Methods, apparatus, devices and storage media for constructing virtual avatar texture maps

Abstract

This application provides a method, apparatus, device, and readable storage medium for constructing a virtual character texture map. The method includes: acquiring a three-dimensional target image; performing face reconstruction based on the three-dimensional target image to obtain a first realistic mesh model; obtaining a cartoon topological structure model based on a cartoon topological structure corresponding to the three-dimensional target image; topologically transferring the first realistic mesh model to the cartoon topological structure model to obtain a second realistic mesh model and a second cartoon topological structure model, such that the number and position of vertices in the second realistic mesh model and the second cartoon topological structure model are the same; and using the mapping relationship of texture coordinates between the second realistic mesh model and the cartoon topological structure model, mapping the texture coordinates of the first realistic mesh model to the second cartoon topological structure model to construct a target texture map in the second cartoon topological structure model. This aims to solve the problem of a large discrepancy between the cartoon topological structure model and the realistic character.

Description

Technical Field

[0001] This application relates to the field of image processing technology, and in particular to methods, apparatus, devices and readable storage media for constructing virtual image texture maps. Background Technology

[0002] The concept of the metaverse is currently quite popular, and one of its key functions is user character modeling. Character modeling takes several different forms: 100% high-precision modeling of a real user's image; simulated real-life images; and cartoon images. High-precision modeling involves accurately replicating the texture features of a real user's image, extracting the user's texture and embedding it into the model. This method requires sophisticated equipment. Simulated real-life images extract texture features from the user's real features and create a model based on these features. The texture features of this model are those extracted from the user's real features and do not undergo cartoon or other special effects processing. Therefore, this method suffers from a similar problem to high-precision modeling: because the created image is too realistic, it can lead to user dissatisfaction when some users are not satisfied with certain features of their own image. Cartoon characters are created by building a cartoon topology model. Although this method is based on the real-life image input by the user, the texture map of the cartoon topology model is very different from the real-life texture map because it has been processed by the cartoon topology. The cartoon character is significantly distorted from the real person. Therefore, the texture features on the model need to be manually adjusted by the user to shape the face parameters, which is a rather cumbersome process. Summary of the Invention

[0003] To overcome the problems existing in related technologies, this application provides a method, apparatus, device and readable storage medium for constructing virtual image texture maps.

[0004] According to a first aspect of the embodiments of this application, a method for constructing a virtual image texture map is provided, the method comprising:

[0005] A three-dimensional target image is acquired, and a face is reconstructed based on the three-dimensional target image to obtain a first realistic mesh model. A cartoon topological structure model is obtained based on the cartoon topological structure corresponding to the three-dimensional target image. The realistic mesh model and the cartoon topological structure model contain different numbers of vertices or have vertices in different positions.

[0006] The first realistic mesh model is topologically transferred to the cartoon topological structure model to obtain the second realistic mesh model, so that the number of vertices and the vertex positions of the second realistic mesh model and the cartoon topological structure model are the same;

[0007] By utilizing the mapping relationship between the texture coordinates of the second realistic mesh model and the cartoon topology model, the texture coordinates of the first realistic mesh model are mapped to the cartoon topology model to construct a target texture map in the cartoon topology model.

[0008] According to a second aspect of the embodiments of this application, an apparatus for constructing a virtual image texture map is provided, comprising:

[0009] The modeling module is used to perform face reconstruction based on the three-dimensional target image to obtain a first realistic mesh model, and to obtain a cartoon topological structure model based on the cartoon topological structure corresponding to the three-dimensional target image; the realistic mesh model and the cartoon topological structure model contain different numbers of vertices or have vertices in different positions;

[0010] The topology migration module is used to migrate the first realistic mesh model to the cartoon topology model to obtain a second realistic mesh model, so that the number of vertices and the vertex positions of the second realistic mesh model and the cartoon topology model are the same.

[0011] The texture transfer module is used to map the texture coordinates of the first real-life mesh model to the cartoon topology model based on the mapping relationship between the texture coordinates of the second real-life mesh model and the cartoon topology model, so as to construct a target texture map in the cartoon topology model.

