Hair area processing method and device of three-dimensional model, equipment and medium

Abstract

The invention discloses a hair area processing method and device of a three-dimensional model, equipment and a medium, and the method comprises the steps: carrying out the five-sense-organ segmentation of a color image of a human face, and obtaining a mask of a hair area; projecting the mask of the hair area to an original face point cloud, and extracting a hair area point cloud and a second face point cloud not including the hair area point cloud; and smoothing the hair area point cloud, and adding the smoothed hair area point cloud to the second face point cloud to obtain a third face point cloud. The method can effectively solve the challenge faced by the depth camera when the depth camera captures hair texture information, improves the quality and reality of point cloud reconstruction, and improves the modeling effect and image reconstruction effect of a three-dimensional reconstruction model.

Description

Technical Field

[0001] The present invention belongs to the technical field of image processing, and particularly relates to a method, device, equipment and medium for processing the hair area of a 3D model. Background Art

[0002] With the rapid development of 3D scanning and reconstruction technologies, depth cameras have become a bridge connecting the physical world and the digital world, and are widely used in multiple fields such as virtual reality, healthcare, and portrait recognition. However, in the digital process of pursuing high fidelity and realism, the accurate reconstruction of hair remains a technical challenge. Hair, especially in the temple area, poses a huge challenge to existing depth cameras due to its fine, dense, complex and changeable structure and the dynamic characteristics of natural fluttering.

[0003] Depth cameras measure the distance of the object surface by emitting and receiving infrared light or structured light, and then generate point cloud data. However, the diameter of hair strands is small, and light is easily transmitted or scattered, resulting in sparse and discontinuous point cloud data. Especially at the hair-skin junction (such as the temple area), the abnormal accumulation of point cloud data often leads to unnatural protrusions or holes in the reconstructed mesh. This not only affects the modeling of the model, but also limits the practicality and user experience of 3D reconstruction in applications such as hairstyle design, virtual try-on, and hair transplant simulation. Summary of the Invention

[0004] In order to overcome one or more of the above technical defects, the present invention provides a method and device for processing the hair area of a 3D model, which can effectively solve the challenges faced by depth cameras in capturing hair texture information, improve the quality and realism of point cloud reconstruction, and enhance the modeling effect and image reconstruction effect of the 3D reconstruction model.

[0005] To solve the above problems, the first aspect of the present invention provides a method for processing the hair area of a 3D model, including:

[0006] Performing facial feature segmentation on the color image of the human face to obtain a mask of the hair area;

[0007] Projecting the mask of the hair area onto the original human face point cloud, and extracting the hair area point cloud and the second human face point cloud that does not include the hair area point cloud;

[0008] Performing smoothing processing on the hair area point cloud, and adding the smoothed hair area point cloud to the second human face point cloud to obtain a third human face point cloud.

[0009] Further, the performing facial feature segmentation on the color image of the human face to obtain a mask of the hair area includes:

[0010] Perform five - sense segmentation on the color image of the human face to obtain the five - sense segmentation result;

[0011] Identify and extract the hair part in the five - sense segmentation result as the mask of the hair region.

[0012] Further, the step of smoothing the hair region point cloud and adding the smoothed hair region point cloud to the second human face point cloud includes:

[0013] Define four different radii and generate neighborhood point clouds with different distances from the hair region point cloud based on the four different radii. The neighborhood point clouds include the first neighborhood point cloud, the second neighborhood point cloud, the third neighborhood point cloud, and the fourth neighborhood point cloud;

[0014] Merge the first neighborhood point cloud, the second neighborhood point cloud, the third neighborhood point cloud, and the fourth neighborhood point cloud to obtain the area to be smoothed;

[0015] Use MLS to smooth the area to be smoothed to obtain the first sub - neighborhood point cloud, the second sub - neighborhood point cloud, the third sub - neighborhood point cloud, and the fourth sub - neighborhood point cloud;

[0016] Perform data processing on the first sub - neighborhood point cloud, the second sub - neighborhood point cloud, the third sub - neighborhood point cloud, and the fourth sub - neighborhood point cloud respectively;

[0017] Add the processed third sub - neighborhood point cloud and fourth sub - neighborhood point cloud to the second human face point cloud.

