Three-dimensional role reconstruction method and related equipment

Abstract

The invention discloses a three-dimensional role reconstruction method and related equipment. Relates to the field of image processing, and the method comprises the steps: generating a three-dimensional head model based on a two-dimensional face picture, the three-dimensional head model being a three-dimensional grid structure comprising vertexes, edges and surfaces, the three-dimensional grid structure comprising a static shape, a facial expression and motion information; according to the three-dimensional head model and the three-dimensional role template, automatic combination operation is carried out through a geometric butt joint algorithm, so that a three-dimensional role model is obtained, and the three-dimensional role template is a whole-body three-dimensional template.

Description

Technical Field

[0001] The present invention relates to the field of image processing, and more specifically, to a three-dimensional character reconstruction method and related equipment. Background Art

[0002] In the field of 3D character reconstruction technology, traditional 3D reconstruction technology usually has the following problems and limitations:

[0003] 1. Multiple photos are required for reconstruction: Many existing 3D character reconstruction technologies (such as advanced reconstruction software such as Avaturn) usually require users to provide three or more photos to generate a high-quality 3D facial model. For users who can only provide a single photo, the generation effect will be limited and it will be difficult to obtain a 3D facial model with complete details.

[0004] 2. The fusion of head and body requires manual operation: Most existing technologies can only generate a 3D head model of a character, such as the technology described in Paper 1. The generated head model needs to be fused with a pre-designed body model, and this process often requires manual operation using professional modeling software, which is not only time-consuming and laborious, but also requires a high level of technical skills from the user and is difficult to automate.

[0005] 3. Lack of skeletal information for animation support: Some 3D character reconstruction technologies (such as those described in Paper 2) can generate full-body models, but the generated 3D models lack skeletal information and cannot be directly used in animation scenes. Users need to manually add skeletal structures, a process that also relies on professional software and manual operation, increasing the difficulty and time cost of model production.

[0006] 4. Poor texture quality: Existing technologies usually map facial texture maps directly from a single facial photo. However, due to limitations such as the photo shooting angle and lighting conditions, texture maps generated by direct mapping usually have noise and unevenness, which affects the visual quality of the final model.

[0007] 5. Single style: Most current 3D character generation technologies are based on real-style character photos for reconstruction, lacking diverse style support (such as cartoon style). This makes it difficult for existing technologies to meet the needs of multiple styles in different application scenarios, such as cartoon character models in games and virtual social networking.

[0008] Therefore, in view of the problems existing in the existing technology when generating three-dimensional characters, such as high requirements for photo input, reliance on manual operation for head and body fusion and skeleton generation, low texture mapping quality and single style, there is an urgent need for an improved three-dimensional character reconstruction technology solution to at least solve some of the above problems. Summary of the invention

[0009] A series of simplified concepts are introduced in the Summary of the Invention, which will be further described in detail in the Detailed Description of the Invention. The Summary of the Invention does not mean to attempt to define the key features and essential technical features of the claimed technical solution, nor does it mean to attempt to determine the scope of protection of the claimed technical solution.

[0010] In a first aspect, the present invention provides a three-dimensional character reconstruction method, the method comprising:

[0011] Generating a three-dimensional head model based on a two-dimensional facial photo, wherein the three-dimensional head model is a three-dimensional mesh structure including vertices, edges and faces, and the three-dimensional mesh structure includes static shape, facial expression and motion information;

[0012] The three-dimensional head model and the three-dimensional character template are automatically combined through a geometric docking algorithm to obtain a three-dimensional character model, wherein the three-dimensional character template is a full-body three-dimensional template.

[0013] In a feasible implementation manner, the above-mentioned generation of a three-dimensional head model based on a two-dimensional facial photo includes:

[0014] Generate a real person's three-dimensional head model based on the above two-dimensional facial photo through the DECA model;

[0015] Generate a cartoon 3D head model based on the 3D real person 3D head model;

[0016] The three-dimensional head model is generated based on the three-dimensional head model of the real person and the three-dimensional head model of the cartoon.

