Digital human skin adjusting method, user interface system and related device

Abstract

The digital human skin adjusting method at least comprises the steps that after a user inputs a to-be-processed skin image, a texture object can be determined from noise objects in a noise layer according to the to-be-processed skin image; according to the embodiment of the invention, texture parameters in the adjustment parameters input by the user can be obtained; determining a target texture object from the texture objects according to the texture parameters; and generating a target skin image according to the target texture object and the to-be-processed skin image. Therefore, the generation process of the texture object does not need user participation, details such as the skin texture are increased, the sense of reality of the digital human is improved, and the working efficiency is improved.

Description

[0001] This application claims the priority of a Chinese patent application with the application number 2024101075009 and the application title "A Skin Simulation Method and Device Based on Digital Human Technology" filed on January 25, 2024, the entire content of which is incorporated herein by reference. Technical Field

[0002] This application relates to the field of artificial intelligence technology, and in particular, to a method for adjusting the skin of a digital human, a user interface system, and related devices. Background Art

[0003] With the development of virtual reality-related technologies, the digital human and digital human live broadcast industries have witnessed great development, bringing huge commercial demands at the same time. In the field of digital humans, the appearance of a digital human is a concrete manifestation of its expressiveness. Skin is an important part of the digital human's sense of reality, so a technical solution for adjusting the skin of a digital human is needed to improve the sense of reality of the digital human. Summary of the Invention

[0004] To solve the above problems, embodiments of this application provide a method for adjusting the skin of a digital human, a user interface system, and related devices, which realize the adjustment of the skin of a digital human, increase texture details such as texture objects, and improve the sense of reality of the digital human.

[0005] Therefore, the following technical solutions are adopted in the embodiments of this application:

[0006] In a first aspect, an embodiment of this application provides a method for adjusting the skin of a digital human, including: after a user inputs a skin image to be processed, obtaining the skin image to be processed input by the user; determining a noise object according to a noise layer; determining a texture object according to the skin image to be processed and the noise object; obtaining a texture parameter in the adjustment parameters input by the user, determining a target texture object from the texture objects according to the texture parameter, and generating a target skin image according to the target texture object and the skin image to be processed, where the target skin image includes the target texture object.

[0007] In the embodiments of the present application, multiple texture objects can be automatically generated through a code program. Furthermore, the user can adjust the texture objects combined with the skin image to be processed by setting different texture parameters. Therefore, the generation process of the texture objects does not require user participation, reducing the labor cost. The user does not need high professional skills, and the entry threshold for digital human skin texture adjustment work is low, expanding the scope of user use. The embodiments of the present application implement the adjustment work of digital human skin, adding details such as skin texture and improving the realism of digital humans. Additionally, since the link of users making skin details is avoided, the labor cost is reduced and the work efficiency is improved. Additionally, since there is no need to collect or create a preset material library, relevant costs are saved.

[0008] In a possible implementation manner, the method further includes: obtaining a plurality of preset noise sub-layers; determining a noise layer according to the plurality of noise sub-layers. The above determining a noise object according to the noise layer includes: extracting the connected regions of the noise layer to determine the noise object. Wherein, each of the plurality of noise sub-layers includes a plurality of noise sub-objects, and the plurality of noise sub-layers are layers of different scales, and the different scales indicate that the area scales of the noise sub-objects are different.

[0009] In this implementation manner, a plurality of noise sub-layers can be obtained according to a noise function. Among them, by adjusting the parameters of the noise function, the area size ranges of the texture sub-objects of different noise sub-layers are different. Therefore, the noise objects in the noise layer have different area scales, making the display effect of the skin texture closer to the real skin texture. For example, the noise layer has freckle objects with different area scales, making the display effect of the freckles closer to real freckles. Exemplarily, the texture object is obtained through a noise layer generated by a noise function, and the value range of the noise function is (0, 1]. That is to say, in the texture object, the texture value at the boundary is 0, the texture value at the center position is 1, and it gradually transitions from the boundary to the center position. Therefore, in the adjusted image, there is a natural transition between the adjusted area and the adjacent area of the skin texture, without obvious boundaries, and it is more realistic.

[0010] In another possible implementation manner, according to the skin image to be processed, the noise layer is processed to obtain a mask layer; according to a random function, the noise objects in the mask layer are processed to obtain a texture object.

[0011] In this implementation, before processing, the noise layer includes multiple noise objects. After processing the noise layer, a mask layer is obtained, and the noise objects in the skin area of the skin image to be processed are retained in the mask layer; after processing the noise objects in the mask layer, processed noise objects are obtained, and the processed noise objects are the texture objects. There are many methods to determine the texture objects from the mask layer, such as discard processing and filtering processing, with the aim of making the distribution of the texture objects in the skin image to be processed close to the texture distribution of real skin. Among them, the discard processing can be performed according to a random function. For example, in the skin area, noise objects are discarded according to a random function, so that the distribution effect of the obtained texture objects is closer to the distribution of texture objects in real skin. Taking the face image as an example of the skin image to be processed and freckles as an example of the texture objects, the random function can use a bivariate Gaussian distribution with the tip of the nose as the center point, where the central discard probability is the smallest and the probability gradually increases towards the periphery, making the distribution effect of the discarded freckles closer to the freckle distribution of a real human face. Among them, the filtering processing can be a method of filtering and screening, making the distribution of the remaining texture objects after screening close to the texture distribution of real skin.

[0012] In another possible implementation, before obtaining the adjustment parameters input by the user, the method further includes dividing the texture objects into N groups with the constraint that the difference between the sum of the areas of each group is minimized, where the sum of the areas represents the sum of the areas of the texture objects in each group.

[0013] In this implementation, the embodiments of the present application provide a method for uniformly changing the adjustment step, enabling the texture density of the texture objects to change uniformly and improving the user experience. Exemplarily, during the process of dividing multiple texture objects into N groups, with the constraint that the difference between the sum of the areas of each group is minimized, where the sum of the areas represents the sum of the areas of all the texture objects in each group. This grouping method minimizes the difference between the sum of the areas of each group, achieving that the area of the skin texture increased or decreased by each adjustment step is approximately equal. In some examples, the texture objects are also grouped according to the granularity, thereby meeting the purpose of the user to add texture objects with different granularities to the skin image to be processed and improving the user experience.

[0014] In another possible implementation, the texture parameter includes the number of texture groups K, where K is an integer greater than or equal to zero and less than or equal to N. Determining the target texture objects from the texture objects according to the texture parameter includes: determining the texture objects in K groups as the target texture objects according to the number of texture groups K.

[0015] In this implementation, the user can adjust the number or density of texture objects to achieve different display effects of skin texture. Taking the adjustment of the density of texture objects as an example for exemplary illustration. Multiple texture objects are divided into N groups, where N is an integer greater than 1. The texture parameters that the user can adjust include the texture density, and the texture density represents the number of groups among the N groups. Furthermore, according to the texture density in the texture parameters, K groups among the N groups are determined, where K is an integer greater than or equal to zero and less than or equal to N. Finally, based on the multiple texture objects in the K groups, the skin image to be processed is processed to obtain the adjusted skin image. That is to say, the larger the texture density is set, the larger the K value is, the more groups of texture objects there are, and the greater the density of the skin texture represented by the texture objects. Taking freckles as an example, the larger the texture density of freckles is set, the more groups of freckles there are, and in the presentation effect of skin texture, the density of freckles is greater. Generally, taking freckle features as an example, the real freckles on a real human face not only have differences in area size (achieved by the characteristics of the noise function), but also the distribution density of real freckles on the faces of different people or the same person at different times is different. That is to say, by setting the number or density of texture objects, different distribution effects of texture objects can be obtained, so as to distinguish the faces of different people or the same person at different times, making the skin texture of the digital human skin closer to the skin texture of real skin, presenting the effect of "thousands of people with thousands of faces".

[0016] In another possible implementation, the method further includes: decomposing the pigments of the image to be processed to obtain basic pigment information.

[0017] In this implementation, since the basic pigment information has practical physical significance, it can make the skin image of the digital human after pigment adjustment close to the real skin of a natural person, thus having a high sense of reality.

[0018] In another possible implementation, the adjustment parameter further includes a pigment parameter, and the method further includes: determining target pigment information according to the pigment parameter and the basic pigment information. Further, based on the target texture object, the target pigment information, and the skin image to be processed, a target skin image is generated.

[0019] In this implementation, since the basic pigment information has practical physical significance, it can make the skin image of the digital human after adjustment close to the real skin of a natural person, thus having a high sense of reality. Compared with only adding skin texture through color, the method of the embodiment of the present application can adjust the basic pigment information in the skin image to be processed. No matter how the user adjusts the pigment information, the image after color space reconstruction belongs to the normal skin color range, solving the problems of skin color distortion and uneven edges during color space editing, and improving the sense of reality of the adjusted image. The pigments of the adjusted image can be adjusted, improving the user experience.

[0020] In another possible implementation, the basic pigment information includes basic melanin information, the target pigment information includes target melanin information, and the pigment parameters include melanin parameters. Based on the pigment parameters and the basic pigment information, the target pigment information is determined. One possible implementation may be: determining the target melanin information based on the melanin parameters and the basic melanin information; and / or, the basic pigment information includes basic hemoglobin information, the target pigment information includes target hemoglobin information, and the pigment parameters include hemoglobin parameters. Based on the pigment parameters and the basic pigment information, the target pigment information is determined. Another possible implementation may be: determining the target hemoglobin information based on the hemoglobin parameters and the hemoglobin information.

