described herein can be downloaded to respective computing / processing devices from a computer readable storage medium or to an external computer or external storage device via a network, for example, the Internet, a local area network, a wide area network and / or a wireless network. The network can comprise copper transmission cables, optical transmission fibers, wireless transmission, routers, firewalls, switches, gateway computers and / or edge servers. A network adapter card or network interface in each computing / processing device receives computer readable program instructions from the network and forwards the computer readable program instructions for storage in a computer readable storage medium within the respective computing / processing device.

[0250] Computer readable program instructions for carrying out operations of the present disclosure can be assembler instructions, instruction-set-architecture (ISA) instructions, machine instructions, machine dependent instructions, microcode, firmware instructions, state-setting data, or either source code or object code written in any combination of one or more programming languages, including an object oriented programming language such as Smalltalk, C++ or the like and conventional procedural programming languages, such as the "C" programming language or similar programming languages. The computer readable program instructions can execute entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on a remote computer or entirely on the remote computer or server. In the latter scenario, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or the connection can be made to an external computer (for example, through the Internet using an Internet Service Provider). In some embodiments, electronic circuitry including, for example, programmable logic circuitry, field-programmable gate array (FPGA), or programmable logic array (PLA) can execute the computer readable program instructions by utilizing state information of the computer readable program instructions to personalize the electronic circuitry, in order to perform aspects of the present disclosure.

[0251] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0252] These computer readable program instructions can be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks. These computer readable program instructions can also be stored in a computer readable storage medium that can include random access memory (RAM), read only memory (ROM), electrically erasable programmable read only memory (EEPROM), flash memory or other data storage device. When the computer readable program instructions are loaded into the computer and other programmable data processing apparatus, a series of operational steps are implemented that provide processes such that the computer or other programmable apparatus provide processes for implementing the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0253] The computer readable program instructions can also be loaded onto a computer, other programmable data processing apparatus, or other device to cause a series of operational steps to be performed on the computer, other programmable data processing apparatus or other device to produce a computer implemented process such that the instructions which execute on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in the flowchart and / or block diagram block or blocks.

[0254] The flow diagrams and the block diagrams in the drawings are presented to illustrate the architecture, functionality, and operation of possible implementations of systems, methods and computer program products according to various embodiments of the present disclosure. In this regard, each block in the flow diagrams and the block diagrams can represent a module, segment, or portion of instructions, which comprises one or more executable instructions for implementing the specified logic functions. In some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures. For example, two blocks shown in succession may, in fact, be executed substantially concurrently, or the blocks may sometimes be executed in the reverse order, depending upon the functionality involved. It will also be noted that each block of the block diagrams and / or flow diagrams, and combinations thereof, can be implemented by dedicated hardware-based systems which perform the specified functions or acts, or combinations of hardware and software.

[0255] The computer program product can be embodied by a hardware, software or a combination thereof. In an optional embodiment, the computer program product is embodied by a computer storage medium, and in another optional embodiment, the computer program product is embodied by a software product, such as a software development kit (SDK) or the like.

[0256] The above description of the various embodiments is intended to be illustrative in all aspects, rather than restrictive. The same or similar features can be employed in other embodiments without departing from the spirit or scope of the disclosure.

[0257] If the technical solutions of the embodiments of the present disclosure involve personal information, the product applying the technical solutions of the embodiments of the present disclosure has been explicitly informed of the personal information processing rules before processing the personal information and has obtained the personal independent consent. If the technical solutions of the embodiments of the present disclosure involve sensitive personal information, the product applying the technical solutions of the embodiments of the present disclosure has obtained the personal independent consent before processing the sensitive personal information and at the same time meets the requirement of "explicit consent". For example, at the personal information collection device such as a camera, a clear and prominent mark is set to inform that the personal information collection range has been entered and the personal information will be collected. If the person voluntarily enters the collection range, it is regarded as the consent to collect the personal information. Or on the device for processing personal information, the personal information processing rules are informed by using obvious marks / information, and the personal authorization is obtained by means of pop-up information or asking the person to upload the personal information by himself / herself. The personal information processing rules can include the personal information processor, the processing purpose of personal information, the processing method, the type of processed personal information and the like.

[0258] The above has described the embodiments of the present disclosure, and the above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and changes are obvious to those skilled in the art without departing from the scope and spirit of the described embodiments. The selection of the terms used herein is intended to best explain the principles of the embodiments, practical application or improvement of the technology in the market, or to enable other ordinary skilled persons in the art to understand the embodiments disclosed herein.

Claims

1. A method for editing a three-dimensional model map, characterized in that: include: Get the target 3D model texture to be edited; Obtaining target text description information for editing the target three-dimensional model texture; The target three-dimensional model texture is edited according to the target text description information to obtain an edited three-dimensional model texture.

