A method and related apparatus for generating digital humans

Abstract

This application provides a digital human generation method and related apparatus. Responding to a digital human generation command, an initial digital human is generated based on a user's image. Using a standard digital human as a reference, the initial digital human is adjusted to match the standard digital human, resulting in an adjusted digital human. A target accessory corresponding to a target part is then assembled with the target part of the adjusted digital human to generate an assembled digital human. In this application, using a standard digital human as a reference, the initial digital human is adjusted to match the standard digital human, thus obtaining the adjusted digital human. Since the adjusted digital human maintains the same scale and shape as the standard digital human, and the target accessory is made based on the target part of the standard digital human, the target accessory corresponding to the target part can be directly assembled with the target part of the adjusted digital human, effectively improving the generation efficiency of the assembled digital human.

Description

Technical Field

[0001] This application relates to the field of computer technology, specifically to the field of digital human technology, and in particular to a method and related apparatus for generating a digital human. Background Technology

[0002] Digital humans are digitized human figures created using digital technology that closely resemble human appearances. Accessories, as an important component of the digital human's image, can enhance its visual realism and reinforce its individual characteristics.

[0003] Currently, by inputting user images into a model, digital humans with user characteristics can be generated based on deep learning and image processing algorithms. However, digital humans generated from user images are difficult to match directly with accessories, requiring technicians to manually adjust the accessories and equip the digital human with the adjusted accessories, resulting in reduced generation efficiency.

[0004] Therefore, there is an urgent need for a solution to address the aforementioned technical problems. Summary of the Invention

[0005] Based on the above problems, this application provides a digital human generation method and related apparatus, aiming to solve the problem of low efficiency in digital human generation.

[0006] The embodiments of this application disclose the following technical solutions:

[0007] Firstly, this application provides a digital human generation method, comprising: responding to a digital human generation instruction, generating an initial digital human based on a user's image; adjusting the initial digital human to match the standard digital human, using a standard digital human as a reference, to obtain an adjusted digital human; the scale and shape of the adjusted digital human being consistent with the scale and shape of the standard digital human being; and assembling a target accessory corresponding to a target part with the target part of the adjusted digital human to generate an assembled digital human, wherein the target accessory is made based on the target part of the standard digital human being.

[0008] In this embodiment, a standard digital human is used as a reference. The initial digital human is adjusted to match the standard digital human to obtain an adjusted digital human. Since the adjusted digital human maintains the same scale and shape as the standard digital human, and the target accessory is made based on the target parts of the standard digital human, the target accessory corresponding to the target part can be directly assembled with the target part of the adjusted digital human, effectively improving the generation efficiency of the assembled digital human.

[0009] In one possible implementation, adjusting the initial digital human to match the standard digital human as a reference to obtain an adjusted digital human includes: adjusting the scale of the initial digital human to match the scale of the standard digital human as a reference to obtain a scale-matched digital human; and adjusting the position information of the index points of the target parts of the scale-matched digital human based on the position information of the index points of the target parts of the standard digital human to generate the adjusted digital human.

[0010] In this embodiment, the initial digital human is adjusted in terms of both scale and shape so that the scale and shape of each part are consistent with the standard digital human, so that the target accessories can be directly assembled, which effectively improves the generation efficiency of the assembled digital human.

[0011] In one possible implementation, adjusting the position information of the index points of the target part of the digital human based on the position information of the index points of the target part of the digital human based on the standard, and generating the adjusted digital human, includes: for the target part, adjusting the position information of the index points of the digital human based on the position information of the index points of the standard digital human, and adjusting the position information of the index points of the scale-matched digital human that correspond to the standard digital human, and generating the adjusted digital human.

[0012] In this embodiment of the application, after obtaining a scale-matched digital human, adjustments were made to the target parts that still had some differences, so that the scale and shape of each part of the adjusted digital human were consistent with the standard digital human.

[0013] In one possible implementation, adjusting the scale of the initial digital human to match the scale of the standard digital human, based on a standard digital human, to obtain a scale-matched digital human includes: determining the scaling ratio, rotation matrix, and translation amount with the minimum value of an energy function as the optimization objective; the energy function is used to indicate the distance between each index point of the initial digital human and the index points of the standard digital human after scaling, rotation, and translation; and adjusting the scale of the initial digital human to match the scale of the standard digital human based on the scaling ratio, rotation matrix, and translation amount to obtain a scale-matched digital human.