[0021] In this implementation, the diffuse reflection map of the color space is decomposed into melanin information and hemoglobin information with physical meanings. Users can edit skin pigments, such as skin color and patterns, by adjusting the melanin information and hemoglobin information. This can solve the problem of poor realism caused by only adjusting at the color level, making the skin of the digital human more realistic. Additionally, after the user adjusts the melanin information and hemoglobin information, since the melanin information and hemoglobin information have physical meanings, no matter how the user adjusts the melanin information and hemoglobin information, the image reconstructed in the color space belongs to the normal skin color range, solving the problem of skin color distortion and improving the user experience. When adjusting the color space, it is easy for users to adjust the skin color to a color beyond the normal skin color range. For example, yellow skin is easily adjusted to orange skin close to yellow, resulting in the problem of skin color distortion.

[0022] In another possible implementation, the method further includes: dividing the image to be processed into at least one image region. The adjustment parameters further include region parameters, and the method further includes: determining the region to be processed from the at least one image region according to the region parameters. Further,

[0023] Determining the target texture object from the texture objects in the region to be processed according to the texture parameters.

[0024] In this implementation, the skin image to be processed input by the user is processed in regions. The user can select the image regions that need to be edited, achieving the purpose of region-based adjustment of skin details (such as skin texture, skin pigment, etc.), rather than the overall editing method, making the adjustment region more flexible and improving the user experience.

[0025] In a second aspect, an embodiment of the present application provides an adjustment device for a digital human skin, including: a first acquisition module, configured to acquire a skin image to be processed input by a user; a noise object determination module, configured to determine a noise object according to a noise layer, where the noise object is a connected region of the noise layer; a texture object determination module, configured to determine a texture object according to the skin image to be processed and the noise object, where the texture object is used to indicate the skin texture in the skin image to be processed; a second acquisition module, configured to acquire adjustment parameters input by the user, where the adjustment parameters include texture parameters; a target texture object determination module, configured to determine a target texture object from the texture objects according to the texture parameters; and a target skin image generation module, configured to generate a target skin image according to the target texture object and the skin image to be processed, where the target skin image includes the target texture object.

[0026] In a possible implementation manner, the noise object determination module is specifically configured to: acquire a plurality of preset noise sub-layers, each noise sub-layer including a plurality of noise sub-objects, where the plurality of noise sub-layers are layers of different scales, and different scales indicate different area scales of the noise sub-objects; determine the noise layer according to the plurality of noise sub-layers; and extract the connected domain of the noise layer to obtain the noise object.

[0027] In another possible implementation manner, the texture object determination module is specifically configured to: process the noise layer according to the skin image to be processed to obtain a mask layer; and process the noise object in the mask layer according to a random function to obtain the texture object.

[0028] In another possible implementation manner, the device further includes: a texture object grouping module, configured to divide the texture objects into N groups with the constraint that the difference in the sum of the areas between each group is the smallest, where the sum of the areas represents the sum of the areas of the texture objects in each group.

[0029] In another possible implementation manner, the texture parameters include the number of texture groups K, where K is an integer greater than or equal to zero and less than or equal to N. The target texture object determination module is specifically configured to: determine the texture objects in the K groups as the target texture objects according to the number of texture groups K.

[0030] In another possible implementation manner, the device further includes: a basic pigment information determination module, configured to decompose the pigment of the skin image to be processed to determine the basic pigment information.

[0031] In another possible implementation manner, the adjustment parameters further include pigment parameters. The device further includes: a target pigment information determination module, configured to determine the target pigment information according to the pigment parameters and the basic pigment information. The target skin image generation module is specifically configured to: generate a target skin image according to the target texture object, the target pigment information, and the skin image to be processed.

[0032] In another possible implementation, the basic pigment information includes basic melanin information, the target pigment information includes target melanin information, and the pigment parameter includes a melanin parameter; the target pigment information determination module is specifically configured to: determine the target melanin information according to the melanin parameter and the basic melanin information. And / or, the basic pigment information includes basic hemoglobin information, the target pigment information includes target hemoglobin information, and the pigment parameter includes a hemoglobin parameter; the target pigment information determination module is specifically configured to: determine the target hemoglobin information according to the hemoglobin parameter and the hemoglobin information.

[0033] In another possible implementation, the device further includes: an image region division module, configured to divide the skin image to be processed into at least one image region. The adjustment parameter further includes a region parameter, and the device further includes: a region to be processed determination module, configured to determine the region to be processed from the at least one image region according to the region parameter. The target texture object determination module is specifically configured to: determine the target texture object from the texture objects in the region to be processed according to the texture parameter.

[0034] In a third aspect, an embodiment of the present application provides a user interface system, including: a client, configured to receive the skin image to be processed and the adjustment parameter input by the user; a server, configured to obtain the skin image to be processed and the adjustment parameter in the client, and execute the algorithm function embodied by the method of any one of the above or the device of any one of the above. Among them, the editing result is fed back to the user in real time through the user interface system, so that the user can perform further skin image adjustment work according to the fed-back editing result, improving the user experience.

[0035] In a fourth aspect, an embodiment of the present application provides a computing device cluster, including at least one computing device. Each of the at least one computing device includes a memory and a processor, and instructions are stored in the memory. When the instructions are executed by the processor, the functions embodied by the method of any one of the above or the device of any one of the above are implemented.

[0036] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by the processor, the functions embodied by the method of any one of the above or the device of any one of the above are implemented.

[0037] In a sixth aspect, an embodiment of the present application provides a computer program product, which includes program instructions. When the program instructions are executed by a computer, the computer is caused to execute the functions embodied by the method of any one of the above or the device of any one of the above. Description of the Drawings

[0038] The drawings required for use in the following description of the embodiments or technologies are briefly introduced below.

[0039] Figure 1 It is a schematic diagram of the composition of a digital human system architecture provided in an embodiment of the present application;

[0040] Figure 2 It is a schematic flowchart of a method for adjusting a digital human skin provided in an embodiment of the present application;

[0041] Figure 3 It is a schematic diagram of a noise sub-layer and a noise layer provided in an embodiment of the present application;

[0042] Figure 4 It is a schematic diagram of the noise layer before and after masking and discarding processing provided in an embodiment of the present application;

[0043] Figure 5 It is a schematic flowchart of a process for grouping multiple texture objects provided in an embodiment of the present application;

[0044] Figure 6 It is a schematic flowchart of a method for adjusting a freckle texture provided in an embodiment of the present application;

[0045] Figure 7 It is a schematic diagram of the composition of a device for adjusting a digital human skin provided in an embodiment of the present application;

[0046] Figure 8 It is a schematic diagram of the composition of a user interface system provided in an embodiment of the present application;

[0047] Figure 9 It is a schematic diagram of the structure of a computing device provided in an embodiment of the present application;

[0048] Figure 10 It is a schematic diagram of the structure of a computing device cluster provided in an embodiment of the present application;

[0049] Figure 11 It is a schematic diagram of the structure of another computing device cluster provided in an embodiment of the present application. Detailed implementation manners

[0050] In this article, the term "and / or" is an association relationship describing associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In this article, the symbol " / " represents that the associated objects are in an "or" relationship, for example, A / B represents A or B.

[0051] In the description and claims of this article, terms such as "first" and "second" are used to distinguish different objects, rather than to describe a specific order of the objects. For example, the first response message and the second response message, etc., are used to distinguish different response messages, rather than to describe a specific order of the response messages.

[0052] In the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design solution described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or having more advantages than other embodiments or design solutions. Rather, the use of words such as "exemplary" or "for example" is intended to present relevant concepts in a specific manner.

[0053] In the description of the embodiments of this application, unless otherwise specified, the meaning of "a plurality of" refers to two or more. For example, a plurality of processing units refers to two or more processing units, etc.; a plurality of elements refers to two or more elements, etc.

[0054] To facilitate understanding of the solution provided by the embodiments of this application, some terms involved in this solution are briefly introduced first.

[0055] Digital human: Also known as virtual digital human, it is a comprehensive product created and used by computer means such as computer graphics, graphics rendering, motion capture, deep learning, and speech synthesis, and has multiple human characteristics (such as appearance characteristics, human performance ability, human interaction ability, etc.). Digital humans can not only perform certain actions or expressions instead of real people, but also imitate real people through means such as sound, image, and video. Representative sub-applications include virtual assistants, virtual customer service, virtual idols, virtual anchors, etc.

[0056] Mask processing: Use a selected image, graphic, or object to block the processed image to control the area or process of image processing.

[0057] Texture object: In real skin, generally, there are many texture details on the skin, such as freckles, tattoos, eye makeup, spots, scars, etc. In a skin image, if only the skin itself exists without texture details, it does not conform to real skin, resulting in a poor sense of reality of the skin image. Therefore, adding texture details to the skin image can make the skin in the skin image closer to real skin and improve the sense of reality of the skin in the skin image. A texture object is the smallest unit that can represent texture details in a skin image. For example, the texture object in a freckle refers to an independent freckle or multiple connected freckles, and the texture object in a tattoo refers to a single connected stripe, etc.

[0058] Noise layer: It refers to a layer obtained according to a noise function or a noise function. The layer mentioned in this application does not refer to the layer in the image processing method, but refers to the data distribution result related to the image position. For example, a noise layer generated according to Poisson noise refers to a data distribution result with a value range of (0, 1] set at different positions of the image, and gradually changing from 1 in the center to 0 around. This distributed data can be fused with the skin image to be processed to obtain the processed skin image.