[0012] According to a third aspect of the embodiments of this application, an apparatus is provided, comprising:

[0013] processor;

[0014] Memory used to store processor-executable instructions;

[0015] The processor is configured to perform the operation described in any of the methods described in the first aspect above. According to a fourth aspect of the present application, a readable storage medium is provided, comprising:

[0016] When executed by the processor, this instruction performs the operation as described in any of the methods described in the first aspect above.

[0017] The technical solutions provided by the embodiments of this application may include the following beneficial effects:

[0018] In this embodiment, a cartoon topological model and a first realistic mesh model are first established based on the 3D target image. The cartoon topological model is obtained through a cartoon topological model. Since the first realistic mesh model is obtained by reconstructing a face from the 3D target image, it possesses realistic facial features. However, the cartoon topological model obtained from the cartoon topological model differs significantly from realistic facial features; the number and position of vertices in the two models are not identical. This solution then topologically transfers the first realistic mesh model to the cartoon topological model to transfer some realistic facial features to the cartoon topological model, giving it some features similar to those of a real person. Next, using the mapping relationship between the texture coordinates of the two models, the texture coordinates of the first realistic mesh model are mapped to the cartoon topological model, and a texture map is constructed based on these coordinates. This achieves the effect of the cartoon topological model possessing realistic texture features, while the process of transferring realistic texture features to the cartoon topological model also enhances the real-life texture.

[0019] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0020] The accompanying drawings, which are incorporated in and form part of this application, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.

[0021] Figure 1A This is a schematic diagram of a simulated real person in one embodiment of this application;

[0022] Figure 1B This is a schematic diagram of a cartoon character in one embodiment of this application;

[0023] Figure 2 This is a flowchart of a method for constructing a texture map of a virtual avatar in one embodiment of this application;

[0024] Figure 3 This is a schematic diagram of a real-life mesh model and its texture map in one embodiment of this application;

[0025] Figure 4 This is a flowchart of the process of reconstructing mesh and texture coordinate mapping in one embodiment of this application;

[0026] Figure 5 This is a framework diagram of an apparatus for constructing a texture map of a virtual image according to an embodiment of this application;

[0027] Figure 6 This is a flowchart of the application in a live streaming network scenario;

[0028] Figure 7This is a schematic diagram of an electronic device according to an embodiment of this application. Detailed Implementation

[0029] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0030] The terminology used in this application is for the purpose of describing particular embodiments only and is not intended to be limiting of the application. The singular forms “a,” “the,” and “the” used in this application and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any or all possible combinations of one or more of the associated listed items.

[0031] It should be understood that although the terms first, second, third, etc., may be used in this application to describe various information, such information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the word "if" as used herein may be interpreted as "when," "when," or "in response to determination."

[0032] When creating virtual avatars, existing technologies typically choose one of three modeling methods to model the target: 1. High-precision modeling; 2. Realistic modeling; 3. Cartoon topological modeling. While all three methods can create a 3D model of the target's face, the inventors have found that these methods have some drawbacks in application. For example, high-precision modeling of the target, or using realistic modeling to create a model that resembles a real person, largely retains the texture features of a real person. Figure 1A As shown; therefore, when some users are dissatisfied with certain features of their own appearance, using the above two methods for character modeling will lead to user dissatisfaction with the modeling results. Furthermore, using a cartoon topology model will completely alter the user's own texture features, such as... Figure 1B As shown, if you want to create a cartoon topology model that is very similar to the user's own image, you need to make a lot of adjustments to the texture feature parameters of the cartoon topology model.

[0033] To address the aforementioned problems, embodiments of this application provide a method for constructing virtual avatar texture maps, referring to... Figure 2 , Figure 2 These are some of the steps in this embodiment.