[0017] In a feasible implementation, it also includes:

[0018] The three-dimensional head model is subjected to noise identification and separation operations, template matching and noise removal operations, and detail repair and adjustment operations to obtain an optimized three-dimensional head model, wherein the three-dimensional head model is based on facial details extracted from the two-dimensional photo and converted into a texture map that can be applied to the surface of the three-dimensional model.

[0019] In a feasible implementation manner, the above-mentioned noise recognition and separation operation is based on artificial intelligence algorithm for recognition, and noise removal is performed through edge detection and feature point recognition.

[0020] In a feasible implementation manner, the above template matching and noise removal operations are performed based on Poisson image editing technology.

[0021] In a feasible implementation, the above-mentioned detail repair and adjustment operations include further refined processing of complex texture problems based on an AI model.

[0022] In a feasible implementation manner, the automatic combination operation based on the three-dimensional head model and the three-dimensional character template through a geometric docking algorithm includes:

[0023] Performing vertex information modification and standardization processing operations on the three-dimensional head model to obtain a processed three-dimensional head model;

[0024] Based on the processed three-dimensional head model, the head template in the three-dimensional character template is replaced according to the index order of the vertices to perform an automatic combination operation.

[0025] In a second aspect, the present invention further provides a three-dimensional character reconstruction device, comprising:

[0026] A generating unit, configured to generate a three-dimensional head model based on a two-dimensional facial photograph, wherein the three-dimensional head model is a three-dimensional mesh structure including vertices, edges and faces, and the three-dimensional mesh structure includes static shape, facial expression and motion information;

[0027] The acquisition unit is used to automatically combine the three-dimensional head model and the three-dimensional character template through a geometric docking algorithm to obtain a three-dimensional character model, wherein the three-dimensional character template is a full-body three-dimensional template.

[0028] In a third aspect, an electronic device comprises: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor is configured to implement the steps of any one of the three-dimensional character reconstruction methods of the first aspect when executing the computer program stored in the memory.

[0029] In a fourth aspect, the present invention further proposes a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the three-dimensional character reconstruction method of any one of the above items in the first aspect is implemented.

[0030] In summary, the present invention overcomes the problems of large data requirements and cumbersome manual operations in traditional three-dimensional character reconstruction by introducing artificial intelligence and image synthesis technology. The present invention proposes a technical solution for generating three-dimensional characters based on artificial intelligence and realizing automatic fusion with body templates. A three-dimensional character model can be generated with only one two-dimensional photo, and the input data requirement is low. Using a pre-designed body template, the three-dimensional character head model can automatically connect the geometric structure with the body model, eliminating manual adjustment. The generated model supports animation and can be used in a variety of application scenarios such as games and virtual reality. Through the synthesis of Poisson image editing technology and noise-free texture templates, redundant texture noise is removed, ensuring that the noise removal process is automated and refined, and achieving high-quality three-dimensional character reconstruction. Compared with the prior art, the present invention requires less input data, includes a processing process for removing map noise, and automatically merges with the body model, so it can be used directly in the desired virtual environment. ,

[0031] The three-dimensional character reconstruction method of the present invention, and other advantages, objectives and features of the present invention will be partially reflected through the following description, and will also be partially understood by those skilled in the art through research and practice of the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0032] Various other advantages and benefits will become apparent to those of ordinary skill in the art by reading the detailed description of the preferred embodiments below. The accompanying drawings are only for the purpose of illustrating the preferred embodiments and are not to be considered as limiting the present specification. Also, the same reference symbols are used throughout the accompanying drawings to represent the same components. In the accompanying drawings:

[0033] Figure 1 A schematic diagram of a three-dimensional character reconstruction method provided in an embodiment of the present application;

[0034] Figure 2 A schematic diagram of the principle of a three-dimensional character reconstruction method provided in an embodiment of the present application;

[0035] Figure 3 A scenario schematic diagram of a three-dimensional character reconstruction method provided in an embodiment of the present application;

[0036] Figure 4 A schematic diagram of the structure of a three-dimensional character reconstruction device provided in an embodiment of the present application;

[0037] Figure 5 A schematic diagram of the structure of an electronic device for three-dimensional character reconstruction provided in an embodiment of the present application. DETAILED DESCRIPTION

[0038] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices. The technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all of the embodiments.