[0059] Noise object: It refers to an independent connected domain in the noise layer. If there are multiple connected domains in the noise layer, it means there are multiple noise objects in the noise layer. For example, for a noise layer generated according to Poisson noise, its noise object refers to a certain connected area that is not 0.

[0060] Pigment: The color of real skin is mainly determined by the pigments in it. The main pigments in real skin are melanin and hemoglobin. Melanin is distributed at different depths of the skin and is the main component of skin color. Hemoglobin appears in the vascular structure of the dermal papilla in anaerobic and aerobic forms, constituting the red color of the skin. Therefore, the pigments in the skin have physical significance. If relevant features of pigments are added to the skin image, the skin in the skin image can be made closer to real skin, and the realism of the skin in the skin image can be improved. Optionally, the skin texture map to be processed can be decomposed through a pigment decomposition algorithm to obtain the basic pigment information of the skin texture map to be processed, such as the melanin component and the hemoglobin component. Exemplarily, a machine learning algorithm combining PCA (Principal Component Analysis) and ICA (Independent Component Analysis) can be used to perform pigment decomposition on the skin image to be processed. It is also possible to use the color and pigment components as a pair of training data samples, and through supervised learning, an encoder can be obtained. Through the trained encoder for pigment decomposition, the skin image to be processed with color representation is used as the input of the encoder, and the basic pigment information of the skin image to be processed is output.

[0061] Random function: It refers to a function that generates random numbers. The random effect brought by the random function is more in line with the real situation in reality. For example, using a random function to process the skin image to be processed can make the skin in the processed skin image closer to real skin.

[0062] Discarding process: A data processing method that can make the remaining objects conform to a certain distribution law by discarding multiple objects. For example, assuming a noise layer containing multiple noise objects, performing a discarding process on the noise objects in the noise layer based on a random function of a binary Gaussian distribution can make the distribution of the noise objects in the noise layer after the discarding process conform to the binary Gaussian distribution.

[0063] As people's demand for digital life continues to increase, virtual digital people have also ushered in a broad market prospect. The appearance of digital people is the specific presentation of their expressiveness. And the skin is an important part of the digital people's realism. Therefore, a technical solution for adjusting the skin of digital people is needed to improve the realism of digital people.

[0064] Currently, the method of editing the color space of the skin image by image processing method only considers the transformation of the color space. Therefore, the adjusted skin image lacks realism, and there is an obvious boundary between the adjusted area and the adjacent area in the adjusted image, and the realism is poor. In addition, according to the preset skin material library, the skin details of the skin material library are fused with the original skin image by image fusion. Since it is necessary to collect or create a preset material library, the cost is high. In addition, the robustness of this solution is insufficient, and users cannot customize the color, size and distribution of skin details.

[0065] In order to solve the technical problems existing in the above scheme, the embodiment of the present application provides another solution. According to the skin image to be processed input by the user, skin details are generated, and the user can directly merge the skin details into the skin image to be processed. Optionally, the skin details include a texture object representing the skin texture, basic pigment information representing the skin pigment, etc. The embodiment of the present application adds skin details to the skin image to be processed, realizes the adjustment of the digital human skin, improves the realism of the skin in the skin image to be processed, thereby improving the realism of the digital human and improving the user experience.

[0066] In the solution of the embodiment of the present application, the generation process of skin details is completed by relevant software. After the user inputs the skin image to be processed, the generated skin details can be directly obtained, so the user does not need to make the skin details, which lowers the entry threshold for skin image adjustment work and expands the user range for skin image adjustment work. Among them, the present application does not limit the relevant software, and the relevant software can be a management platform or functional module running on a computing device. In addition, since the link of users making skin details is avoided, labor costs are reduced and work efficiency is improved. In some implementations, the editing results can also be fed back to the user in real time through the user interface system, so that the user can perform further skin image adjustment work based on the fed-back editing results, thereby improving the user experience.

[0067] In some implementation manners of the embodiments of the present application, by performing pigment decomposition on the to-be-processed skin image input by the user, basic pigment information such as melanin information and hemoglobin information is obtained; the user can edit the adjustment effect of the skin image and skin details at the pigment level by adjusting the basic pigment information. Since the basic pigment information has actual physical meanings, the adjusted skin image of the digital human can be made close to the real skin of a natural person, and thus has a high sense of reality. That is to say, the problem of poor sense of reality when adjusting the skin image at the color level is solved. In some implementation manners of the embodiments of the present application, the skin texture is related to a noise function, and the value range of the noise function is (0, 1]. That is to say, in the skin texture, the texture value at the boundary is 0, the texture value at the center position is 1, and it gradually transitions from the boundary to the center position. Therefore, in the adjusted image, there is a natural transition between the adjusted area and the adjacent area of the skin texture, and there is no obvious boundary, which is more realistic.

[0068] In the solution of the embodiments of the present application, the generation process of the skin details is completed by a computing device. After the user inputs the to-be-processed skin image, the generated skin details can be directly obtained. Therefore, there is no need to collect or create a preset material library, which reduces the cost. In some implementation manners of the embodiments of the present application, the user can also input adjustment parameters, and the computing device adjusts the color, size, and distribution of the skin details of the skin image according to the adjustment parameters input by the user. For example, in some implementation manners, by performing bucketing processing on multiple texture objects, the user can select the number of buckets of the texture objects, so as to fuse the texture objects in the selected number of buckets with the to-be-processed skin image and output the adjusted image, solving the problem of adjusting the density of the skin texture.

[0069] In order to make the objectives, technical solutions, and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described below with reference to the accompanying drawings.

[0070] See Figure 1 shown in Figure 1 is a schematic diagram of the composition of a digital human system architecture provided by the embodiments of the present application. In the embodiments of the present application, the digital human system architecture 100 may include a terminal 110 and a server 120. Among them, the server 120 may include one or more servers ( Figure 1 taking including one server as an example for illustration), and the server 120 may provide the methods or devices provided by the embodiments of the present application for one or more terminals.

[0071] Among them, an application related to digital human skin adjustment can be installed on the terminal 110. The above application or web page can provide an interface. The terminal 110 can receive the skin image to be processed, adjustment parameters, etc. input by the user on the interface, and send the above adjustment parameters to the server 120. The server 120 can adjust the skin image to be processed based on the received adjustment parameters to obtain processing results such as the target skin image, and return the processing results to the terminal 110 to feedback the processing results to the user.

[0072] It should be understood that in some alternative implementations, the terminal 110 can also complete the actions of obtaining processing results such as the target skin image based on the received adjustment parameters by itself, without the need for the cooperation of the server. The embodiments of the present application do not limit this.

[0073] Next, describe Figure 1 the product form of the terminal 110. The terminal 110 in the embodiments of the present application can be a mobile phone, a tablet computer, a wearable device, a vehicle-mounted device, an augmented reality (AR) / virtual reality (VR) device, a laptop computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), etc. The embodiments of the present application do not make any restrictions on this.

[0074] Next, describe Figure 1 the product form of the server 120. It can be further understood that the server 120 can be various servers, such as servers with an X86 architecture. Specifically, it can be a whole cabinet server, a blade server, a high-density server, a rack server, or a high-performance server, etc. In other words, the embodiments of the present application do not specifically limit the specific categories of the server. Further, it can be understood that Figure 1 the structure of the server shown does not constitute a limitation on the structure of the server. The server may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0075] See Figure 2 as shown Figure 2 is a schematic flowchart of a method for adjusting the digital human skin provided by the embodiments of the present application. As Figure 2 shown, the embodiments of the present application provide a method for adjusting the digital human skin, which mainly includes the following steps:

[0076] Step S201, obtain the skin image to be processed input by the user.

[0077] Step S202: Determine a noise object based on the noise layer, where the noise object is a connected region of the noise layer.

[0078] Step S203: Determine a texture object based on the skin image to be processed and the noise object, where the texture object is used to indicate the skin texture in the skin image to be processed.

[0079] Step S204: Obtain the adjustment parameters input by the user, where the adjustment parameters include texture parameters.

[0080] Step S205: Determine a target texture object from the texture object according to the texture parameters.

[0081] Step S206: Generate a target skin image according to the target texture object and the skin image to be processed, where the target skin image includes the target texture object.

[0082] Optionally, the method at least includes a preprocessing stage and an adjustment stage. The preprocessing stage includes Step S201, Step S202, and Step S203, and the adjustment stage includes Step S204, Step S205, and Step S206. Dividing the method into a preprocessing stage and an adjustment stage can more clearly describe the functions of the above method in different stages, but does not constitute a limitation on the above steps. Exemplarily, after the user inputs the skin image to be processed, the steps of the preprocessing stage can be executed once to obtain a texture object. The user can adjust the skin image to be processed multiple times through texture parameters to obtain multiple target skin images until the user obtains a satisfactory target skin image. That is to say, for the same skin image to be processed, the steps of the preprocessing stage can be executed once, and the steps of the adjustment stage can be executed multiple times.

[0083] In the embodiments of the present application, Figure 1 Taking the digital human system architecture 100 shown as an example, an exemplary description is given. The terminal 110 may include an interface or a web page for digital human skin adjustment, and the user inputs the skin image to be processed and adjustment parameters in the digital human skin image. The server 120 determines a texture object indicating the skin texture according to the skin image to be processed and the noise object in the noise layer. Determine a target texture object from the texture object according to the adjustment parameters input by the user; generate a target skin image according to the target texture object and the skin image to be processed. Through the interface or web page of the terminal 110, the adjusted target skin image is output to the user.