[0018] Further, the step of performing data processing on the first sub - neighborhood point cloud, the second sub - neighborhood point cloud, the third sub - neighborhood point cloud, and the fourth sub - neighborhood point cloud respectively includes:

[0019] Perform replacement processing on the first sub - neighborhood point cloud and the second sub - neighborhood point cloud respectively;

[0020] Perform weighted summation on the third sub - neighborhood point cloud;

[0021] Replace the fourth sub - neighborhood point cloud with the fourth neighborhood point cloud.

[0022] The second aspect of the present invention provides a hair region processing device for a three - dimensional model, which is used to implement the above - mentioned hair region processing method for a three - dimensional model, and includes:

[0023] A five - sense segmentation module, which is used to perform five - sense segmentation on the color image of the human face to obtain the mask of the hair region;

[0024] A projection module, which is used to project the mask of the hair region onto the original human face point cloud, extract the hair region point cloud and the second human face point cloud that does not include the hair region point cloud;

[0025] A hair region smoothing module, which is used to smooth the point cloud of the hair region, and add the smoothed point cloud of the hair region to the second face point cloud to obtain a third face point cloud.

[0026] The third aspect of the present invention provides a method for three-dimensional face reconstruction, including:

[0027] Collecting face image data, where the face image data includes a color image and a depth image of the face;

[0028] Using the above-mentioned hair region processing method of the three-dimensional model to process the hair region of the color image of the face;

[0029] Inputting the third face point cloud including the smoothed point cloud of the hair region into the three-dimensional reconstruction model to obtain the three-dimensional face image data.

[0030] The fourth aspect of the present invention provides a computer device, including a memory and a processor, where the memory stores a computer program, and when the processor executes the computer program, the steps of the above-mentioned method are implemented.

[0031] The fifth aspect of the present invention provides a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the above-mentioned method are implemented.

[0032] Compared with the prior art, the present invention has the following beneficial effects:

[0033] The present invention discloses a method and device for processing the hair region of a three-dimensional model, which performs five sense organs segmentation on the color image of the face to obtain a mask of the hair region; combines the mask of the hair region and the original face point cloud to obtain the point cloud of the hair region, smooths the point cloud of the hair region, and adds the smoothed point cloud of the hair region to the second face point cloud that does not include the point cloud of the hair region to obtain the face point cloud after hair smoothing. The method proposed by the present invention specifically targets complex regions such as sideburns, can effectively solve the challenges faced by depth cameras in capturing hair texture information, improve the quality and realism of point cloud reconstruction, and enhance the modeling effect and image reconstruction effect of the three-dimensional reconstruction model. Description of the Drawings

[0034] The following further details the specific implementation manners of the present invention in conjunction with the drawings, where:

[0035] Figure 1 is the flow of the hair region processing method of the three-dimensional model described in Embodiment 1 Figure 1 ;

[0036] Figure 2 is the flow of the hair region processing method of the three-dimensional model described in Embodiment 1 Figure 2 ;

[0037] Figure 3 Schematic diagram of the structure of the hair region processing device for the 3D model described in Embodiment 2;

[0038] Figure 4 Flowchart of the 3D face reconstruction method described in Embodiment 3;

[0039] Figure 5 Schematic diagram of the structure of the computer device described in Embodiment 4;

[0040] Marking description: 110, acquisition module; 120, data processing module; 130, smoothing module. Detailed implementation manners

[0041] The following describes the preferred embodiments of the present invention with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

[0042] Embodiment 1

[0043] This embodiment discloses a method for processing the hair region of a 3D model, as shown in Figure 1 and 2 , including:

[0044] S1. Perform five - sense segmentation on the color image color_image of the human face to obtain a mask of the hair region.

[0045] In this embodiment, step S1 includes:

[0046] Perform five - sense segmentation on the color image color_image of the human face to obtain a five - sense segmentation result. The five - sense segmentation algorithm is a prior art and will not be elaborated here.