[0039] See also Figures 1 to 3 , Figure 1 2 is a schematic diagram of a three-dimensional character reconstruction method provided in an embodiment of the present application; 3 is a schematic diagram of a three-dimensional character reconstruction method provided in an embodiment of the present application; Figure 3 A scenario schematic diagram of a three-dimensional character reconstruction method provided in an embodiment of the present application;

[0040] The method may specifically include:

[0041] S110, generating a three-dimensional head model based on the two-dimensional facial photo, wherein the three-dimensional head model is a three-dimensional grid structure including vertices, edges and faces, and the three-dimensional grid structure includes static shape, facial expression and motion information;

[0042] Exemplarily, in this step, a 3D head model is generated from a single 2D facial photo uploaded by a user through an artificial intelligence algorithm. First, an artificial intelligence model (such as "DECA: Detailed Expression Capture and Animation") is used to recognize the face in the photo and extract key image features. These features are used to construct a 3D mesh structure containing vertices, edges, and faces, which forms the basis of the 3D head model.

[0043] The generated 3D mesh structure includes not only the static shape of the face, but also expression and motion information. This dynamic information provides support for subsequent animation applications. In addition, unlike the standard "DECA" model (which only supports the generation of real people), this scheme supports the generation of 3D characters in multiple styles (such as real photo style, cartoon style). Therefore, during the 3D mesh generation process, the system constructs a 3D head basic model that supports different styles to adapt to different types of photo inputs.

[0044] S120, automatically combining the three-dimensional head model and the three-dimensional character template through a geometric docking algorithm to obtain a three-dimensional character model, wherein the three-dimensional character template is a full-body three-dimensional template.

[0045] Exemplarily, in this step, the generated three-dimensional head model is automatically combined with a pre-designed full-body three-dimensional character template through a geometric docking algorithm to complete the construction of the three-dimensional character model.

[0046] The 3D head model is automatically combined with the 3D character template through a geometric docking algorithm. The algorithm ensures that the head model and the body template fully match the geometric structure of the contact surface (including vertices, edges and faces).

[0047] To achieve precise docking, the system-designed full-body 3D template uses a standardized 3D grid structure at the connection between the head and the body to ensure that the vertices and grids on the contact surface are completely consistent. In this way, the system can automatically complete the fusion operation of the head and the body without manual adjustment. Finally, a complete 3D character model is generated, and the character model supports animation applications. Through the above steps, the system realizes the automatic generation and combination operation from 2D photos to 3D characters, ensuring high efficiency and animation.

[0048] In a feasible implementation manner, the above-mentioned generation of a three-dimensional head model based on a two-dimensional facial photo includes:

[0049] Generate a real person's three-dimensional head model based on the above two-dimensional facial photo through the DECA model;

[0050] Generate a cartoon 3D head model based on the 3D real person 3D head model;

[0051] The three-dimensional head model is generated based on the three-dimensional head model of the real person and the three-dimensional head model of the cartoon.

[0052] Exemplarily, a three-dimensional real person head model is generated from a two-dimensional facial photo uploaded by a user through the artificial intelligence model DECA (Detailed Expression Capture and Animation). The DECA model first performs face recognition on the uploaded two-dimensional photo to identify the main facial features, such as the outlines of the eyes, nose, mouth and facial expression features. After extracting the facial features, the DECA model converts the two-dimensional image information into a three-dimensional mesh model. The three-dimensional mesh model contains a structure composed of vertices, edges, and faces, which can faithfully restore the static shape of the face as well as expression and motion information, so that the generated model has a foundation for animation support. The head model generated by the DECA model truly reflects the real facial features in the photo and is a three-dimensional head model with a realistic style.