[0084] Optionally, the embodiments of the present application do not limit the content input by the user. The user can input a digital human or a digital human image. After the server 120 performs a masking process on the digital human or the digital human image, a skin image of the digital human is obtained, and this skin image can be confirmed as the skin image to be processed. Optionally, the terminal 110 can also display the skin image of the digital human after the masking process to the user, and the user determines the area to be processed in the skin image, thereby obtaining the skin image to be processed. Alternatively, the user can also directly input the skin image of the digital human, and the server 120 thus uses this skin image as the skin image to be processed.

[0085] Optionally, for determining the noise object according to the noise layer, a specific implementation manner may be: obtaining a plurality of preset noise sub-layers; determining the noise layer according to the plurality of noise sub-layers; performing connected component extraction on the noise layer to determine the noise object. Wherein, each of the noise sub-layers includes a plurality of noise sub-objects, and the plurality of noise sub-layers are layers of different scales, and the different scales indicate that the area scales of the noise sub-objects are different.

[0086] Furthermore, a plurality of noise sub-layers can be obtained according to the noise function. Among them, by adjusting the parameters of the noise function, the area size ranges of the texture sub-objects of different noise sub-layers are different. Therefore, the noise objects in the noise layer have different area scales, so that the display effect of the skin texture is closer to the real skin texture.

[0087] See Figure 3 as shown Figure 3 is a schematic diagram of the noise sub-layer and the noise layer provided by the embodiments of the present application. As Figure 3 shown, by setting different parameters of the noise function, a plurality of noise sub-layers of different scales are correspondingly generated. That is to say, each noise sub-layer includes a plurality of texture sub-objects, and the scales of the texture sub-objects in different noise sub-layers are different. The scale of the texture sub-object can be understood as the area of the connected component in the noise sub-layer. Exemplarily, as Figure 3 shown, the scale of the texture sub-object in the noise sub-layer a is tiny; the scale of the texture sub-object in the noise sub-layer b is small; the scale of the texture sub-object in the noise sub-layer c is medium; the scale of the texture sub-object in the noise sub-layer d is large.

[0088] In one example, as Figure 3 shown, by fusing a plurality of noise sub-layers, a noise layer is obtained, which can make the noise objects in the noise layer have multiple area scales, so that the display effect of the skin texture is closer to the real skin texture. For example, the noise layer has a plurality of area scales of freckle objects, so that the display effect of the freckles is closer to the real freckles.

[0089] Optionally, with regard to determining a texture object based on the skin image to be processed and the noise object, a specific implementation manner may be: processing the noise layer based on the skin image to be processed to obtain a mask layer; processing the noise object in the mask layer according to a random function to obtain a texture object.

[0090] Optionally, with regard to processing the noise layer to obtain a mask layer, a possible implementation manner is mask processing, and it can also be methods such as filtering and screening processing, discarding processing, etc. With regard to processing the mask layer to obtain a texture object, a possible implementation manner is discarding processing, and it can also be filtering processing or screening processing according to preset rules. For example, by setting an area threshold, filtering or screening the noise objects in the mask layer.

[0091] See Figure 4 as shown in Figure 4 FIG. is a schematic diagram of the noise layer provided in the embodiment of the present application before and after mask and discard processing. As Figure 4 shown, taking the mask method as an example, processing the noise layer to obtain a mask layer; taking the discard processing as an example, processing the mask layer to obtain a texture object. Before the mask and discard processing, the noise layer includes multiple noise objects. During the mask processing, the noise objects in the skin area of the skin image to be processed are retained to obtain a mask layer; and for the retained texture objects, discard processing is performed according to a random function to obtain a texture object. Therefore, after the mask and discard processing, a texture layer is obtained, and the texture layer includes several texture objects. Among them, for the sake of convenience in distinction, after the mask and discard processing, the noise layer is called a texture layer, and after the mask and discard processing, the retained noise objects are called texture objects. As Figure 4 shown, a noise object represents a certain connected domain in the noise layer, and a texture object represents a certain connected domain in the texture layer.

[0092] As Figure 4 shown, taking the skin image to be processed as a face image as an example, and taking freckles as an example for the noise object and the texture object. In the skin area of the face image, at least one of the multiple noise objects is discarded according to a random function, so that the distribution effect of the discarded freckles is closer to the freckle distribution of a real face. As Figure 4 shown, the purpose of the mask processing of the noise layer and the prefabricated face image mask is to mask process or discard process all the noise objects in the non-skin area of the face image, so that only texture objects are retained in the skin area.

[0093] Optionally, the random function includes a two-dimensional Gaussian distribution function centered at a preset position, and the preset position includes a position in the skin image to be processed. Taking a face image as an example, as Figure 4As shown in the texture layer on the right, the random function can use a bivariate Gaussian distribution centered on the tip of the nose. The bivariate Gaussian distribution has the characteristics of the smallest central discard probability and gradually increasing probability towards the periphery. Therefore, the freckle distribution effect after discarding is closer to the freckle distribution of a real human face.

[0094] Generally, taking freckle features as an example, the real freckles on a real human face not only have different area sizes, but also have different distribution densities of real freckles on the faces of different people or on the face of the same person at different times. That is to say, by setting the number or density of texture objects, different distribution effects of texture objects can be obtained, so as to distinguish the faces of different people or the faces of the same person at different times, making the skin texture of the digital human skin closer to the skin texture of real skin, presenting the effect of "thousands of people with thousands of faces".

[0095] Optionally, the method further includes: dividing the texture object into N groups, where N is an integer greater than 1.

[0096] Optionally, further, the texture parameter includes the number of texture groups K, where K is an integer greater than or equal to zero and less than or equal to N. Among them, according to the texture parameter, a target texture object is determined from the texture objects. A possible implementation method may be: according to the texture group number K parameter, determine the group number K, where K is an integer greater than or equal to zero and less than or equal to N; determine the texture objects in the K groups as the target texture objects.

[0097] That is to say, the user can adjust the number or density of texture objects, so that the number of skin textures in the skin image to be processed is different. Taking the adjustment of the density of texture objects as an example for exemplary illustration. Multiple texture objects are divided into N groups, where N is an integer greater than 1. The texture parameters that the user can adjust include the texture density, and the texture density represents the number of groups in the N groups. Furthermore, according to the texture density in the texture parameters, K groups in the N groups are determined, where K is an integer greater than or equal to zero and less than or equal to N. Finally, the multiple texture objects in the K groups are fused with the skin image to be processed to obtain the adjusted skin image. That is to say, the larger the texture density is set, the larger the K value is, the more groups of texture objects are, and the greater the density of the skin texture represented by the texture objects is. Taking freckles as an example, the larger the texture density of freckles is set, the more groups of freckles are, and in the presentation effect of the skin texture, the greater the freckle density is.

[0098] Optionally, a possible implementation method for dividing the texture object into N groups may be: taking the minimum difference in the sum of areas between each group as a constraint condition, dividing the texture object into N groups, and the sum of areas represents the sum of the areas of the texture objects in each group.

[0099] That is to say, the embodiment of the present application provides a method for uniformly changing the adjustment step length, enabling the texture density of the texture object to change uniformly and improving the user experience. Exemplarily, when dividing multiple texture objects into N groups, the constraint condition is that the difference between the area sums of each group is minimized, where the area sum represents the sum of the areas of all texture objects in each group. This grouping method minimizes the difference between the area sums of each group, achieving that the areas of the skin textures increased or decreased by each adjustment step are approximately equal. The following will be further described with reference to Figure 5 for exemplary purposes.

[0100] Refer to Figure 5 as shown in Figure 5 which is a schematic flowchart of the grouping process for multiple texture objects provided by the embodiment of the present application. As shown in Figure 5 the method for grouping multiple texture objects provided by the embodiment of the present application mainly includes the following steps:

[0101] S501, sort the multiple texture objects in descending order of area. This step is an optional step because when grouping all texture objects, a traversal method can be used, so area sorting may not be necessary and all texture objects can still be accessed. The beneficial effect of sorting the multiple texture objects in descending order of area is to preferentially group the texture objects with larger areas, reducing the number of traversals and improving the grouping efficiency.

[0102] S502, calculate the average area avg of each group. The average area of each group is the ratio of the total area sum S all of all texture objects in the texture layer to the number of divided groups N. The average area avg of each group is equivalent to the sum of the areas of the texture objects in each group on average. The calculation formula for the average area avg of each group is shown in formula (1).

[0103] avg = S all / N (1)

[0104] It should be further noted that since each texture object cannot be divided, there is a certain difference between the total area sum of the texture objects in each group after division and the average area avg of each group. The smaller this difference is, the smaller the difference between the adjustment step lengths of the texture density. During the user adjustment process, the change process of the skin texture becomes more uniform, and the user experience is better. Therefore, minimizing this difference can be used as a grouping constraint condition or an optimization object.

[0105] Exemplarily, as shown in formula (2), the difference between the area sums of the texture objects between each group can be used as a constraint condition. In formula (2), S i represents the sum of the areas of all texture objects in the i-th group, and S jDenote the sum of the areas of all texture objects in the j-th group. By summing the absolute values of the differences in the sums of areas between any two groups, minimizing the sum can be used as the optimization objective for grouping.