[0014] In this embodiment, minimizing the energy function is the optimization objective, which enables rapid and accurate scale alignment between the initial digital human and the standard digital human.

[0015] In one possible implementation, adjusting the position information of the index points of the target parts of the digital human based on the position information of the index points of the target parts of the digital human based on the standard, and generating an adjusted digital human, includes: in response to an adjustment instruction, adjusting the position information of the index points of the target parts of the digital human based on the position information of the index points of the target parts of the digital human based on the standard, and generating an adjusted digital human, wherein the adjustment instruction is used to indicate the target parts corresponding to the target accessories.

[0016] In this embodiment of the application, the target part can be determined based on the adjustment command, and the target part can be adjusted more specifically, which can effectively improve the adjustment efficiency and flexibility, thereby improving the generation efficiency of the assembled digital human.

[0017] In one possible implementation, the initial digital human and the standard digital human are in different scale spaces. The step of adjusting the initial digital human to match the standard digital human based on the standard digital human to obtain an adjusted digital human includes: converting the initial digital human and the standard digital human to the same scale space; and in the scale space, adjusting the initial digital human to match the standard digital human based on the standard digital human to obtain an adjusted digital human.

[0018] In this embodiment of the application, by converting the initial digital human and the standard digital human to the same scale space, the initial digital human and the standard digital human can be in the same coordinate system, which facilitates the subsequent adjustment process.

[0019] In one possible implementation, adjusting the initial digital human to match the standard digital human in the scale space, based on the standard digital human, to obtain the adjusted digital human includes: registering the index points of the target parts of the initial digital human with the index points of the target parts of the standard digital human in the scale space, based on the standard digital human, to obtain index point pairs; and adjusting the initial digital human to match the standard digital human based on the index point pairs to obtain the adjusted digital human. In this embodiment, index point pairing facilitates the subsequent adjustment process.

[0020] In one possible implementation, the target body part is the head; the target accessory includes one or more of hair, glasses, hats, head-mounted rehabilitation devices, surgical helmets, and goggles.

[0021] Generally, the initial digital human's head features are quite complex. The method provided in this application embodiment can adjust the head details so that the scale and shape of the initial digital human's head are consistent with the scale and shape of the standard digital human's head. This allows head accessories to be directly fitted onto the adjusted digital human's head, thereby improving the generation efficiency of the digital human while ensuring the overall visual effect of the digital human.

[0022] Second aspect: Embodiments of this application provide a digital human generation device, including:

[0023] Production unit, adjustment unit, and assembly unit;

[0024] The generation unit is configured to generate an initial digital human based on the user's image in response to a digital human generation command;

[0025] The adjustment unit is used to adjust the initial digital human to match the standard digital human based on the standard digital human, thereby obtaining an adjusted digital human; the scale and shape of the adjusted digital human are consistent with the scale and shape of the standard digital human.

[0026] The assembly unit is used to assemble the target accessory corresponding to the target part with the adjusted target part of the digital human to generate the assembled digital human. The target accessory is made based on the target part of the standard digital human.

[0027] Third aspect: This application provides a computer device, which includes a processor and a memory;

[0028] The memory is used to store program code and transmit the program code to the processor;

[0029] The processor is configured to execute the steps of a digital human generation method as described in the first aspect above, according to instructions in the program code.

[0030] Fourth aspect: Embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a digital human generation method as described in the first aspect above.

[0031] Fifth aspect: This application provides a computer program product, which, when run on a computer, executes the steps of a digital human generation method as described in the first aspect above. Attached Figure Description

[0032] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0033] Figure 1 A schematic diagram illustrating an application scenario provided in an embodiment of this application;

[0034] Figure 2 A flowchart illustrating a digital human generation method provided in this application embodiment;

[0035] Figure 3 A schematic diagram of adjusting an initial digital human as provided in an embodiment of this application;

[0036] Figure 4 A schematic diagram illustrating a digital human generation method provided in an embodiment of this application;

[0037] Figure 5 This is a schematic diagram of the structure of a digital human generation device provided in an embodiment of this application;

[0038] Figure 6 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0039] The terms "first," "second," and "third," etc., used in this application specification, claims, and drawings are used to distinguish different objects, not to limit a specific order.