[0053] After obtaining the 3D head model of a real person, the system further generates a cartoon-style 3D head model. In order to create a cartoon effect, the system performs style conversion on the real person head model. This may include using a cartoonization algorithm to adjust the shape features of the model so that it has cartoon features (for example, exaggerated eyes, simplified facial contours, etc.). In this conversion process, a generative adversarial network (GAN) or other cartoonization algorithms can be used to adjust the positions of vertices, edges, and faces of the 3D mesh to make it more consistent with the visual effects of the cartoon style. This process may involve simplifying and smoothing real facial features to achieve a cartoon-style effect. When generating a cartoon style, the basic structure of the cartoon model still maintains a certain consistency with the real person head model, so that the style can be easily switched later and the compatibility of the character between cartoon and real styles is guaranteed.

[0054] The system combines the real person and cartoon-style 3D head models to generate the final 3D head model, ensuring that the model is both realistic and can support cartoon effects. The structure of the real person 3D head model and the cartoon 3D head model are combined to create a comprehensive model that can switch between the two styles. The model contains the details of the real person model (such as facial expressions and movement information), while also having the style characteristics of the cartoon model.

[0055] In a feasible implementation, it also includes:

[0056] The three-dimensional head model is subjected to noise identification and separation operations, template matching and noise removal operations, and detail repair and adjustment operations to obtain an optimized three-dimensional head model, wherein the three-dimensional head model is based on facial details extracted from the two-dimensional photo and converted into a texture map that can be applied to the surface of the three-dimensional model.

[0057] Exemplarily, facial details are first extracted from a two-dimensional photo and converted into a texture map suitable for a three-dimensional model. This texture map contains the texture details of the face, but may also contain some noise, so the three-dimensional head model needs to be processed.

[0058] Detect and mark the noisy areas in the texture map for noise identification and separation. Remove excess noise by comparing with the standard template. Further process the complex areas (such as the hair) to ensure that the final texture map is clear and free of interference. Through the above steps, a high-quality, noise-free optimized 3D head model is obtained.

[0059] In a feasible implementation manner, the above-mentioned noise recognition and separation operation is based on artificial intelligence algorithm for recognition, and noise removal is performed through edge detection and feature point recognition.

[0060] Exemplarily, the system uses artificial intelligence algorithms to identify and separate noise from texture maps. AI algorithms are used to identify noise areas in texture maps, such as areas with uneven textures due to shooting angles or uneven lighting. Edge detection and feature point recognition techniques are used to further separate noise areas in texture images to distinguish facial details from other unnecessary interference information.

[0061] Through this step, the noise in the 3D head model is effectively identified, preparing for the subsequent noise removal operation.

[0062] In a feasible implementation manner, the above template matching and noise removal operations are performed based on Poisson image editing technology.

[0063] Exemplarily, the Poisson image editing technique is used to perform template matching and noise removal operations. The generated texture map is matched and compared with a pre-prepared noise-free standard template. Through the Poisson image editing technique, the noise area is synthesized with the standard template to remove the noise in the texture map, making the texture smoother and more natural. This method ensures the texture quality of the generated 3D head model and eliminates unnecessary interference caused by the angle of the photo or the lighting.

[0064] In a feasible implementation, the above-mentioned detail repair and adjustment operations include further refined processing of complex texture problems based on an AI model.

[0065] Exemplarily, in this embodiment, the system performs detail repair and adjustment operations on complex areas in the texture map (such as hair interference on the forehead). For example, an AI model such as a generative adversarial network (GAN) is used to generate a pure face image without hair interference. The generated interference-free face image is mapped to the three-dimensional model to make the texture map cleaner and avoid the influence of complex areas (such as hair) on the facial texture. In this way, the system can further refine and optimize the texture map to ensure the high visual quality and authenticity of the three-dimensional character model.