[0106]

[0107] S503, calculate the remaining area area of the i-th group. Each group may include multiple texture objects. During the process of dividing multiple texture objects, it is necessary to update the remaining area area of the i-th group and match suitable texture objects according to the updated remaining area area. The calculation formula for the remaining area area is shown in formula (3), S h Denote the sum of the areas of the multiple texture objects that have been divided in the i-th group.

[0108] area = avg - S h (3)

[0109] S504, traverse the remaining multiple texture objects. The purpose of traversing is to find one or more texture objects that match the remaining area area of the i-th group.

[0110] S505, determine whether the area of the texture object is greater than or equal to avg. That is, compare the area of the current texture object with the average area avg of each group. When the area of the texture object is greater than or equal to avg, execute step S506; when the area of the texture object is less than avg, execute step S507.

[0111] S506, divide it into the i-th group, i = i + 1. When the area of the current texture object is greater than or equal to the average area of each group, the current texture object forms an independent group to avoid the sum of the areas of multiple texture objects in the current group being too large. i = i + 1 means that the number of groups increases by 1. After this step is executed, execute step S502 to recalculate the average area avg of each group so that the remaining multiple texture objects can be evenly grouped.

[0112] S507, divide it into the i-th group and update the remaining area area of the i-th group. Since the area of the texture object is less than avg, after dividing the current texture object into the i-th group, the remaining area area decreases and is greater than zero, and it is necessary to update the remaining area area of the i-th group.

[0113] S508, determine whether there is a texture object with an area equal to area. When there is a texture object with an area equal to area, execute step S509; when there is no texture object with an area equal to area, execute step S510.

[0114] S509, Divide it into the i-th group, where i = i + 1. Since there is a texture object with an area equal to area, after dividing the current texture object into the i-th group, the remaining area area is equal to zero, and the division of the current group is completed. That is to say, after this step is executed, step S502 is executed.

[0115] S510, Divide the texture object closest to area into the i-th group.

[0116] That is to say, when the area of the current texture object is less than the average area avg of each group, match the remaining area area of each group with the remaining texture objects to determine the group to which the current texture object belongs. When there is no texture object with an area equal to area, find the texture object closest to area and divide it into the i-th group.

[0117] S511, Determine whether area is greater than the minimum threshold. When area is greater than the minimum threshold, execute step S512. Exemplarily, the minimum threshold can be zero or other values. Explain the reason for setting the minimum threshold. When the remaining area area of the i-th group is less than or equal to the minimum threshold, there may be no texture object that can be placed in the i-th group, that is, the grouping of the i-th group is completed, and it is necessary to execute step S512 currently to perform the grouping work of the (i + 1)-th group.

[0118] S512, i = i + 1, indicating that the current grouping is completed and the grouping work of the next group starts.

[0119] It should be further noted that the user can adjust the display density of the skin texture by adjusting the number of display groups of the texture objects, so as to adjust the skin texture and achieve the display effect of "unique for each person". During the adjustment process by the user, the difference between the adjustment step sizes of the texture density is very small. During the adjustment process by the user, the change process of the skin texture becomes more uniform, and it is easier for the user to find the required texture density, providing a good user experience.

[0120] Further, after grouping multiple texture objects according to the grouping method as Figure 5 shown, if arranged in the front and back according to the group number, since the multiple texture objects are arranged in descending order of area, in the groups with a smaller group number, the area of a single texture object is larger, that is, the granularity of the texture object is larger; in the groups with a larger group number, the area of a single texture object is smaller, that is, the granularity of the texture object is smaller.

[0121] In one example, assume that the user selects the texture density as K. Then, fuse the texture objects of the first K groups with the skin image to be processed to obtain the adjusted image. Therefore, when the value of K changes from small to large, the granularity of the texture object changes from large to small; when the value of K changes from large to small, the granularity of the texture object changes from small to large.

[0122] In another example, assume that the user selects a texture density of K, and can select the positions of K groups of texture objects. Then, the K groups of texture objects at the corresponding positions are fused with the skin image to be processed to obtain an adjusted image. Thus, when the user selects the positions of the K groups of texture objects to be forward, an image with larger-granularity texture objects after fusion can be obtained; when the user selects the positions of the K groups of texture objects to be in the middle, an image with medium-granularity texture objects after fusion can be obtained; when the user selects the positions of the K groups of texture objects to be backward, an image with smaller-granularity texture objects after fusion can be obtained. That is to say, the texture objects can be grouped according to the granularity, so as to meet the purpose of the user adding texture objects with different granularities to the skin image to be processed, and the user experience is improved.

[0123] Optionally, it further includes: performing pigment decomposition on the image to be processed to determine basic pigment information.

[0124] Optionally, the adjustment parameter further includes a pigment parameter, and the method further includes: determining target pigment information according to the pigment parameter and the basic pigment information. A possible implementation manner of step S206 may be: generating a target skin image according to the target texture object, the target pigment information, and the skin image to be processed.

[0125] In one example, the basic pigment information includes basic melanin information, the target pigment information includes target melanin information, and the pigment parameter includes a melanin parameter. Determining the target pigment information according to the pigment parameter and the basic pigment information, a possible implementation manner may be: determining the target melanin information according to the melanin parameter and the basic melanin information.

[0126] In another example, the basic pigment information includes basic hemoglobin information, the target pigment information includes target hemoglobin information, and the pigment parameter includes a hemoglobin parameter. Determining the target pigment information according to the pigment parameter and the basic pigment information, another possible implementation manner may be: determining the target hemoglobin information according to the hemoglobin parameter and the hemoglobin information.

[0127] Furthermore, the basic pigment information has physical significance. Therefore, the skin rendering effect of the generated target skin image is closer to that of real skin, making the digital human more realistic.

[0128] Furthermore, after the user edits the pigment information of the skin image, the color space can be reconstructed according to the edited pigment information to obtain a skin image with edited skin details. Compared with editing from a pure color space, the pigment information editing in the embodiments of the present application can make the image reconstructed from the color space free of skin color distortion problems.

[0129] Exemplarily, after the user adjusts the melanin information and hemoglobin information, since the melanin information and hemoglobin information have physical meanings, no matter how the user adjusts the melanin information and hemoglobin information, the image after color space reconstruction belongs to the normal skin color range, solving the problem of skin color distortion and improving the user experience. When adjusting the color space, it is easy for the user to adjust the skin color to a color beyond the normal skin color range. For example, yellow skin is easily adjusted to orange skin close to yellow, resulting in the problem of skin color distortion.

[0130] Optionally, the method further includes: dividing the image to be processed into at least one image region. The adjustment parameter further includes a region parameter, and the method further includes: determining a region to be processed from the at least one image region according to the region parameter. A possible implementation manner of step S205 may be: determining a target texture object from the texture objects in the region to be processed according to the texture parameter.

[0131] Optionally, the related work of dividing the image region can be automatically completed by a computing device according to the image to be processed, or can be determined by the user customarily. Or use the image regions divided by the computing device as the default setting or preferred setting; at the same time, there is a module for user customization, which is used for the user to customarily determine the division of the image region.

[0132] Furthermore, by performing sub-region processing on the skin image to be processed input by the user, the user can select the image region to be edited, achieving the purpose of sub-region adjustment of skin details (such as skin texture, skin pigment, etc.), rather than the overall editing method, making the adjustment region more flexible and improving the user experience.

[0133] See Figure 6 as shown Figure 6 is a schematic flowchart of a method for adjusting freckle texture provided by an embodiment of the present application. As Figure 6 shown, in order to further illustrate how to generate a target skin image including a target texture object, an embodiment of the present application takes freckle texture as an example to illustrate the texture object. As Figure 6 shown, a method for adjusting freckle texture mainly includes the following steps:

[0134] Step S601, generating a noise layer through a selected noise function.

[0135] It should be further noted that the noise function is related to the pre-added skin texture. For example, when the skin texture is freckles, the Poisson noise can be selected as the noise function. The value range of the freckle texture generated by the Poisson noise is (0, 1], and it shows a gradient form from the center to the periphery. This form is brought by the Poisson noise and is close to the display effect of the freckle texture, making the display effect of the freckle texture more realistic. Another example is that when the skin texture is wrinkles, the eddy current noise can be selected as the noise function.

[0136] Optionally, multiple noise sub-layers of different scales are generated through the selected Poisson noise, and a noise layer is obtained based on the multiple noise sub-layers. Different noise layers with different scale ranges can be obtained by setting different parameters of the noise function. Different scales mean that the area ranges of the noise objects in each noise layer are different. Refer to Figure 3 the four noise sub-layers of different scales shown in the example in

[0137] Step S602: Extract connected regions from the noise layer to obtain multiple noise objects.

[0138] Exemplarily, taking the freckle objects shown in Figure 3 as an example, extract the connected regions from the four noise sub-layers of different scales to obtain multiple noise sub-objects in each noise sub-layer. The four noise sub-layers can be fused by addition to obtain a noise layer. By extracting the connected regions from the noise layer, multiple noise objects can be obtained.

[0139] Step S603: Discard the extracted noise objects to obtain texture objects.

[0140] Furthermore, the skin image to be processed includes a skin area and a non-skin area. During the discarding process, randomly discard the multiple noise objects corresponding to the skin area, and discard the multiple noise objects in the non-skin area. For example, in one discarding method, set the positions where the discarded noise objects are located to 0, which means there are no noise objects at these positions.