[0034] In step S201, a three-dimensional target image is acquired;

[0035] In step S202, face reconstruction is performed based on the 3D target image to obtain a first realistic mesh model, and a cartoon topology model is obtained based on the cartoon topology corresponding to the 3D target image; the realistic mesh model and the cartoon topology model contain different numbers of vertices or have vertices in different positions;

[0036] In step S203, the topology of the first realistic mesh model is transferred to the cartoon topology model to obtain the second realistic mesh model, so that the number of vertices and the vertex positions of the second realistic mesh model and the cartoon topology model are the same.

[0037] In step S204, the texture coordinates of the first real-life mesh model are mapped to the cartoon topology model using the mapping relationship between the texture coordinates of the second real-life mesh model and the cartoon topology model, so as to construct the target texture map in the cartoon topology model.

[0038] It should be noted that PCA (Principal Component Analysis) is a feature reduction algorithm commonly used in fields such as face recognition to reduce the dimensionality of feature points in face models. Since cartoon topology lacks a PCA basis, it lacks feature points suitable for dimensionality reduction. Therefore, the feature points of the cartoon topology model will differ from those of the first realistic mesh model with a PCA basis. Consequently, the number of vertices in the first realistic mesh model and the cartoon topology model may differ, and the vertex positions may not all be identical.

[0039] As an example, when performing topology transfer on the first realistic mesh model, the alignment parameters between the first realistic mesh model and the cartoon topology model can be obtained first. This alignment parameter, known as the first alignment parameter, refers to the spatial relationship parameters between the vertices of the first realistic mesh model and the vertices of the cartoon topology model during the topology transfer process. Based on the obtained first alignment parameter, the number and position of vertices in the first realistic mesh model are aligned with the vertices in the cartoon topology model to obtain the second realistic mesh model. After obtaining the second realistic mesh model, the mapping relationship of texture coordinates between the second realistic mesh model and the cartoon topology model is determined based on the first alignment parameter. In one example, the specific approach could be as follows: Based on the first alignment parameter, determine the texture coordinates corresponding to the texture coordinates of the cartoon topology model within the second realistic mesh model; determine the mapping relationship between the texture coordinates of the second realistic mesh model and the texture coordinates of the cartoon topology model. Alternatively, based on the first alignment parameter, find the correspondence between the vertices of the cartoon topology model and the second realistic mesh model. Since texture coordinates are composed of model vertices, a texture map of the cartoon topology model can be constructed on the second realistic mesh model based on the correspondence between model vertices. This texture map can then be compared with the texture map of the second realistic mesh model to obtain the mapping relationship of the texture coordinates. It should be noted that the method for determining the mapping relationship of texture coordinates using the first alignment parameter is not unique and will not be elaborated upon here.

[0040] In another embodiment, when constructing the target texture map in the cartoon topology model, the texture map of the first realistic mesh model can be obtained first through texture sampling, and then the texture map of the first realistic mesh model can be mapped onto the cartoon topology model through the mapping relationship of texture coordinates, such as... Figure 3 As shown, 301 and 302 represent the first realistic texture feature and the first realistic mesh model, respectively. 303 is the texture map of the mesh model, and the coordinate points contained in 302 are the texture coordinates. After mapping these texture coordinates onto the cartoon topological model according to the mapping relationship, the cartoon topological structure model can be made to have realistic texture features. It should be noted that the texture sampling method can be texture sampling based on the pixel points where the model vertices are located, or texture sampling based on the pixel points within the mesh of the mesh model. Either sampling method can achieve the final purpose of this application. In this embodiment, texture sampling based on the pixel points within the mesh of the first realistic mesh model is used.