[0066] In a feasible implementation manner, the automatic combination operation based on the three-dimensional head model and the three-dimensional character template through a geometric docking algorithm includes:

[0067] Performing vertex information modification and standardization processing operations on the three-dimensional head model to obtain a processed three-dimensional head model;

[0068] Based on the processed three-dimensional head model, the head template in the three-dimensional character template is replaced according to the index order of the vertices to perform an automatic combination operation.

[0069] For example, the generated new head model may have different geometric structures, such as the number of vertices, vertex arrangement order or face connection method, which are inconsistent with the template head. In order to be compatible with the template head, the vertex information of the new head model needs to be adjusted to adapt to the skeleton binding structure of the template model.

[0070] The geometry of the new head model is standardized using the Vertex Reordering Based on FaceReferences algorithm.

[0071] The specific operations include: making sure the number of vertices of the new head model is consistent with the template head model. Adjusting the vertex order and the arrangement of the faces to make the two completely match. This step ensures that the new head model can be seamlessly connected to the template character without losing the bone binding information.

[0072] After the standardization process is completed, the newly generated head model replaces the head vertex information of the original template according to the vertex index order of the template model's head. This replacement only involves the position adjustment of the geometric vertices, without changing the bone structure and binding information of the original head model. The association information (such as weight distribution) between the bones and head vertices in the template model has been pre-established. The replacement is only a change in the position of the geometric vertices, and the binding relationship between the bones and the geometric vertices is not affected. Therefore, there is no need to re-perform the bone binding operation, saving a lot of time and effort for manual adjustment. The newly generated head model is perfectly connected to the head position of the template model through vertex position replacement.

[0073] In summary, the present invention overcomes the problems of large data requirements and cumbersome manual operations in traditional three-dimensional character reconstruction by introducing artificial intelligence and image synthesis technology. The present invention proposes a technical solution for generating a three-dimensional character based on artificial intelligence and realizing automatic fusion with a body template. A three-dimensional character model can be generated with only one two-dimensional photo, and the input data requirement is low. Using a pre-designed body template, the three-dimensional character head model can automatically connect with the body model for geometric structure, eliminating manual adjustment. The generated model supports animation and can be used in various application scenarios such as games and virtual reality. Through the Poisson image editing technology and the noise-free texture template for synthesis, the redundant texture noise is removed, the noise removal process is ensured to be automated and refined, and high-quality three-dimensional character reconstruction is achieved. Compared with the prior art, the present invention requires less input data, includes a process for removing map noise, and automatically merges with the body model, so it can be directly used in the desired virtual environment.

[0074] See also Figure 2 , an embodiment of the three-dimensional character reconstruction device in the embodiment of the present application may include:

[0075] A generating unit 21, configured to generate a three-dimensional head model based on a two-dimensional facial photo, wherein the three-dimensional head model is a three-dimensional grid structure including vertices, edges and faces, and the three-dimensional grid structure includes static shape, facial expression and motion information;

[0076] The acquisition unit 22 is used to automatically combine the three-dimensional head model and the three-dimensional character template through a geometric docking algorithm to obtain a three-dimensional character model, wherein the three-dimensional character template is a full-body three-dimensional template.

[0077] like Figure 3 As shown, an embodiment of the present application also provides an electronic device 300, including a memory 310, a processor 320, and a computer program 311 stored in the memory 320 and executable on the processor, and when the processor 320 executes the computer program 311, the steps of any of the above-mentioned methods for three-dimensional character reconstruction are implemented.

[0078] Since the electronic device introduced in this embodiment is a device used to implement a three-dimensional character reconstruction method in the embodiment of the present application, based on the method introduced in the embodiment of the present application, technical personnel in this field can understand the specific implementation method of the electronic device of this embodiment and its various variations. Therefore, how the electronic device implements the method in the embodiment of the present application is not introduced in detail here. As long as the equipment used by technical personnel in this field to implement the method in the embodiment of the present application is within the scope of protection of this application.