[0141] Optionally, for the process of discarding the multiple noise objects in the non-skin area, it can be achieved by multiplying the noise layer by the mask of the freckle-free area of the face. The freckle-free area refers to the non-skin area where no freckles are set. That is to say, through the mask processing, the noise objects in the non-skin areas such as eyes, lips, and ears are removed from the noise layer.

[0142] Optionally, taking the four noise sub-layers of different scales shown in Figure 3 as an example, the mask and discard processing can be performed before fusing to obtain the noise layer. The implementation methods of the mask and discard processing are not elaborated here.

[0143] Exemplarily, when setting freckle textures in the face region, the probability distribution of the discard process can be a bivariate Gaussian distribution centered on the tip of the nose, that is, the discard probability at the tip of the nose is the smallest, and the discard probability gradually increases towards the surroundings. That is to say, the probability distribution of the discard process can be determined according to the real freckle distribution in the face. The present application does not limit the probability distribution of the discard process.

[0144] Step S604: Group multiple texture objects to obtain N groups of texture objects and store them in a storage system, where N is an integer greater than zero. Exemplarily, the grouping method can be the method as Figure 5 shown. The storage system refers to any medium, device, etc. with a storage function, and the present application does not limit it.

[0145] Optionally, in step S605, classify according to a preset region, and group multiple textures in each region.

[0146] Optionally, divide the corresponding skin detail objects into N groups so that the sum of the areas of multiple texture objects in each group is approximately equal. That is to say, when increasing each adjustment step, the area of the increased texture objects is approximately equal.

[0147] Exemplarily, according to the preset face regions, all freckle objects are divided into five regions: forehead, nose, cheeks, chin, and other regions. The freckle objects in these five regions are grouped respectively. In each region, the areas of the freckle objects in each group are approximately equal. Exemplarily, for the binning algorithm, N is taken as 100 and numbered from 1 to 100.

[0148] Optionally, in step S606, through a pigment decomposition algorithm, decompose the basic pigment information corresponding to each pixel position in the skin image to be processed and store it in the storage system. Optionally, the basic pigment information includes melanin information and hemoglobin information.

[0149] Exemplarily, a machine learning algorithm combining PCA (Principal Component Analysis) and ICA (Independent Component Analysis) can be used to perform pigment decomposition on the skin image to be processed. It is also possible to use the color and pigment component as a training data sample pair, and through supervised learning, obtain an encoder. Perform pigment decomposition through the trained encoder.

[0150] Optionally, in step S607, the adjustment parameters input by the user include region parameters. Obtain the region parameters input by the user and determine the region to be processed from multiple image regions. The method of dividing regions is as shown in step S604 and will not be elaborated here.

[0151] Step S608: Read multiple preprocessed texture objects from the storage system.

[0152] Step S609: The texture parameter includes texture density, and the texture density represents the number of groups among N groups of texture objects. By obtaining the texture density input by the user, determine the number of groups K, where K is an integer greater than or equal to zero and less than or equal to N. That is to say, in the area to be processed, all texture objects included in 1 to K groups need to be added.

[0153] In other examples, the texture parameter includes the number of textures. That is to say, the user can select the number of texture objects, so as to fuse the corresponding number of texture objects with the skin image to be processed to obtain an adjusted image. For example, the number of texture objects is displayed to the user in the form of a percentage, and the user can select the percentage of texture objects to determine the number of texture objects to be fused with the skin image to be processed.

[0154] Optionally, in step S610, the adjustment parameter further includes a pigment parameter, and the pigment parameter further includes a melanin parameter and a hemoglobin parameter. Read the basic pigment information of the skin image to be processed from the storage system. And / or, linearly adjust the melanin information in the basic pigment information according to the melanin parameter; linearly adjust the hemoglobin information in the basic pigment information according to the hemoglobin parameter.

[0155] Exemplarily, assume that the area to be processed is selected as the cheek area, and the grouped freckle objects stored in the cheek area are taken out. Assume that the texture density input by the user is 40, the melanin parameter is 1.3, and the hemoglobin parameter is 0.8. Then, respectively weight the melanin component (an example of melanin information) in the basic pigment information by 1.3 and the hemoglobin component (an example of hemoglobin information) by 0.8. Select all freckle objects included in groups 1 to 40 in the cheek area. In the area of the freckle objects, it is necessary to adjust the melanin component and hemoglobin component of each pixel point in the freckle objects.

[0156] Exemplarily, in a freckle object (a specific example of freckle texture), assume that the freckle texture value corresponding to a certain pixel point is α, the melanin component is C m , and the hemoglobin component is C h , the melanin parameter is 1.3, and the hemoglobin parameter is 0.8. Then the adjusted melanin component of this pixel point is 1.3*C m , and the hemoglobin component is 0.8*C h . After the adjusted melanin component and hemoglobin component are fused with the skin texture of this pixel point (that is, the freckle texture value is α), the fused melanin component of this pixel point is 1.3*C m *α + 1*Cm *(1 - α), the hemoglobin component is 0.8*C h *α + 1*C h *(1 - α). Since the value range of α is (0, 1] and shows a gradient from the center to the periphery. Therefore, after adjustment, the freckle texture value in the edge area of the freckle object is α ≈ 0, and the melanin component is approximately equal to C m and the hemoglobin component is approximately equal to C h , which is the same as the melanin component C m and the hemoglobin component C h in the edge area of the adjacent non - freckle object. Therefore, the transition between the adjusted area of the freckle object and the non - adjusted area of the non - freckle object is natural, close to the transition process of freckle texture in real skin, and has a strong sense of reality. At the same time, the freckle texture value in the center area of the freckle object is α ≈ 1, the melanin component is approximately equal to 1.3*C m and the hemoglobin component is approximately equal to 0.8*C h . Compared with the melanin component C m and the hemoglobin component C h of normal skin color, the melanin component is larger and the hemoglobin component is smaller, which can present the effect of freckles. That is to say, the boundary of the freckle object is close to the freckle boundary of real skin, and the transition from non - freckled skin to freckle texture is more natural and has a stronger sense of reality.

[0157] That is to say, compared with only increasing skin texture through color, the method of the embodiment of the present application solves the problems of distortion and uneven edges existing in color space editing, and improves the sense of reality of the adjusted image.

[0158] Optionally, in step S611, through the method of color reconstruction, the processing result is reconstructed into the color space to obtain the adjusted image. The method of color reconstruction can be obtained through the inverse transformation of the pigment decomposition algorithm, which is not elaborated in the embodiment of the present application here.

[0159] In other examples, by selecting different noise functions and discard functions, the method of the embodiment of the present application can also be used in the adjustment process of tattoos, eye makeup, spots, scars and other textures, which is not elaborated in the embodiment of the present application here. Taking scars as an example, through mask processing, the scar can be used as a separate area to be processed, and the method provided by the embodiment of the present application is implemented on the area to be processed where the scar is located to adjust the skin details of the scar, such as adjusting the skin pigment and skin texture of the scar.

[0160] Optionally, the method at least includes a pre - processing stage and an adjustment stage. The pre - processing stage includes steps S601 to S606, and the adjustment stage includes steps S607 to S611.

[0161] Based on the same concept as the foregoing embodiments, an adjustment device for a digital human skin is further provided in an embodiment of the present application.

[0162] Please refer to Figure 7 , Figure 7 which is a schematic diagram of the composition of an adjustment device for a digital human skin provided in an embodiment of the present application. As Figure 7 shown, an adjustment device 700 for a digital human skin at least includes: a first acquisition module 701, configured to acquire a to-be-processed skin image input by a user; a noise object determination module 702, configured to determine a noise object according to a noise layer, where the noise object is a connected region of the noise layer; a texture object determination module 703, configured to determine a texture object according to the to-be-processed skin image and the noise object, where the texture object is used to indicate the skin texture in the to-be-processed skin image; a second acquisition module 704, configured to acquire an adjustment parameter input by the user, where the adjustment parameter includes a texture parameter; a target texture object determination module 705, configured to determine a target texture object from the texture objects according to the texture parameter; and a target skin image generation module 706, configured to generate a target skin image according to the target texture object and the to-be-processed skin image, where the target skin image includes the target texture object.

[0163] Optionally, for ease of description, the various modules in the adjustment device 700 for a digital human skin are grouped. The first acquisition module 701, the noise object determination module 702, and the texture object determination module 703 belong to a preprocessing unit; the second acquisition module 704, the target texture object determination module 705, and the target skin image generation module 706 belong to an adjustment unit. This grouping does not constitute a limitation to the present application. According to different usage scenarios, the various modules in the adjustment device 700 for a digital human skin may also not be grouped or grouped in other ways.

[0164] In a possible implementation manner, the noise object determination module 702 is specifically configured to: acquire a plurality of preset noise sub-layers, each noise sub-layer including a plurality of noise sub-objects, where the plurality of noise sub-layers are layers of different scales, and different scales indicate different area scales of the noise sub-objects; determine the noise layer according to the plurality of noise sub-layers; and extract the connected domain of the noise layer to obtain the noise object.

[0165] In another possible implementation manner, the texture object determination module 703 is specifically configured to: process the noise layer according to the to-be-processed skin image to obtain a mask layer; and process the noise object in the obtained mask layer according to a random function to obtain the texture object.

[0166] In another possible implementation, the apparatus 700 further includes: a texture object grouping module, configured to divide texture objects into N groups with the constraint that the difference in the sum of areas between each group is minimized, where the sum of areas represents the sum of the areas of the texture objects in each group.