[0040] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of the terms "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0041] Currently, by inputting user images into a model, digital humans with user characteristics can be generated based on deep learning and image processing algorithms. However, due to limitations such as the shooting angle and distance differences of user images, as well as the complexity of some user features, digital humans generated from user images suffer from inconsistent scale in some parts, which affects the subsequent process of equipping the digital humans with accessories.

[0042] For example, for digital humans with large heads, the pre-designed hair model may be too small to fully cover the head, resulting in an unnatural connection between the hair and the head. The size of the glasses may also be too small to match the digital human, affecting the overall visual effect of the digital human. For digital humans with small heads, the hair and glasses may be too large, which will also disrupt the image harmony of the digital human.

[0043] To ensure the overall visual effect of the digital human, technicians can currently manually adjust the accessories and equip the digital human with the adjusted accessories. However, this method has the following problems:

[0044] Firstly, this method requires technicians to perform meticulous manual operations, resulting in high labor costs and low efficiency. Secondly, the process of technicians manually adjusting accessories makes it difficult to establish a standardized process and a replicable template. Whenever a new digital human needs to be generated, technicians need to readjust the accessories, affecting the generation efficiency of digital humans. Thirdly, when faced with a large demand for digital human generation, this method requires a significant amount of human resources and time, failing to meet the needs of rapid production and large-scale application.

[0045] Based on this, embodiments of this application provide a digital human generation method. In response to a digital human generation command, an initial digital human is generated based on a user's image. Using a standard digital human as a reference, the initial digital human is adjusted to match the standard digital human, resulting in an adjusted digital human. A target accessory corresponding to a target part is then assembled with the target part of the adjusted digital human to generate an assembled digital human. The target accessory is made based on the target part of the standard digital human.

[0046] In this embodiment, a standard digital human is used as a reference. The initial digital human is adjusted to ensure that the adjusted digital human maintains the same scale and shape as the standard digital human. Furthermore, since the target accessories are made based on target parts of the standard digital human, the target accessories corresponding to the target parts can be directly assembled with the target parts of the adjusted digital human. This effectively improves the generation efficiency of the assembled digital human, meeting the needs of rapid production and large-scale application.

[0047] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present application.

[0048] See Figure 1 This figure is a schematic diagram of an application scenario provided by an embodiment of this application.

[0049] The digital human generation method provided in this application can be applied to computer devices. In some embodiments, the computer device may include, but is not limited to, mobile phones, tablets, laptops, handheld computers, wearable electronic devices, smartwatches, etc. This application does not impose any special restrictions on the specific form of the above-mentioned electronic devices.

[0050] In this embodiment of the application, computer device 1101 can interact with server 1102. Computer device 1101 can respond to user operations, acquire user images, send user images to server 1102, and generate digital human instructions.

[0051] After receiving the user's image and generating a digital human, server 1102 responds to the digital human generation command by generating an initial digital human based on the user's image. Using a standard digital human as a reference, the initial digital human is adjusted to match the standard digital human, resulting in an adjusted digital human. Target accessories corresponding to the target body parts are then assembled with the target body parts of the adjusted digital human to generate an assembled digital human. The target accessories are made based on the target body parts of the standard digital human.

[0052] After generating the assembled digital human, the server 1102 can send the assembled digital human to the computer device 1101 so that the computer device 1101 can display the assembled digital human to the user.

[0053] It is understood that the server can be deployed separately from the computer equipment, or the server can be integrated into the computer equipment. This application embodiment does not specifically limit this.

[0054] The following description, in conjunction with the accompanying drawings, introduces a digital human generation method provided by an embodiment of this application. (See attached drawings.) Figure 2 The figure is a flowchart of a digital human generation method provided in an embodiment of this application, including S201-S203.

[0055] S201. In response to the instruction to generate a digital human, an initial digital human is generated based on the user's image.

[0056] In this embodiment of the application, the user's image can be captured in real time or it can be a historically captured image retrieved from a database.

[0057] In one possible implementation, in response to a digital human generation instruction, a user's image can be input into a digital human generation model, which then generates an initial digital human with the user's features based on deep learning and image processing algorithms.

[0058] The initial digital human possesses the user's characteristics, but due to limitations in the shooting angle and distance of the user's image, as well as the complexity of some user features, the initial digital human suffers from inconsistent scale in some parts, making it difficult to directly match with accessories.

[0059] In order to equip accessories while ensuring the overall visual effect of the digital human, in this embodiment of the application, after generating the initial digital human, S202 can be executed.