[0079] In the specific implementation process, when the computer program 311 is executed by the processor, it can achieve Figure 1 Any implementation manner in the corresponding embodiments.

[0080] It should be noted that in the above embodiments, the description of each embodiment has its own emphasis, and for parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0081] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

[0082] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded computer, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0083] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0084] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0085] The present application also provides a computer program product, which includes computer software instructions. When the computer software instructions are executed on a processing device, the processing device is caused to execute the following Figure 1 The process of 3D character reconstruction in the corresponding embodiment.

[0086] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on the computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website site, a computer, a server or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (digital subscriber line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website site, computer, server or data center. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a server or a data center that includes one or more available media integration. The available medium can be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid state drive (SSD)), etc.

[0087] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0088] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0089] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0090] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0091] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), disk or optical disk and other media that can store program codes.

[0092] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A three-dimensional character reconstruction method, characterized in that: include: Generating a three-dimensional head model based on a two-dimensional facial photo, wherein the three-dimensional head model is a three-dimensional mesh structure including vertices, edges and faces, and the three-dimensional mesh structure includes static shape, facial expression and motion information; The three-dimensional head model and the three-dimensional character template are automatically combined through a geometric docking algorithm to obtain a three-dimensional character model, wherein the three-dimensional character template is a full-body three-dimensional template.

2. The three-dimensional character reconstruction method according to claim 1, characterized in that: The method of generating a three-dimensional head model based on a two-dimensional facial photo comprises: Generate a three-dimensional head model of a real person through a DECA model based on the two-dimensional facial photo; Generate a cartoon 3D head model based on the 3D real person 3D head model; The three-dimensional head model is generated based on the real person three-dimensional head model and the cartoon three-dimensional head model.

3. The three-dimensional character reconstruction method according to claim 1, characterized in that: Also includes: The three-dimensional head model is subjected to noise identification and separation operations, template matching and noise removal operations, and detail repair and adjustment operations to obtain an optimized three-dimensional head model, wherein the three-dimensional head model is converted into a texture map that can be applied to the surface of the three-dimensional model based on facial details extracted from the two-dimensional photo.

4. The three-dimensional character reconstruction method according to claim 3, characterized in that: The noise recognition and separation operation is based on artificial intelligence algorithm for recognition, and noise removal is performed through edge detection and feature point recognition.

5. The three-dimensional character reconstruction method according to claim 3, characterized in that: The template matching and noise removal operations are performed based on the Poisson image editing technology.

6. The three-dimensional character reconstruction method according to claim 3, characterized in that: The detail repair and adjustment operations include further refined processing of complex texture problems based on the AI ​​model.

7. The three-dimensional character reconstruction method according to claim 1, characterized in that: The automatic combination operation based on the three-dimensional head model and the three-dimensional character template through a geometric docking algorithm includes: Performing vertex information modification and standardization processing operations on the three-dimensional head model to obtain a processed three-dimensional head model; Based on the processed three-dimensional head model, the head template in the three-dimensional character template is replaced according to the index order of the vertices to perform an automatic combination operation.

8. A three-dimensional character reconstruction device, characterized in that: include: A generating unit, configured to generate a three-dimensional head model based on a two-dimensional facial photograph, wherein the three-dimensional head model is a three-dimensional mesh structure including vertices, edges and faces, and the three-dimensional mesh structure includes static shape, facial expression and motion information; An acquisition unit is used to automatically combine the three-dimensional head model and the three-dimensional character template through a geometric docking algorithm to acquire a three-dimensional character model, wherein the three-dimensional character template is a full-body three-dimensional template.

9. An electronic device, comprising: A memory and a processor, wherein the processor is used to implement the steps of the three-dimensional character reconstruction method according to any one of claims 1 to 7 when executing the computer program stored in the memory.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the three-dimensional character reconstruction method according to any one of claims 1 to 7 is implemented.

Images