[0047] Identify and extract the hair part in the five - sense segmentation result as the mask hair_segmentation_result of the hair region.

[0048] S2. Project the mask of the hair region onto the original human face point cloud, and extract the hair region point cloud and the second human face point cloud that does not include the hair region point cloud.

[0049] In order to smooth the 3D hair region point cloud, project the two - dimensional segmentation result of the hair region, that is, the mask hair_segmentation_result, onto the original human face point cloud to obtain the hair region point cloud hair_region. Filter the hair region point cloud hair_region in the original human face point cloud to obtain the second human face point cloud cloud_filtered that does not include the hair region point cloud.

[0050] S3. Smooth the point cloud of the hair region, and add the smoothed point cloud of the hair region to the second face point cloud to obtain a third face point cloud.

[0051] Specifically, step S3 includes:

[0052] Define four radii with different sizes, namely radius_1, radius_2, radius_3, and radius_4, and radius_1 < radius_2 < radius_3 < radius_4. Based on the four radii with different sizes and the distance search module based on the kd-tree index, generate four neighborhood point clouds with different distances from the point cloud of the hair region hair_region, including the first neighborhood point cloud neighbour_1, the second neighborhood point cloud neighbour_2, the third neighborhood point cloud neighbour_3, and the fourth neighborhood point cloud neighbour_4, and there is no intersection between the first neighborhood point cloud neighbour_1, the second neighborhood point cloud neighbour_2, the third neighborhood point cloud neighbour_3, and the fourth neighborhood point cloud neighbour_4. In this implementation, radius_1 = 0, that is, neighbour_1 = hair_region.

[0053] Merge the first neighborhood point cloud neighbour_1, the second neighborhood point cloud neighbour_2, the third neighborhood point cloud neighbour_3, and the fourth neighborhood point cloud neighbour_4 to obtain the region to be smoothed smoothing_regions.

[0054] Use MLS (moving least square) to smooth the region to be smoothed smoothing_regions to obtain the first sub-neighborhood point cloud neighbour_1', the second sub-neighborhood point cloud neighbour_2', the third sub-neighborhood point cloud neighbour_3', and the fourth sub-neighborhood point cloud neighbour_4'.

[0055] For the first sub-neighborhood point cloud neighbour_1', the second sub-neighborhood point cloud neighbour_2', the third sub-neighborhood point cloud neighbour_3' and the fourth sub-neighborhood point cloud neighbour_4', replacement processing is performed on neighbour_1' and neighbour_2' respectively, weighted summation is performed on neighbour_3', and neighbour_4' is reserved and replaced with neighbour_4 to obtain neighbour_1", neighbour_2", neighbour_3" and neighbour_4". Specifically, after replacement, neighbour_1" = neighbour_1', neighbour_2" = neighbour_2', that is, neighbour_1' and neighbour_2' remain unchanged during this process; neighbour_4" = neighbour_4, that is, the most original point cloud is reserved. neighbour_3" is the weighted summation of neighbour_3 and neighbour_3', and its weight is related to the shortest distance from each point in neighbour_3 to the neighbour_2 point cloud. The formula is:

[0056] point”=(1-α)*point+α*point'

[0057] where α = distance(point,neighbour_2) / (radius_3 - radius_2).

[0058] Add the processed third sub-neighborhood point cloud and fourth sub-neighborhood point cloud to the second face point cloud. Specifically, add neighbour_3" and neighbour_4" to the second face point cloud cloud_filtered.

[0059] Through the replacement, weighted summation and reservation processing of the first sub-neighborhood point cloud neighbour_1', the second sub-neighborhood point cloud neighbour_2', the third sub-neighborhood point cloud neighbour_3' and the fourth sub-neighborhood point cloud neighbour_4', the connection between the smoothed hair and the other point clouds of the original face is realized.