[0060] S202. Using a standard digital human as a reference, adjust the initial digital human to match the standard digital human to obtain an adjusted digital human.

[0061] Among them, the standard digital human is a pre-constructed digital human whose scale and shape meet the current requirements.

[0062] In this embodiment, after obtaining an initial digital human, the initial digital human can be adjusted to match the standard digital human, based on the standard digital human, to obtain an adjusted digital human. The adjusted digital human has a high degree of consistency with the scale and shape of the standard digital human.

[0063] In one possible implementation, S202 can be achieved through steps 211-212.

[0064] Step 211: Using a standard digital human as a reference, adjust the scale of the initial digital human to match the scale of the standard digital human to obtain a scale-matched digital human.

[0065] In this embodiment of the application, when the scale of the initial digital human is adjusted to match the scale of the standard digital human based on a standard digital human, the minimum value of the energy function can be used as the optimization target to determine the scale parameters.

[0066] When the scale parameters include scaling factor, rotation matrix, and translation amount, the energy function can be expressed by equation (1) as follows:

[0067] (1)

[0068] In the formula, s is a scalar representing the scaling ratio; R is the rotation matrix, representing the rotation amount; t represents the translation amount; n is the number of index points; P i Represents the index point in the initial digital human; Q j The index point represents the standard digital human.

[0069] It is understood that the dimensions of the rotation matrix and translation are not specifically limited in the embodiments of this application. For example, R can be a 3×3 rotation matrix and t can be a three-dimensional vector.

[0070] In this embodiment, the energy function is used to indicate the distance between the index points of the initial digital human and the index points of the standard digital human after scaling, rotation, and translation. The scale parameter can be obtained by solving for the minimum value of the energy function, as shown in equation (2) below:

[0071] (2)

[0072] In this embodiment, when the energy function reaches its minimum value, the corresponding s, R, and t can be determined as the scaling ratio, rotation matrix, and translation amount for adjusting the initial digital human scale, respectively.

[0073] Based on this, using the scaling ratio, rotation matrix, and translation amount, the scale of the initial digital human can be adjusted to match the scale of the standard digital human, thus obtaining a scale-matched digital human.

[0074] In this embodiment, the scale-matched digital human has the same scale as the standard digital human, and most of their areas overlap. However, due to image quality issues, some parts of the scale-matched digital human may still differ from those of the standard digital human. Based on this, the differing local parts can be identified as target parts, and the target parts of the scale-matched digital human can be adjusted according to step 212.

[0075] Step 212: Based on the position information of the index points of the target parts of the digital human according to the standard, adjust the position information of the index points of the target parts of the scale-matched digital human to generate the adjusted digital human.

[0076] In this embodiment of the application, for the target part, the position information of the index points of the scale-matched digital human and the standard digital human that correspond to each other can be adjusted based on the position information of the index points of the standard digital human to generate the adjusted digital human.

[0077] For example, in this embodiment of the application, the target part can be adjusted based on the Laplacian mesh deformation algorithm, as shown in equation (3) below:

[0078] (3)

[0079] In the formula, the first part is Where L is a matrix and x is a vector. As a vector, the first part is used to constrain the index points of the scale-matched digital human and the standard digital human for the target part, where there is no corresponding relationship, and to maintain the original position information.

[0080] Part Two is Where ω is the weight parameter, x j c represents the index point corresponding to the scale-matched digital human. j This represents the index points corresponding to the standard digital human. The second part is used to adjust the position information of the index points of the scale-matched digital human and the standard digital human that correspond to the target area; that is, to adjust the position information of the index points of the scale-matched digital human to be consistent with the position information of the index points corresponding to the standard digital human.

[0081] See Figure 3 The figure is a schematic diagram of adjusting the initial digital human provided in an embodiment of this application. Taking the head as the target part, and using the standard digital human 3101 as a reference, the scale of the initial digital human 3102 is adjusted to match the scale of the standard digital human 3101 to obtain a scale-matched digital human 3103.

[0082] The scale of the scale-matched digital human 3103 is highly consistent with that of the standard digital human 3101, with only some differences in head details. Based on this, the position information of the index points of the head of the scale-matched digital human 3103 can be adjusted according to the position information of the index points of the head of the standard digital human to generate an adjusted digital human 3104.