2. The method according to claim 1, characterized in that The step of editing the target three-dimensional model texture according to the target text description information to obtain the edited three-dimensional model texture includes: Generating an editable latent code corresponding to the target three-dimensional model map; generating an edited latent code according to the target text description information and the editable latent code; An edited three-dimensional model map is generated according to the edited latent code.

3. The method according to claim 2, characterized in that Generating the editable latent code corresponding to the target three-dimensional model map includes: In response to the target text description information being used to edit the target three-dimensional model texture as a whole, an editable latent code corresponding to the target three-dimensional model texture is generated.

4. The method according to claim 2 or 3, characterized in that Generating an edited latent code according to the target text description information and the editable latent code includes: Training a mapping model based on the target text description information and the editable latent code; In response to the completion of the training of the mapping model, the target text description information and the editable latent code are input into the trained mapping model, and the edited latent code is generated by the trained mapping model.

5. The method according to claim 4, characterized in that The step of training a mapping model according to the target text description information and the editable potential code includes: Inputting the target text description information and the editable latent code into a mapping model, and generating a predicted edited latent code through the mapping model; Obtaining a predicted rendered image corresponding to the predicted edited latent code; The mapping model is trained according to the predicted rendered image and the target text description information.

6. The method according to claim 5, characterized in that The obtaining of the predicted rendered image corresponding to the predicted edited latent code includes: generating a predicted edited three-dimensional model map corresponding to the predicted edited latent code; The predicted edited three-dimensional model texture is pasted on a grid of a target three-dimensional model corresponding to the target three-dimensional model texture to obtain a predicted rendered image corresponding to the predicted edited latent code.

7. The method according to claim 5, characterized in that The step of training the mapping model according to the predicted rendered image and the target text description information includes: Determining a value of a loss function corresponding to the mapping model according to the similarity between the predicted rendered image and the target text description information; The mapping model is trained according to the value of the loss function.

8. The method according to claim 2, characterized in that Generating an edited latent code according to the target text description information and the editable latent code includes: Determine the target mapping model; The target text description information and the editable latent code are input into the target mapping model, and the edited latent code is generated by the target mapping model.

9. The method according to claim 8, characterized in that Determining the target mapping model includes: Determining keywords for the target text description information; A target mapping model is determined according to the keyword.

10. The method according to claim 2 or 3, characterized in that Generating the editable latent code corresponding to the target three-dimensional model map includes: Obtaining a potential code to be optimized corresponding to the target three-dimensional model map; Generating a three-dimensional model map corresponding to the potential code to be optimized; According to the three-dimensional model map corresponding to the potential code to be optimized and the target three-dimensional model map, the potential code to be optimized is optimized until a preset optimization stop condition is met, thereby obtaining an editable potential code corresponding to the target three-dimensional model map.

11. The method according to claim 2 or 3, characterized in that Generating the editable latent code corresponding to the target three-dimensional model map includes: The target three-dimensional model map is input into a pre-trained inverse model, and an editable latent code corresponding to the target three-dimensional model map is generated by the inverse model.

12. The method according to claim 2 or 3, characterized in that Generating an edited three-dimensional model map according to the edited latent code includes: The edited latent code is input into a first preset generation model, and the edited three-dimensional model map is generated by the first preset generation model.

13. The method according to claim 12, characterized in that The first preset generative model is a generative adversarial model.

14. The method according to claim 1, wherein The step of editing the target three-dimensional model texture according to the target text description information to obtain the edited three-dimensional model texture includes: Obtaining position information of a target editing area of ​​the target three-dimensional model map; The target three-dimensional model texture, the position information of the target editing area and the target text description information are input into a second preset generation model, and the edited three-dimensional model texture is generated by the second preset generation model.

15. The method according to claim 14, characterized in that The obtaining of the position information of the target editing area of ​​the target three-dimensional model map includes: According to the area painted by the user on the target three-dimensional model texture, position information of the target editing area of ​​the target three-dimensional model texture is determined.

16. The method according to claim 14 or 15, characterized in that The second preset generation model is an image generation model based on a diffusion process.

17. The method according to claim 1, wherein The target three-dimensional model map is a three-dimensional face map.

18. A device for editing a three-dimensional model map, characterized in that: include: A first acquisition module is used to acquire a target three-dimensional model texture to be edited; A second acquisition module is used to acquire target text description information for editing the target three-dimensional model map; The editing module is used to edit the target three-dimensional model texture according to the target text description information to obtain the edited three-dimensional model texture.

19. An electronic device, characterized in that: include: one or more processors; a memory for storing executable instructions; The one or more processors are configured to call the executable instructions stored in the memory to execute the method according to any one of claims 1 to 17.

20. A computer-readable storage medium having computer program instructions stored thereon, characterized in that: When the computer program instructions are executed by a processor, the method according to any one of claims 1 to 17 is implemented.

21. A computer program product comprising computer-readable code, or a non-volatile computer-readable storage medium carrying computer-readable code, characterized in that: When the computer-readable code is executed in an electronic device, a processor in the electronic device executes the method according to any one of claims 1 to 17.

Images