[0060] Specifically, the replacement, weighted summation and reservation processing can be performed n times as needed to obtain a better smoothing and connection effect, and the processed third sub-neighborhood point cloud and fourth sub-neighborhood point cloud are added to the second face point cloud to obtain the third face point cloud, that is, the three-dimensional face point cloud data with a smoothed hair area.

[0061] The 3D model hair area processing method provided by the present invention acquires the color image of a human face and the original human face point cloud, extracts the hair area as a mask through an existing facial feature segmentation algorithm, combines the mask of the hair area and the original human face point cloud to obtain the hair area point cloud, performs smoothing processing on the hair area point cloud, and adds the smoothed hair area point cloud to the second human face point cloud that does not include the hair area point cloud to obtain the human face point cloud with smoothed hair. The method proposed by the present invention specifically targets complex areas such as sideburns, can effectively solve the challenges faced by depth cameras in capturing hair texture information, improve the quality and realism of point cloud reconstruction, and enhance the modeling effect and image reconstruction effect of 3D reconstruction models.

[0062] Embodiment 2

[0063] This embodiment discloses a hair area processing device for a 3D model, as Figure 3 , which is used to implement the hair area processing method of the 3D model described in Embodiment 1, and includes a facial feature segmentation module 110, a projection module 120, and a hair area smoothing module 130. Specifically, the facial feature segmentation module 110 is used to perform facial feature segmentation on the color image of a human face to obtain a mask of the hair area; the projection module 120 is used to project the mask of the hair area onto the original human face point cloud to extract the hair area point cloud and the second human face point cloud that does not include the hair area point cloud; the hair area smoothing module 130 is used to perform smoothing processing on the hair area point cloud and add the smoothed hair area point cloud to the second human face point cloud to obtain the third human face point cloud.

[0064] For other specific implementation details, please refer to Embodiment 1 and will not be elaborated here.

[0065] Each module in the above-mentioned hair area processing device for a 3D model can be implemented in whole or in part by software, hardware, and their combination. The above-mentioned modules can be embedded in the processor of a computer device in hardware form or be independent of it, or can be stored in the memory of the computer device in software form, so that the processor can call and execute the operations corresponding to the above-mentioned modules.

[0066] Embodiment 3

[0067] This embodiment discloses a human face 3D reconstruction method, as Figure 4 , including:

[0068] Collect human face image data, where the human face image data includes the color image and depth image of a human face.

[0069] Use the hair area processing method of the 3D model described in Claim 1 to perform hair area processing on the color image of a human face.

[0070] Input the third - person face point cloud including the smoothed hair - region point cloud into the 3D reconstruction model to obtain the 3D face image data.

[0071] For other specific implementation details, please refer to Embodiment 1 and will not be elaborated here.

[0072] Combined with the hair - region processing method of the 3D model provided in Embodiment 1, it can improve the unnatural protrusions or holes in the mesh caused by complex regions such as sideburns during 3D face reconstruction in the prior art, and improve the modeling effect of the 3D reconstruction model and the image reconstruction effect.

[0073] Embodiment 4

[0074] This embodiment discloses a computer device, which can be a server or a terminal integrated with a scheduler, and its internal structure diagram can be as Figure 5 shown. The computer device includes a processor, a memory, and a network interface connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non - volatile storage medium and an internal memory. The non - volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non - volatile storage medium. The database of the computer device is used to store data. The network interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, it realizes a hair - region processing method of a 3D model.

[0075] Those skilled in the art can understand that Figure 5 the structure shown in

[0076] is only a block diagram of some structures related to the solution of this application, and does not constitute a limitation on the computer device to which the solution of this application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine some components, or have different component arrangements.

[0077] Perform five - sense segmentation on the color image of the face to obtain a mask of the hair region;

[0078] Project the mask of the hair region onto the original face point cloud, and extract the hair - region point cloud and the second face point cloud that does not include the hair - region point cloud;

[0079] Smooth the hair - region point cloud, and add the smoothed hair - region point cloud to the second face point cloud to obtain a third - person face point cloud.

[0080] In this embodiment, when the processor executes the computer program, the following steps are further implemented:

[0081] Perform facial feature segmentation on the color image of the face to obtain the facial feature segmentation result;

[0082] Identify and extract the hair part in the facial feature segmentation result as the mask of the hair region.