[0083] For example, based on the position information of index point A of the head of the standard digital human, the position information of index point A' of the head of the scale-matched digital human 3103 is adjusted; based on the position information of index point B of the head of the standard digital human, the position information of index point B' of the head of the scale-matched digital human 3103 is adjusted; based on the position information of index point C of the head of the standard digital human, the position information of index point C' of the head of the scale-matched digital human 3103 is adjusted, thereby generating the adjusted digital human 3104.

[0084] It is understood that the number of index points is not specifically limited in the embodiments of this application, and the above is only used as an example to introduce the concept, with index points A, B, C, A', B', and C' as examples.

[0085] like Figure 3 As shown, the scale and shape of each part of the adjusted digital human 3104 are highly consistent with the scale and shape of the corresponding parts of the standard digital human 3101.

[0086] In one example, target parts can be preset. After obtaining a scale-matched digital human, the position information of the index points of the target parts of the scale-matched digital human can be adjusted based on the position information of the index points of the target parts of the standard digital human to generate an adjusted digital human.

[0087] It is understood that the target location is not specifically limited in the embodiments of this application, and can be set according to the actual situation.

[0088] In another example, in response to an adjustment command, the position information of the index points of the target parts of a scale-matched digital human can be adjusted based on the position information of the index points of the target parts of the standard digital human, thereby generating an adjusted digital human.

[0089] The adjustment command is used to indicate the target part corresponding to the target accessory.

[0090] For example, when the target accessory is glasses, the target area can be the eye and ear regions. In this case, in response to the adjustment command, the position information of the index points of the eye and ear regions of the scale-matched digital human can be adjusted based on the position information of the index points of the standard digital human's eye and ear regions to generate an adjusted digital human. When the target accessory is a mask, the target area can be the mouth region. In this case, in response to the adjustment command, the position information of the index points of the mouth region of the scale-matched digital human can be adjusted based on the position information of the index points of the standard digital human's mouth region to generate an adjusted digital human.

[0091] In the method provided in this application embodiment, the target part can be determined based on the adjustment command, and the scale-matched digital human can be adjusted more specifically, thereby improving the efficiency and flexibility of the adjustment.

[0092] In another possible implementation, S202 can be achieved through steps 221-222 when the initial digital human and the standard digital human are in different scale spaces.

[0093] Step 221: Convert the initial digital human and the standard digital human to the same scale space.

[0094] In this embodiment of the application, by converting the initial digital human and the standard digital human to the same scale space, it is convenient to adjust the initial digital human in the same coordinate system later.

[0095] Step 222: In the scale space, using a standard digital human as a reference, adjust the initial digital human to match the standard digital human to obtain the adjusted digital human.

[0096] In this embodiment of the application, in the scale space, using a standard digital human as a reference, the index points of the target parts of the initial digital human can be registered with the index points of the target parts of the standard digital human to obtain index point pairs.

[0097] For example, in this embodiment of the application, the index points of the target parts of the initial digital human can be registered with the index points of the target parts of the standard digital human using a retopology tool to obtain index point pairs. These index point pairs can be represented as (i, j), indicating a mapping relationship between the i-th index point of the initial digital human and the j-th index point of the standard digital human.

[0098] During the registration process between the index points of the initial digital human's target body parts and the index points of the standard digital human's target body parts, the target region of the initial digital human's target body parts can be labeled, and each index point P of the target region can be labeled. i The nearest neighbor search algorithm is used to find the nearest index point Q in the standard digital human. j As corresponding points, we obtain index point pairs.

[0099] For example, when the target part is the head, the target area of ​​the target part may include, but is not limited to, one or more parts such as the top of the head, forehead, ears, face, mouth, and chin.

[0100] In some examples, index point pairs can be saved as files for later use when adjusting the initial digital human.

[0101] In this embodiment of the application, after obtaining the index point pair, the initial digital human can be adjusted to match the standard digital human based on the index point pair to obtain the adjusted digital human.

[0102] S203. Assemble the target accessory corresponding to the target part with the adjusted digital human to generate the assembled digital human.

[0103] Among them, the target accessories are made based on the target parts of the standard digital human.

[0104] Taking the head as an example, the target accessory may include, but is not limited to, hair, glasses, hats, head-mounted rehabilitation equipment, surgical helmets, and goggles, or one or more of these.