[0041] Optionally, in step S204, before constructing the target texture map, a mesh can be reconstructed based on the mapped texture coordinates, and the target texture map can be constructed based on the pixels within the mesh. In one embodiment, the reconstructed mesh can be a Delaunay triangle (i.e., a Delaunay triangulation, a set of connected but non-overlapping triangles, where the circumcircles of these triangles do not contain any other points of the region, meaning that any triangle in the triangulation generated using this method cannot contain the vertices of other triangles), and the target texture map can be constructed based on the pixels within the Delaunay triangle mesh; alternatively, other mesh shapes can be constructed, such as equilateral triangles, quadrilaterals, etc., without particular limitation. In this embodiment, some steps of constructing the target texture map based on the Delaunay triangle are as follows: Figure 4 As shown,

[0042] In step S401, the texture coordinates of the first realistic mesh model and the cartoon topological structure model are obtained;

[0043] In step S402, the Drone triangle is constructed based on the texture coordinates of the cartoon topology model;

[0044] In step S403, since the texture coordinates are composed of model vertices, the vertex coordinates of the corresponding cartoon topology model can be determined by the texture coordinates of the cartoon topology model, thereby converting the De Röne triangle composed of the texture coordinates of the cartoon topology model into the De Röne triangle composed of the vertices of the cartoon topology model.

[0045] In step S404, since the second realistic mesh model is obtained by topological transfer of the first realistic mesh model and the cartoon topology model, the vertices of the cartoon topology model are mapped to the second realistic mesh model in combination with the first alignment parameter. Thus, the De Röne triangle formed by the vertices of the cartoon topology model is transformed into the De Röne triangle formed by the vertices of the second realistic mesh model. Since the texture coordinates are formed by the model vertices, it is equivalent to obtaining the De Röne triangle formed by the texture coordinates of the second realistic mesh model.

[0046] In step S405, all De Röhne triangles are traversed to obtain the mapping relationship between the De Röhne triangles in the second realistic mesh model and the De Röhne triangles in the cartoon topology mesh model, and then the mapping relationship between their texture coordinates is obtained.

[0047] In step S406, the De Röne triangle constructed based on the texture coordinates of the first real-life mesh model is warped, and its triangular facet region is mapped to the De Röne triangle of the cartoon topological model according to the mapping relationship. The above warping process refers to the processing of image affine transformation. Its principle is: for a coordinate point (x, y) in a two-dimensional coordinate system, a 2x2 matrix can be used to adjust the values ​​of x and y. By adjusting x and y, a linear transformation (rotation, scaling) of the two-dimensional shape can be achieved. Therefore, the entire transformation process is the process of adjusting (x, y). Affine transformation refers to the process of performing a linear transformation and a translation in a vector space to transform to another vector space. The linear transformation is to multiply (x, y) by a matrix, which can be a matrix containing the image transformation feature dimension; the translation transformation is to add (x, y) to a vector, which can be a vector containing the translation position transformation information. In affine transformations, the target image retains its straightness and parallelism. Straightness means that a straight line remains a straight line after an affine transformation, and an arc remains an arc. Parallelism means that the relative positions of straight lines remain unchanged, parallel lines remain parallel, and the order of coordinate points on a straight line does not change, although the angles between vectors may change. In other words, an affine transformation is a linear transformation from two-dimensional coordinates (x, y) to two-dimensional coordinates (u, v), and its mathematical expression is as follows:

[0048]

[0049] Where a1a2 are the transformation coefficients of the image with respect to x, such as the displacement coefficients on the x-axis; b1b2 are the transformation coefficients of the image with respect to y, such as the displacement coefficients on the y-axis; and c1c2 are constants.

[0050] When an image needs to be scaled and rotated, it needs to be done using matrices. To avoid changing the image size after the transformation, the 2x3 matrix constructed from the above formula is converted into a homogeneous matrix, resulting in the homogeneous matrix representation of the image affine transformation:

[0051]

[0052] In step S407, after processing all the Drone triangles, the target texture map is completed.

[0053] It should be noted that, Figure 4The real-life texture coordinate space referred to in A includes the coordinate spaces of the first real-life mesh model and the second real-life mesh model. The vertices corresponding to the texture coordinates of the first real-life mesh model are the same as those corresponding to the texture coordinates of the second real-life mesh model, but the positions of these vertices are not all the same. After the first real-life mesh model is transformed into the second real-life mesh model through topological transfer, the texture coordinates based on the first real-life mesh model will not change as the topological transfer proceeds.