[0167] In another possible implementation, the texture parameter includes the number of texture groups K, where K is an integer greater than or equal to zero and less than or equal to N. The target texture object determination module 705 is specifically configured to: determine the texture objects in K groups as target texture objects according to the number of texture groups K.

[0168] In another possible implementation, the apparatus 700 further includes: a basic pigment information determination module, configured to decompose the pigments of the skin image to be processed and determine the basic pigment information.

[0169] In another possible implementation, the adjustment parameter further includes a pigment parameter, and the apparatus 700 further includes: a target pigment information determination module, configured to determine the target pigment information according to the pigment parameter and the basic pigment information. The target skin image generation module 706 is specifically configured to: generate a target skin image according to the target texture object, the target pigment information, and the skin image to be processed.

[0170] In another possible implementation, the basic pigment information includes basic melanin information, the target pigment information includes target melanin information, and the pigment parameter includes a melanin parameter; the target pigment information determination module is specifically configured to: determine the target melanin information according to the melanin parameter and the basic melanin information. And / or, the basic pigment information includes basic hemoglobin information, the target pigment information includes target hemoglobin information, and the pigment parameter includes a hemoglobin parameter; the target pigment information determination module is specifically configured to: determine the target hemoglobin information according to the hemoglobin parameter and the hemoglobin information.

[0171] In another possible implementation, the apparatus 700 further includes: an image area division module, configured to divide the skin image to be processed into at least one image area. The adjustment parameter further includes an area parameter, and the apparatus 700 further includes: a to-be-processed area determination module, configured to determine the to-be-processed area from the at least one image area according to the area parameter. The target texture object determination module 705 is specifically configured to: determine the target texture objects from the texture objects in the to-be-processed area according to the texture parameter.

[0172] Based on the same concept as the foregoing embodiments, an embodiment of the present application further provides a user interface system.

[0173] Please refer to Figure 8 , Figure 8 which is a schematic diagram of the composition of a user interface system provided by an embodiment of the present application. As Figure 8As shown, a user interface system 800 mainly includes:

[0174] A client 810, configured to receive a skin image to be processed and adjustment parameters input by a user. Optionally, the client 810 is provided with a display interface as Figure 8 shown.

[0175] A server 820, configured to obtain the skin image to be processed and the adjustment parameters in the client, and execute the algorithm functions embodied by the method of any one of the above or the device of any one of the above.

[0176] As Figure 8 shown, the adjustment parameters input by the user include texture parameters. The texture parameters include texture density and texture quantity, and are used to adjust the quantity of texture objects displayed.

[0177] Optionally, as Figure 8 shown, the adjustment parameters input by the user further include pigment parameters. The pigment parameters include melanin parameters, which are used to adjust the melanin information in the skin image to be processed. The pigment parameters include hemoglobin parameters, which are used to adjust the hemoglobin information in the skin image to be processed.

[0178] Optionally, as Figure 8 shown, the adjustment parameters input by the user further include region parameters. Exemplarily, as Figure 8 shown, the skin image to be processed is divided into multiple image regions, such as a forehead region, a nose region, a cheek region, a mouth region, a chin region, etc. The user can determine one or more of the image regions as the regions to be processed.

[0179] Optionally, the user interface system 800 can be an interface in an existing production engine. Existing production engines, such as Unreal Engine, Unity Creation Engine, Blender rendering software, Maya software, etc. The user interface system 800 can also be an independently developed production tool. The user interface system 800 can provide a fast skin detail editing function to improve the efficiency of skin editing.

[0180] Optionally, the user interface system 800 can also be integrated into a client facing the user. That is to say, the client 810 and the server 820 can be respectively deployed in the client 110 and the server 120, or the client 810 and the server 820 can be integrated and deployed in the client 110.

[0181] Optionally, the user interface system 800 can also be directly provided to the user as an independent module, facilitating the user to integrate the user interface system 800 into a specific application scenario.

[0182] Further, as Figure 8As shown, in the client 810, there is a control for adjusting the pigment parameter. During the adjustment process, the change process of skin pigment from light to dark or from dark to light can be presented. The implementation method of the control is not limited and can be, for example, Figure 8 the progress slider shown, or other forms of controls such as text boxes.

[0183] Furthermore, as Figure 8 shown, in the client 810, there is a control for adjusting the texture parameter. During the adjustment process, the change process of skin texture from more to less or from less to more can be presented. The implementation method of the control is not limited.

[0184] Exemplarily, as Figure 8 shown, in the client 810, Figure 8 on the right is the obtained digital human image. By masking non-skin areas such as the eyes, lips, and ears of the digital human image, the skin image to be processed shown on the left is obtained. Alternatively, the user can also directly input the skin image to be processed shown on the left as Figure 8 . As Figure 8 shown on the left, the human face is divided into a forehead area, an eye area, a nose area, a mouth area, and a chin area, and different image areas are represented by corresponding masks. For example, when the forehead area is the area to be processed, the eye area, the nose area, the mouth area, and the chin area are covered. Optionally, after successfully importing the skin image to be processed, the server 820 preprocesses the skin image to be processed to obtain a processing result, such as multiple texture objects and basic pigment information, etc., and stores them in the storage system. When the user determines the adjustment parameters in the client 810, the server 820 obtains the adjusted image according to the adjustment parameters and the processing result of the preprocessing. Optionally, in Figure 8 the area of the skin image to be processed on the left, the adjusted image is displayed in real time; the skin image to be processed and the adjusted image can also be displayed simultaneously so that the user can further understand the changes before and after the adjustment. Figure 8 shown on the left

[0185] Based on the same concept as the foregoing embodiments, an embodiment of the present application also provides a computing device 900. As Figure 9 shown, the computing device 900 includes: a bus 902, a processor 904, a memory 906, and a communication interface 908. The processor 904, the memory 906, and the communication interface 908 communicate with each other through the bus 902. The computing device 900 can be a server, such as a central server, an edge server, or a local server in a local data center, or an electronic device such as a desktop computer, a laptop computer, or a smart phone. It should be understood that the present application does not limit the number of processors and memories in the computing device 900.

[0186] The bus 902 can be a Peripheral Component Interconnect (PCI) bus, an Extended Industry Standard Architecture (EISA) bus, or the like. The bus can be divided into an address bus, a data bus, a control bus, etc. For ease of representation, Figure 9 only one line is used in Figure 9 , but it does not mean that there is only one bus or one type of bus. The bus 804 can include a path for transmitting information between various components of the computing device 900 (for example, the memory 906, the processor 904, the communication interface 908).

[0187] The processor 904 can include any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0188] The memory 906 can include volatile memory, such as random access memory (RAM). The processor 904 can also include non-volatile memory, such as read-only memory (ROM), flash memory, a hard disk drive (HDD), or a solid state drive (SSD).

[0189] The memory 906 stores executable program code, and the processor 904 executes the executable program code to respectively implement the functions of one or more modules in the digital human skin adjustment device 700 shown in the foregoing Figure 7 so as to implement the digital human skin adjustment method described in the above embodiments. That is, the memory 906 stores instructions for executing the digital human skin adjustment method described in the above embodiments.

[0190] Alternatively, the memory 906 stores executable code, and the processor 904 executes the executable code to implement the functions of the digital human skin adjustment device 700 shown in the foregoing Figure 7 so as to implement the digital human skin adjustment method described in the above embodiments. That is, the memory 906 stores instructions for executing the digital human skin adjustment method described in the above embodiments.

[0191] The communication interface 908 uses a transceiver module such as, but not limited to, a network interface card or a transceiver to implement communication between the computing device 900 and other devices or communication networks.

[0192] Embodiments of the present application also provide a computing device cluster. The computing device cluster includes at least one computing device. The computing device may be a server, such as a central server, an edge server, or a local server in a local data center. In some embodiments, the computing device may also be an electronic device such as a desktop computer, a laptop computer, or a smart phone.

[0193] As Figure 10 shown, the computing device cluster includes at least one computing device 900. Instructions for executing the adjustment method of the digital human skin described in the foregoing embodiments may be stored in the memory 906 of one or more of the computing devices 900 in the computing device cluster.

[0194] In some possible implementation manners, partial instructions for executing the adjustment method of the digital human skin described in the foregoing embodiments may also be stored separately in the memory 906 of one or more of the computing devices 900 in the computing device cluster. In other words, a combination of one or more computing devices 900 may jointly execute the instructions for executing the adjustment method of the digital human skin described in the foregoing embodiments.

[0195] It should be noted that the memories 906 in different computing devices 900 in the computing device cluster may store different instructions, respectively for executing partial functions of the adjustment device 700 of the digital human skin shown in the foregoing Figure 7 shown. That is, the instructions stored in the memories 906 of different computing devices 900 may implement the functions of one or more modules of the adjustment device 700 of the digital human skin.

[0196] In some possible implementation manners, one or more computing devices in the computing device cluster may be connected through a network. Among them, the network may be a wide area network or a local area network, etc. Figure 11 Shows a possible implementation manner. As Figure 11As shown, two computing devices 900A and 900B are connected via a network. Specifically, they are connected to the network through the communication interfaces in each computing device. In this possible implementation, the memory 906 in computing device 900A stores instructions for executing the functions of one or more modules in the adjustment device 700 of the digital human skin. At the same time, the memory 906 in computing device 900B stores instructions for executing the functions of one or more modules in the adjustment device 700 of the digital human skin. That is to say, the combination of computing devices 900A and 900B can jointly execute the instructions for performing the adjustment method of the digital human skin described in the above embodiments.