[0105] Since the target accessory is made based on the target parts of a standard digital human, it matches the target parts of the standard digital human. In the method provided in this application embodiment, the initial size and shape of the digital human have been adjusted to be consistent with the size and shape of the standard digital human through S201-S202. In this case, the target accessory corresponding to the target part can be directly assembled with the adjusted target part of the digital human to generate the assembled digital human.

[0106] The method provided in this application can be applied to scenarios such as virtual social platforms, AR try-on technology, virtual character creation, virtual rehabilitation training, and medical teaching and surgical simulation.

[0107] Taking the head as an example again, when a user creates a personalized digital human avatar on a virtual social platform, the method provided in this application embodiment can quickly generate a digital human avatar with accessories such as hair and glasses that match the digital human head, thereby improving the realism of the digital human avatar and enhancing the user's interactive experience on the virtual social platform.

[0108] When a user tries on virtual glasses, hats, or other head accessories using augmented reality (AR) technology, the method provided in this application embodiment can accurately fit the virtual target accessory onto the corresponding part of the user's digital persona, providing the user with a more realistic and natural try-on effect and enhancing the user's experience.

[0109] In film, television series, or animation production, the method provided by the embodiments of this application can quickly generate high-quality digital humans while ensuring that accessories such as hair and glasses match the digital human's head, reducing the workload and cost of post-production special effects and improving the efficiency of digital human generation.

[0110] In virtual rehabilitation training, a user's digital human is created based on the method provided in the embodiments of this application. Various rehabilitation assistive devices (such as head-mounted rehabilitation devices) can be equipped on the digital human to provide the user with a more realistic and personalized virtual rehabilitation training experience.

[0111] In medical teaching and surgical simulation scenarios, the digital human generated based on the method provided in this application has a high degree of matching between its head and various medical devices (such as surgical helmets, goggles, etc.), which can provide medical students with a more realistic operating experience and improve learning outcomes.

[0112] It is understood that the method provided in this application embodiment can be applied to various scenarios that require the generation of digital humans, and the above is only an example.

[0113] To facilitate understanding, the following example uses the head as the target area, and combines it with... Figure 4 The methods provided in the embodiments of this application will be described in general.

[0114] See Figure 4 The figure is a schematic diagram of a digital human generation method provided in an embodiment of this application. In this embodiment, a standard digital human is generated in advance, and target accessories such as hair and glasses are made based on the head of the standard digital human.

[0115] When it is necessary to generate a digital avatar of a user, an image of the user can be obtained by capturing the user's image or retrieving the user's image from a database. In response to the command to generate a digital avatar, the user's image can be input into the digital avatar generation model to generate an initial digital avatar.

[0116] Due to factors such as image quality and the complexity of head features, it is difficult to directly match the target head accessory with the initial digital human's head. Therefore, in this embodiment, by using a standard digital human as a benchmark, scale matching and head deformation processing are performed on the initial digital human to obtain an adjusted digital human.

[0117] For example, to perform scale matching and head deformation processing on an initial digital human, the initial digital human and a standard digital human can first be converted to the same scale space. In this scale space, the index points of the initial digital human's head are registered with the index points of the standard digital human's head using a retopology tool to obtain index point pairs. After obtaining the index point pairs, based on these pairs and using the standard digital human as a reference, scale matching and head deformation processing are performed on the initial digital human to obtain the adjusted digital human.

[0118] It is understood that the scale matching and head deformation processing processes in this embodiment are the same as those in S202 above, and will not be repeated here.

[0119] After obtaining the adjusted digital human, you can call the target accessory corresponding to the standard digital human and equip the target accessory to the corresponding position on the head of the adjusted digital human to generate the assembled digital human.

[0120] In summary, this embodiment uses a standard digital human as a benchmark, adjusting the initial digital human to match the standard digital human to obtain an adjusted digital human. Since the adjusted digital human maintains the same scale and shape as the standard digital human, and the target accessories are made based on the target parts of the standard digital human, the target accessories corresponding to the target parts can be directly assembled with the target parts of the adjusted digital human. This effectively improves the generation efficiency of the assembled digital human while ensuring the realism and harmony of the overall image.

[0121] Meanwhile, the method provided in this application embodiment can adjust the new target accessory or the original target accessory based on the standard digital human when it is necessary to add or update the target accessory, without the need for complicated adaptation operations, thus improving the flexibility and scalability of the generated assembled digital human.