[0054] In another embodiment, an apparatus for constructing a texture map of a virtual avatar is provided, such as... Figure 5 As shown:

[0055] The modeling module 501 is used to perform face reconstruction based on the three-dimensional target image to obtain a first realistic mesh model, and to obtain a cartoon topological structure model based on the cartoon topological structure corresponding to the three-dimensional target image; the realistic mesh model and the cartoon topological structure model contain different numbers of vertices or have vertices in different positions; it should be noted that, since the first realistic mesh model and the cartoon topological structure model are modeled in different ways, and the cartoon topological model does not have a PCA basis, their model vertices are inconsistent, and the positions of the vertices are not all the same;

[0056] The topology transfer module 502 is used to transfer the topology of the first real-life mesh model established above to the cartoon topology model to obtain the second real-life mesh model, so that the number of vertices and the position of vertices of the second real-life mesh model after topology transfer are exactly the same as those of the cartoon topology model.

[0057] The texture transfer module 503 is used to map the texture coordinates of the first real-life mesh model to the cartoon topology model based on the mapping relationship between the texture coordinates of the second real-life mesh model and the cartoon topology model, so as to construct the target texture map in the cartoon topology model.

[0058] It should be noted that when the topology transfer module 502 performs topology transfer on the first realistic mesh model, it can first obtain the alignment parameters between the first realistic mesh model and the cartoon topology model. Based on the obtained first alignment parameters, the number and position of vertices in the first realistic mesh model are aligned with the vertices in the cartoon topology model to obtain the second realistic mesh model. After obtaining the second realistic mesh model, the mapping relationship between the texture coordinates of the second realistic mesh model and the cartoon topology model is determined based on the first alignment parameters. Specifically, the texture coordinates of the second realistic mesh model are determined based on the texture coordinates of the cartoon topology model and the first alignment parameters. The mapping relationship is then determined based on the texture coordinates of the second realistic mesh model and the texture coordinates of the cartoon topology model.

[0059] Before the texture transfer module 503 performs texture transfer, the texture map of the first real-life mesh model can be obtained through texture sampling. Then, the texture map of the first real-life mesh model can be mapped onto the cartoon topology model through the mapping relationship of texture coordinates. Optionally, before constructing the texture map, the mesh can be re-established based on the texture coordinates of the cartoon topology model. In one embodiment, the re-established mesh is the De Röne triangle. The specific process is as described above.

[0060] Optionally, before the texture transfer module 503 constructs the target texture map, a mesh can be reconstructed based on the mapped texture coordinates, and the target texture map can be constructed based on the pixels within the mesh. In one embodiment, the reconstructed mesh is a de Röhne triangle, and the target texture map is constructed based on the pixels within the de Röhne triangle mesh.

[0061] One application scenario for the texture map scheme for constructing virtual avatars provided in this application is in a live streaming scenario. The network environment includes a broadcaster client, several viewer clients, and a live streaming server. The broadcaster client distributes the live video stream to the viewer clients through the live streaming server. The functionality of the texture map scheme for constructing virtual avatars in this application can be configured on the broadcaster client or on the live streaming server. Figure 6 The following explanation uses the method of executing this application on a broadcaster client as an example.

[0062] In step S601, the broadcaster client captures the broadcaster's image through the camera;

[0063] In step S602, the broadcaster client obtains a first real-person mesh model by reconstructing the face of the broadcaster image; and performs cartoon topology modeling based on the broadcaster image to obtain a cartoon topology model.

[0064] In step S603, the broadcaster client migrates the topology of the first real-person mesh model to the cartoon topology model to obtain the second real-person mesh model. After this topology migration process, the first real-person mesh model, which was not aligned with the number and position of the vertices of the cartoon topology model, is transformed into the second real-person mesh model that is aligned with the number and position of the vertices of the cartoon topology model.

[0065] In step S604, the broadcaster client can obtain the mapping relationship between the texture coordinates of the second real-person mesh model and the cartoon topology model based on the first alignment parameter during the topology migration. Based on this mapping relationship, the texture coordinates of the first real-person mesh model containing the facial texture features of the photo input by the user are mapped onto the cartoon topology model.