[0083] In this embodiment, when the processor executes the computer program, the following steps are further implemented:

[0084] Define four different radii, and generate neighborhood point clouds with different distances from the hair region point cloud based on the four different radii. The neighborhood point clouds include the first neighborhood point cloud, the second neighborhood point cloud, the third neighborhood point cloud, and the fourth neighborhood point cloud;

[0085] Merge the first neighborhood point cloud, the second neighborhood point cloud, the third neighborhood point cloud, and the fourth neighborhood point cloud to obtain the area to be smoothed;

[0086] Use MLS to smooth the area to be smoothed to obtain the first sub-neighborhood point cloud, the second sub-neighborhood point cloud, the third sub-neighborhood point cloud, and the fourth sub-neighborhood point cloud;

[0087] Perform data processing on the first sub-neighborhood point cloud, the second sub-neighborhood point cloud, the third sub-neighborhood point cloud, and the fourth sub-neighborhood point cloud respectively;

[0088] Add the third sub-neighborhood point cloud and the fourth sub-neighborhood point cloud after data processing to the second face point cloud.

[0089] In this embodiment, when the processor executes the computer program, the following steps are further implemented:

[0090] Perform replacement processing on the first sub-neighborhood point cloud and the second sub-neighborhood point cloud respectively;

[0091] Perform weighted summation on the third sub-neighborhood point cloud;

[0092] Replace the fourth sub-neighborhood point cloud with the fourth neighborhood point cloud.

[0093] Embodiment 5

[0094] This embodiment discloses a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0095] Perform facial feature segmentation on the color image of the face to obtain the mask of the hair region;

[0096] Project the mask of the hair region onto the original face point cloud, and extract the hair region point cloud and the second face point cloud that does not include the hair region point cloud;

[0097] Smooth the point cloud of the hair region and add the smoothed point cloud of the hair region to the second face point cloud to obtain the third face point cloud.

[0098] In this embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0099] Perform five - sense segmentation on the color image of the face to obtain the five - sense segmentation result;

[0100] Identify and extract the hair part in the five - sense segmentation result as the mask of the hair region.

[0101] In this embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0102] Define four radii of different sizes and generate neighborhood point clouds with different distances from the hair region point cloud based on the four radii of different sizes. The neighborhood point clouds include the first neighborhood point cloud, the second neighborhood point cloud, the third neighborhood point cloud, and the fourth neighborhood point cloud;

[0103] Merge the first neighborhood point cloud, the second neighborhood point cloud, the third neighborhood point cloud, and the fourth neighborhood point cloud to obtain the area to be smoothed;

[0104] Use MLS to smooth the area to be smoothed to obtain the first sub - neighborhood point cloud, the second sub - neighborhood point cloud, the third sub - neighborhood point cloud, and the fourth sub - neighborhood point cloud;

[0105] Perform data processing on the first sub - neighborhood point cloud, the second sub - neighborhood point cloud, the third sub - neighborhood point cloud, and the fourth sub - neighborhood point cloud respectively;

[0106] Add the third sub - neighborhood point cloud and the fourth sub - neighborhood point cloud after data processing to the second face point cloud.

[0107] In this embodiment, when the computer program is executed by the processor, the following steps are further implemented:

[0108] Perform replacement processing on the first sub - neighborhood point cloud and the second sub - neighborhood point cloud respectively;

[0109] Perform weighted summation on the third sub - neighborhood point cloud;

[0110] Replace the fourth sub - neighborhood point cloud with the fourth neighborhood point cloud.

[0111] Those of ordinary skill in the art can understand that all or part of the processes in the methods of the above embodiments can be completed by instructing relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above methods. Among them, any reference to a memory, database, or other medium used in the embodiments provided in the present application can include at least one of non-volatile and volatile memories. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetoresistive random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM), etc. The databases involved in the embodiments provided in the present application can include at least one of relational databases and non-relational databases. Non-relational databases can include distributed databases based on blockchain, etc., without limitation. The processors involved in the embodiments provided in the present application can be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, data processing logics based on quantum computing, etc., without limitation.