[0122] This application provides a digital human generation device, see [link]. Figure 5 The figure is a schematic diagram of the structure of a digital human generation device provided in an embodiment of this application. Its specific implementation method is consistent with the implementation method and the technical effect achieved in the above-described method embodiment, and some contents will not be repeated.

[0123] A digital human generation device 5100 includes:

[0124] The unit includes a generation unit 5101, an adjustment unit 5102, and an assembly unit 5103.

[0125] The generation unit 5101 is configured to generate an initial digital human based on the user's image in response to a digital human generation command;

[0126] The adjustment unit 5102 is used to adjust the initial digital human to match the standard digital human based on the standard digital human, so as to obtain the adjusted digital human; the scale and shape of the adjusted digital human are consistent with the scale and shape of the standard digital human.

[0127] The assembly unit 5103 is used to assemble the target accessory corresponding to the target part with the adjusted target part of the digital human to generate the assembled digital human. The target accessory is made based on the target part of the standard digital human.

[0128] In one possible implementation, the adjustment unit is specifically used for:

[0129] Using a standard digital human as a benchmark, the scale of the initial digital human is adjusted to match the scale of the standard digital human to obtain a scale-matched digital human; based on the position information of the index points of the target parts of the standard digital human, the position information of the index points of the target parts of the scale-matched digital human is adjusted to generate an adjusted digital human.

[0130] In one possible implementation, the adjustment unit is specifically used for:

[0131] For the target area, based on the position information of the index points of the standard digital human, the position information of the index points of the scale-matched digital human that correspond to the standard digital human is adjusted to generate an adjusted digital human.

[0132] In one possible implementation, the adjustment unit is specifically used for:

[0133] The scaling ratio, rotation matrix, and translation amount are determined with the goal of minimizing the energy function; the energy function is used to indicate the distance between the index points of the initial digital human and the index points of the standard digital human after scaling, rotation, and translation.

[0134] Based on the scaling ratio, rotation matrix, and translation amount, the scale of the initial digital human is adjusted to match the scale of the standard digital human, thus obtaining a scale-matched digital human.

[0135] In one possible implementation, the adjustment unit is specifically used for:

[0136] In response to the adjustment command, the position information of the index points of the target parts of the digital human with the standard is adjusted based on the position information of the index points of the target parts of the digital human with the scale matching, and the adjusted digital human is generated. The adjustment command is used to indicate the target parts corresponding to the target accessories.

[0137] In one possible implementation, the initial digital human and the standard digital human exist in different scale spaces, and the adjustment unit is specifically used for:

[0138] Transform the initial digital human and the standard digital human into the same scale space;

[0139] In the scale space, using a standard digital human as a reference, the initial digital human is adjusted to match the standard digital human to obtain an adjusted digital human.

[0140] In one possible implementation, the adjustment unit is specifically used for:

[0141] In the scale space, using a standard digital human as a reference, the index points of the target parts of the initial digital human are registered with the index points of the target parts of the standard digital human to obtain index point pairs.

[0142] Based on the index point pair, the initial digital human is adjusted to match the standard digital human to obtain the adjusted digital human.

[0143] In one possible implementation, the target body part is the head; the target accessory includes one or more of hair, glasses, hats, head-mounted rehabilitation devices, surgical helmets, and goggles.

[0144] In summary, the apparatus provided in this application uses a standard digital human as a reference, adjusting the initial digital human to match the standard digital human to obtain an adjusted digital human. Since the adjusted digital human maintains the same scale and shape as the standard digital human, and the target accessory is made based on the target parts of the standard digital human, the target accessory corresponding to the target part can be directly assembled with the target part of the adjusted digital human. This effectively improves the generation efficiency of the assembled digital human while ensuring the realism and harmony of the overall image of the assembled digital human.

[0145] Meanwhile, the device provided in this application embodiment only needs to adjust the new target accessory or the original target accessory when it is necessary to add or update the target accessory, without the need for complicated adaptation operations, which improves the flexibility and scalability of the generated assembled digital human.

[0146] This application provides a computer device, such as... Figure 6 As shown in the figure, this figure is a schematic diagram of the structure of a computer device provided in an embodiment of this application. The computer device 6100 includes a processor 6101 and a memory 6102.

[0147] The memory 6102 is used to store program code and transmit the program code to the processor 6101;

[0148] The processor 6101 is used to execute the steps of a digital human generation method as described above according to the instructions in the program code.

[0149] This application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of a digital human generation method as described above.