[0066] In step S605, the broadcaster client reconstructs the mesh on the cartoon topology model based on the texture coordinates of the mapped first real-person mesh model, and constructs a texture map based on the pixels in the reconstructed mesh.

[0067] In step S606, the cartoon topology model with the constructed texture map is output. The output cartoon topology model is a model that can create a cartoon character with a certain similarity to the anchor's image.

[0068] In step S607, the broadcaster client generates a cartoon character using a cartoon topology model according to the needs of the scene, renders the live broadcast screen, and sends it to the live broadcast server so that the live broadcast server can distribute it to each viewer client.

[0069] Correspondingly, such as Figure 7 As shown, this application also provides a device 70 for constructing texture maps of virtual avatars, including a processor 71; a memory 72 for storing executable instructions, the memory 72 including a computer program; wherein, the processor 71 is configured to:

[0070] A three-dimensional target image is acquired, and a face is reconstructed based on the target three-dimensional target image to obtain a first realistic mesh model. A cartoon topological structure model is obtained based on the cartoon topological structure corresponding to the target three-dimensional target image. The aforementioned realistic mesh model and cartoon topological structure model contain different numbers of vertices and / or have vertices in different positions.

[0071] The first realistic mesh model is transferred to the cartoon topology model to obtain the second realistic mesh model, so that the number of vertices and the vertex positions are the same in the second realistic mesh model and the cartoon topology model.

[0072] By utilizing the mapping relationship between the texture coordinates of the second realistic mesh model and the cartoon topology model, the texture coordinates of the first realistic mesh model are mapped onto the aforementioned cartoon topology model, so as to construct the target texture map in the aforementioned cartoon topology model.

[0073] The processor 71 can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or any conventional processor.

[0074] The memory 72 may include at least one type of storage medium, including flash memory, hard disk, multimedia card, card-type memory (e.g., SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disk, optical disk, etc. Furthermore, the device can cooperate with network storage devices that perform storage functions via a network connection. The memory 72 may be an internal storage unit of the device 70, such as the hard disk or RAM of the device 70. The memory 72 may also be an external storage device of the device 70, such as a plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the device 70.

[0075] Furthermore, memory 72 may include both internal storage units of device 70 and external storage devices. Memory 72 is used to store computer programs and other programs and data required by the device. Memory 72 can also be used to temporarily store data that has been output or will be output.

[0076] Device 70 can be a computing device such as a desktop computer, laptop, handheld computer, or cloud server. The device may include, but is not limited to, a processor 71 and a memory 72. Those skilled in the art will understand that... Figure 7 This is merely an example of device 70 and does not constitute a limitation on device 70. It may include more or fewer components than shown, or combine certain components, or different components. For example, the device may also include input / output devices, network access devices, buses, etc.

[0077] The specific implementation process of the functions and roles of each unit in the above-mentioned equipment can be found in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

Claims

1. A method for constructing a virtual character texture map, characterized in that, The method includes: A three-dimensional target image is acquired, and a face is reconstructed based on the three-dimensional target image to obtain a first realistic mesh model. A cartoon topological structure model is obtained based on the cartoon topological structure corresponding to the three-dimensional target image. The first realistic mesh model and the cartoon topological structure model contain different numbers of vertices or have vertices in different positions. Based on the first alignment parameter, a topological transfer is performed on the first realistic mesh model to obtain a second realistic mesh model, so that the number of vertices and the vertex positions of the second realistic mesh model and the cartoon topological structure model are the same. After the first realistic mesh model is transformed into the second realistic mesh model through topological transfer, the texture coordinates of the first realistic mesh model will not change during the topological transfer. The first alignment parameter is a parameter that reflects the spatial relationship between the vertices of the first realistic mesh model and the vertices of the cartoon topological structure model during the topological transfer process. Based on the first alignment parameter, the mapping relationship between the texture coordinates of the second realistic mesh model and the cartoon topology model is determined. Using the mapping relationship, the texture coordinates of the first realistic mesh model are mapped to the cartoon topology model to construct a target texture map in the cartoon topology model, so that the cartoon topology model has realistic texture features.