[0112] The technical features of the above embodiments can be combined arbitrarily. For the sake of concise description, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope described in this specification.

[0113] In the description of this specification, the descriptions referring to terms such as "in this embodiment" or "specifically" mean that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic expressions of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described can be combined in a suitable manner in any one or more embodiments or examples. In addition, without contradiction, those skilled in the art can combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0114] As mentioned above, it is only a preferred embodiment of the present invention, and there is no any formal limitation to the present invention. Therefore, any modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A method for processing a hair region of a three-dimensional model, characterized in that: include: Perform facial features segmentation on the color image of the face and obtain the mask of the hair area; Projecting the mask of the hair region onto the original face point cloud, extracting the hair region point cloud and the second face point cloud excluding the hair region point cloud; The hair region point cloud is smoothed, and the smoothed hair region point cloud is added to the second face point cloud to obtain a third face point cloud.

2. The method for processing the hair region of a three-dimensional model according to claim 1, characterized in that: The method of performing facial features segmentation on the color image of the face to obtain a mask of the hair area includes: Perform facial features segmentation on the color image of the face to obtain the facial features segmentation result; Identify and extract the hair part in the facial features segmentation result as a mask for the hair area.

3. The method for processing the hair region of a three-dimensional model according to claim 1, characterized in that: The step of smoothing the hair region point cloud and adding the smoothed hair region point cloud to the second face point cloud includes: Define four radii of different sizes, and generate neighborhood point clouds with different distances from the hair area point cloud based on the four radii of different sizes, the neighborhood point clouds including a first neighborhood point cloud, a second neighborhood point cloud, a third neighborhood point cloud, and a fourth neighborhood point cloud; Merge the first neighborhood point cloud, the second neighborhood point cloud, the third neighborhood point cloud and the fourth neighborhood point cloud to obtain the area to be smoothed; MLS is used to smooth the area to be smoothed, and the first sub-neighborhood point cloud, the second sub-neighborhood point cloud, the third sub-neighborhood point cloud and the fourth sub-neighborhood point cloud are obtained; Performing data processing on the first sub-neighborhood point cloud, the second sub-neighborhood point cloud, the third sub-neighborhood point cloud and the fourth sub-neighborhood point cloud respectively; The data-processed third sub-neighborhood point cloud and the fourth sub-neighborhood point cloud are added to the second face point cloud.

4. The method for processing the hair region of a three-dimensional model according to claim 1, characterized in that: The data processing of the first sub-neighborhood point cloud, the second sub-neighborhood point cloud, the third sub-neighborhood point cloud and the fourth sub-neighborhood point cloud respectively includes: Respectively perform replacement processing on the first sub-neighborhood point cloud and the second sub-neighborhood point cloud; Perform weighted summation on the third sub-neighborhood point cloud; Replace the fourth sub-neighborhood point cloud with the fourth neighborhood point cloud.

5. A three-dimensional model hair region processing device, used to implement the three-dimensional model hair region processing method according to any one of claims 1 to 4, characterized in that: include: The facial features segmentation module is used to segment the facial features of a color image and obtain a mask of the hair area; A projection module, used to project the mask of the hair area onto the original face point cloud, extract the hair area point cloud and the second face point cloud excluding the hair area point cloud; The hair region smoothing module is used to smooth the hair region point cloud, and add the smoothed hair region point cloud to the second face point cloud to obtain the third face point cloud.

6. A method for three-dimensional reconstruction of a face, characterized in that: include: Collecting facial image data, the facial image data including a color image and a depth image of the face; Using the hair region processing method of the three-dimensional model as described in any one of claims 1 to 4 to perform hair region processing on the color image of the face; The third face point cloud including the smoothed hair area point cloud is input into the three-dimensional reconstruction model to obtain the three-dimensional image data of the face.

7. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 4 are implemented.

8. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 4 are implemented.

Images