[0150] This application provides a computer program product that, when run on a computer, executes the steps of a digital human generation method as described above.

[0151] The above description is merely one specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for generating a digital human, characterized in that, include: In response to the instruction to generate a digital human, an initial digital human is generated based on the user's image; Using a standard digital human as a benchmark, the initial digital human is adjusted to match the standard digital human to obtain an adjusted digital human; The scale and shape of the adjusted digital human are consistent with the scale and shape of the standard digital human; the standard digital human is pre-constructed. The target accessory corresponding to the target part is assembled with the adjusted target part of the digital human to generate the assembled digital human. The target accessory is made based on the target part of the standard digital human.

2. The method according to claim 1, characterized in that, The process of adjusting the initial digital human to match the standard digital human, based on a standard digital human, to obtain an adjusted digital human includes: Using a standard digital human as a benchmark, the scale of the initial digital human is adjusted to match the scale of the standard digital human to obtain a scale-matched digital human; Based on the position information of the index points of the target parts of the digital human according to the standard, the position information of the index points of the target parts of the scale-matched digital human is adjusted to generate the adjusted digital human.

3. The method according to claim 2, characterized in that, The location information of the index points of the target parts of the digital human based on the standard is used to adjust the location information of the index points of the target parts of the scale-matched digital human to generate an adjusted digital human, including: For the target area, based on the position information of the index points of the standard digital human, the position information of the index points of the scale-matched digital human that correspond to the standard digital human is adjusted to generate an adjusted digital human.

4. The method according to claim 2, characterized in that, The process of adjusting the scale of the initial digital human to match the scale of the standard digital human, based on a standard digital human, to obtain a scale-matched digital human includes: The scaling ratio, rotation matrix, and translation amount are determined with the goal of minimizing the energy function; the energy function is used to indicate the distance between the index points of the initial digital human and the index points of the standard digital human after scaling, rotation, and translation. Based on the scaling ratio, rotation matrix, and translation amount, the scale of the initial digital human is adjusted to match the scale of the standard digital human, thus obtaining a scale-matched digital human.

5. The method according to claim 2, characterized in that, The location information of the index points of the target parts of the digital human based on the standard is used to adjust the location information of the index points of the target parts of the scale-matched digital human to generate an adjusted digital human, including: In response to the adjustment command, the position information of the index points of the target parts of the digital human with the standard is adjusted based on the position information of the index points of the target parts of the digital human with the scale matching, and the adjusted digital human is generated. The adjustment command is used to indicate the target parts corresponding to the target accessories.

6. The method according to claim 1, characterized in that, The initial digital human and the standard digital human exist in different scale spaces. The process of adjusting the initial digital human to match the standard digital human, based on the standard digital human, to obtain the adjusted digital human includes: Transform the initial digital human and the standard digital human into the same scale space; In the scale space, using a standard digital human as a reference, the initial digital human is adjusted to match the standard digital human to obtain an adjusted digital human.

7. The method according to claim 6, characterized in that, In the scale space, using a standard digital human as a reference, the initial digital human is adjusted to match the standard digital human to obtain an adjusted digital human, including: In the scale space, using a standard digital human as a reference, the index points of the target parts of the initial digital human are registered with the index points of the target parts of the standard digital human to obtain index point pairs. Based on the index point pair, the initial digital human is adjusted to match the standard digital human to obtain the adjusted digital human.

8. The method according to any one of claims 1-7, characterized in that, The target body part is the head; the target accessories include one or more of the following: hair, glasses, hats, head-mounted rehabilitation devices, surgical helmets, and goggles.

9. A digital human generation device, characterized in that, include: Production unit, adjustment unit, and assembly unit; The generation unit is configured to generate an initial digital human based on the user's image in response to a digital human generation command; The adjustment unit is used to adjust the initial digital human to match the standard digital human based on the standard digital human, so as to obtain the adjusted digital human; The scale and shape of the adjusted digital human are consistent with the scale and shape of the standard digital human; the standard digital human is pre-constructed. The assembly unit is used to assemble the target accessory corresponding to the target part with the adjusted target part of the digital human to generate the assembled digital human. The target accessory is made based on the target part of the standard digital human.

10. A computer device, characterized in that, The computer device includes: a processor and a memory; The memory is used to store program code and transmit the program code to the processor; The processor is configured to execute the steps of a digital human generation method as described in any one of claims 1-8 according to instructions in the program code.

Images