2. The method according to claim 1, characterized in that, Based on the first alignment parameter, a topology transfer is performed on the first real-person mesh model, including: Obtain the first alignment parameter between the first realistic mesh model and the cartoon topology model, and align the number and position of vertices of the first realistic mesh model and the cartoon topology model according to the first alignment parameter.

3. The method according to claim 1, characterized in that: The steps for determining the mapping relationship include: Based on the first alignment parameter, determine the texture coordinates corresponding to the texture coordinates of the cartoon topology model in the second realistic mesh model; determine the mapping relationship based on the texture coordinates of the second realistic mesh model and the texture coordinates of the cartoon topology model.

4. The method according to claim 1, characterized in that, The steps for constructing a target texture map in a cartoon topology model include: The mesh is reconstructed based on the mapped texture coordinates, and the target texture map is constructed based on the pixels within the mesh.

5. The method according to claim 4, characterized in that, The steps for reconstructing the mesh include: A Drone triangle is constructed based on the mapped texture coordinates, and a target texture map is constructed based on the pixels within the grid of the Drone triangle.

6. An apparatus for constructing a virtual image texture map, characterized in that, The device includes: The modeling module is used to reconstruct faces based on a 3D target image to obtain a first realistic mesh model, and to obtain a cartoon topology model based on a cartoon topology corresponding to the 3D target image; the first realistic mesh model and the cartoon topology model contain different numbers of vertices or have vertices in different positions; The topology transfer module is used to perform topology transfer on the first realistic mesh model based on a first alignment parameter to obtain a second realistic mesh model, so that the number of vertices and vertex positions of the second realistic mesh model and the cartoon topology model are the same. After the first realistic mesh model is transformed into the second realistic mesh model through topology transfer, the texture coordinates based on the first realistic mesh model will not change during the topology transfer. The first alignment parameter is a parameter that reflects the spatial relationship between the vertices of the first realistic mesh model and the vertices of the cartoon topology model during the topology transfer process. The texture transfer module is used to determine the mapping relationship between the texture coordinates of the second realistic mesh model and the cartoon topology model based on the first alignment parameter, and to map the texture coordinates of the first realistic mesh model to the cartoon topology model based on the mapping relationship, so as to construct a target texture map in the cartoon topology model and make the cartoon topology model have realistic texture features.

7. The apparatus according to claim 6, characterized in that, The topology transfer module is used to perform topology transfer on the first real-life mesh model based on the first alignment parameter, including: Obtain the first alignment parameter between the first realistic mesh model and the cartoon topology model, and align the number and position of vertices of the first realistic mesh model and the cartoon topology model according to the first alignment parameter.

8. The apparatus according to claim 6, characterized in that: The topology migration module determines the mapping relationship by: determining the corresponding texture coordinates of the second realistic mesh model in the second realistic mesh model based on the texture coordinates of the cartoon topology model and the first alignment parameter; and determining the mapping relationship based on the texture coordinates of the second realistic mesh model and the texture coordinates of the cartoon topology.

9. The apparatus according to claim 6, characterized in that, The texture transfer module is used to construct the target texture map in the cartoon topology model, including: The mesh is reconstructed based on the mapped texture coordinates, and the target texture map is constructed based on the pixels within the mesh.

10. The apparatus according to claim 9, characterized in that, The texture transfer module reconstructs the mesh based on the mapped texture coordinates, and constructs the target texture map based on the pixels within the mesh, including: A Drone triangle is constructed based on the mapped texture coordinates, and a target texture map is constructed based on the pixels within the grid of the Drone triangle.

11. An electronic device, characterized in that, include: processor; A memory for storing processor-executable instructions; wherein the processor is configured to perform the operation as described in any one of claims 1-5.

12. A computer-readable storage medium storing computer instructions thereon, characterized in that: When executed by the processor, this instruction performs the operation described in any one of claims 1-5.

Images