[0197] It should be understood that Figure 11 the functions of the computing device 900A shown in can also be completed by multiple computing devices 900. Similarly, the functions of the computing device 900B can also be completed by multiple computing devices 900.

[0198] The embodiments of the present application also provide another computing device cluster. The connection relationship between the computing devices in this computing device cluster can be similarly referred to Figure 10 and Figure 11 the connection method of the said computing device cluster. The difference is that the memory 906 in one or more computing devices 900 in this computing device cluster can store the same instructions for executing the method in the above embodiments.

[0199] In some possible implementations, the memory 906 of one or more computing devices 900 in this computing device cluster can also separately store partial instructions for executing the foregoing data processing method. In other words, the combination of one or more computing devices 900 can jointly execute the instructions for performing the adjustment method of the foregoing digital human skin.

[0200] Based on the same concept as the foregoing embodiments, the present application provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the functions embodied by the method or device provided by the embodiments of the present application are realized.

[0201] Based on the same concept as the foregoing embodiments, the present application provides a computer program product, which includes program instructions. When the program instructions are executed by a computer, the computer executes the functions embodied by the method or device provided by the embodiments of the present application.

[0202] Those of ordinary skill in the art should also be further aware that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, computer software, or a combination of the two. To clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described according to functions in the above description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those of ordinary skill in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of this application.

[0203] The steps of the methods or algorithms described in combination with the embodiments disclosed herein can be implemented by hardware, software modules executed by a processor, or a combination of the two. The software modules can be placed in a random access memory (RAM), internal memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other form of storage medium known in the technical field.

[0204] The specific embodiments described above further elaborate on the purpose, technical solution, and beneficial effects of this application. It should be understood that the above description is only the specific embodiments of this application and is not used to limit the protection scope of this application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of this application shall be included in the protection scope of this application.

Claims

1. A method for adjusting a digital human skin, characterized in that, The method includes: Obtaining a skin image to be processed input by a user; Determining a noise object according to a noise layer, where the noise object is a connected region of the noise layer; Determining a texture object according to the skin image to be processed and the noise object, where the texture object is used to indicate skin texture in the skin image to be processed; Obtaining an adjustment parameter input by the user, where the adjustment parameter includes a texture parameter; Determining a target texture object from the texture object according to the texture parameter; Generating a target skin image according to the target texture object and the skin image to be processed, where the target skin image includes the target texture object.

2. The method according to claim 1, wherein The method further includes: Obtaining a plurality of preset noise sub-layers, where each noise sub-layer includes a plurality of noise sub-objects, the plurality of noise sub-layers are layers of different scales, and the different scales indicate different area scales of the noise sub-objects; Determining the noise layer according to the plurality of noise sub-layers; The determining a noise object according to the noise layer includes: Extracting the connected regions of the noise layer to obtain a noise object.

3. The method according to claim 1 or 2, characterized in that, The determining a texture object according to the skin image to be processed and the noise object includes: Processing the noise layer according to the skin image to be processed to obtain a mask layer; Processing the noise object in the mask layer according to a random function to obtain a texture object.

4. The method according to any one of claims 1 to 3, characterized in that Before obtaining the adjustment parameter input by the user, the method further includes: Dividing the texture object into N groups with the constraint that the difference in the sum of areas between each group is minimized, where the sum of areas represents the sum of the areas of the texture objects in each group, and N is an integer greater than 1.

5. The method according to claim 4, wherein The texture parameter includes a texture group number K, where K is an integer greater than or equal to zero and less than or equal to N. The determining a target texture object from the texture object according to the texture parameter includes: Determining the texture objects in K groups as target texture objects according to the texture group number K.

6. The method according to any one of claims 1-5, characterized in that Before obtaining the adjustment parameter input by the user, the method further includes: Decomposing the pigments of the skin image to be processed to obtain basic pigment information.

7. The method according to claim 6, wherein The adjustment parameter further includes a pigment parameter, and the method further includes: Determining target pigment information according to the pigment parameter and the basic pigment information; The generating a target skin image according to the target texture object and the skin image to be processed includes: Generating a target skin image according to the target texture object, the target pigment information, and the skin image to be processed.

8. The method according to claim 7, characterized in that, The basic pigment information includes basic melanin information, the target pigment information includes target melanin information, and the pigment parameter includes a melanin parameter; The determining target pigment information according to the pigment parameter and the basic pigment information includes: determining the target melanin information according to the melanin parameter and the basic melanin information; And / or, the basic pigment information includes basic hemoglobin information, the target pigment information includes target hemoglobin information, and the pigment parameter includes a hemoglobin parameter; Determining the target pigment information according to the pigment parameter and the basic pigment information includes: determining the target hemoglobin information according to the hemoglobin parameter and the hemoglobin information.

9. The method according to any one of claims 1-8, characterized in that It further includes: Dividing the to-be-processed skin image into at least one image region; The adjustment parameter further includes a region parameter, and the method further includes: determining a to-be-processed region from the at least one image region according to the region parameter; Determining the target texture object from the texture objects according to the texture parameter includes: Determining the target texture object from the texture objects in the to-be-processed region according to the texture parameter.

10. An adjustment device for a digital human skin, characterized in that, It includes: A first acquisition module, configured to acquire a to-be-processed skin image input by a user; A noise object determination module, configured to determine a noise object according to a noise layer, where the noise object is a connected region of the noise layer; A texture object determination module, configured to determine a texture object according to the to-be-processed skin image and the noise object, where the texture object is used to indicate skin texture in the to-be-processed skin image; A second acquisition module, configured to acquire an adjustment parameter input by the user, where the adjustment parameter includes a texture parameter; A target texture object determination module, configured to determine a target texture object from the texture objects according to the texture parameter; A target skin image generation module, configured to generate a target skin image according to the target texture object and the to-be-processed skin image, where the target skin image includes the target texture object.

11. The device according to claim 10, characterized in that, The noise object determination module is specifically configured to: Acquire a plurality of preset noise sub-layers, each of the noise sub-layers includes a plurality of noise sub-objects, the plurality of noise sub-layers are layers of different scales, and the different scales indicate that the area scales of the noise sub-objects are different; Determine a noise layer according to the plurality of noise sub-layers; Extract the connected domain of the noise layer to obtain a noise object.

12. The device according to claim 10 or 11, characterized in that, The texture object determination module is specifically configured to: Process the noise layer according to the to-be-processed skin image to obtain a mask layer; Process the noise object in the mask layer according to a random function to obtain a texture object.

13. The device according to any one of claims 10 to 12, characterized in that It further includes: A texture object grouping module, configured to divide the texture objects into N groups with the constraint that the difference in the sum of areas between each group is the smallest, where the sum of areas represents the sum of the areas of the texture objects in each group.

14. The device according to claim 13, characterized in that, The texture parameter includes the number of texture groups K, where K is an integer greater than or equal to zero and less than or equal to N; the target texture object determination module is specifically configured to: determine the texture objects in K groups as the target texture objects according to the number of texture groups K.

15. The device according to any one of claims 10-14, characterized in that, It further includes: A basic pigment information determination module, configured to decompose the pigment of the to-be-processed skin image to determine basic pigment information.

16. The device according to claim 15, characterized in that, The adjustment parameter further includes a pigment parameter, and the apparatus further includes: a target pigment information determination module, configured to determine target pigment information according to the pigment parameter and the basic pigment information; The target skin image generation module is specifically configured to: generate a target skin image according to the target texture object, the target pigment information, and the to-be-processed skin image.

17. The device according to claim 16, characterized in that, The basic pigment information includes basic melanin information, the target pigment information includes target melanin information, and the pigment parameter includes a melanin parameter; specifically, the target pigment information determination module is configured to: determine the target melanin information according to the melanin parameter and the basic melanin information; and / or, the basic pigment information includes basic hemoglobin information, the target pigment information includes target hemoglobin information, and the pigment parameter includes a hemoglobin parameter; specifically, the target pigment information determination module is configured to: determine the target hemoglobin information according to the hemoglobin parameter and the hemoglobin information.

18. The device according to any one of claims 10-17, characterized in that It further includes: an image area division module, configured to divide the to-be-processed skin image into at least one image area; The adjustment parameter further includes a region parameter, and the apparatus further includes: a to-be-processed region determination module, configured to determine a to-be-processed region from the at least one image area according to the region parameter; Specifically, the target texture object determination module is configured to: determine a target texture object from the texture objects in the to-be-processed region according to the texture parameter.

19. A user interface system, characterized in that, It includes: a client, configured to receive a to-be-processed skin image and an adjustment parameter input by a user; a server, configured to obtain the to-be-processed skin image and the adjustment parameter in the client, and execute the method according to any one of claims 1-9 or the functions embodied by the apparatus according to any one of claims 10-18.

20. A cluster of computing devices, characterized in that, It includes at least one computing device, and each computing device in the at least one computing device includes a memory and a processor. It is characterized in that instructions are stored in the memory, and when the instructions are executed by the processor, the functions embodied by the method according to any one of claims 1-9 or the apparatus according to any one of claims 10-18 are realized.

21. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the functions embodied by the method according to any one of claims 1-9 or the apparatus according to any one of claims 10-18 are realized.

22. A computer program product, characterized in that, The computer program product includes program instructions, and when the program instructions are executed by a computer, the computer is caused to execute the method according to any one of claims 1-9 or the functions embodied by the apparatus according to any one of claims 10-18.

Images