CN114708374A - Virtual image generation method, device, electronic device and storage medium - Google Patents

Abstract

The present disclosure provides a virtual image generation method, which relates to the technical field of artificial intelligence, and in particular, to the technical field of computer vision and virtual/augmented reality technology. The specific implementation scheme is: determining the style feature map of the target object in the target image; determining a plurality of neural radiation field vectors according to the style feature map; and generating a virtual image corresponding to the target object according to the plurality of neural radiation field vectors. The present disclosure also provides an apparatus for generating a virtual image, an electronic device and a storage medium.

Description

Virtual image generation method, device, electronic device and storage medium

technical field

The present disclosure relates to the field of artificial intelligence technology, in particular to the field of computer vision technology and virtual/augmented reality technology, and can be applied to image processing scenarios. More specifically, the present disclosure provides an avatar generation method, apparatus, electronic device and storage medium.

Background technique

With the development of artificial intelligence technology, in fields such as virtual/augmented reality, deep learning models are widely used in image processing or image generation. In addition, avatars have a wide range of applications in social, live or gaming scenarios. The avatar can be generated in a human-based manner.

SUMMARY OF THE INVENTION

The present disclosure provides a virtual image generation method, apparatus, device and storage medium.

According to an aspect of the present disclosure, there is provided a method for generating an avatar, the method comprising: determining a style feature map of a target object in a target image; determining a plurality of neural radiation field vectors according to the style feature map; and according to the A plurality of neural radiation field vectors are used to generate a virtual image corresponding to the target object.

According to another aspect of the present disclosure, there is provided an apparatus for generating an avatar, the apparatus comprising: a first determining module for determining a style feature map of a target object in a target image; a second determining module for determining according to the style a feature map, for determining a plurality of neural radiation field vectors; and a generating module for generating a virtual image corresponding to the target object according to the plurality of neural radiation field vectors.

According to another aspect of the present disclosure, there is provided an electronic device comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor. The at least one processor executes to enable the at least one processor to perform the methods provided in accordance with the present disclosure.

According to another aspect of the present disclosure, there is provided a non-transitory computer-readable storage medium storing computer instructions for causing a computer to perform a method provided in accordance with the present disclosure.

According to another aspect of the present disclosure, there is provided a computer program product comprising a computer program which, when executed by a processor, implements the method provided according to the present disclosure.

It should be understood that what is described in this section is not intended to identify key or critical features of embodiments of the disclosure, nor is it intended to limit the scope of the disclosure. Other features of the present disclosure will become readily understood from the following description.

Description of drawings

The accompanying drawings are used for better understanding of the present solution, and do not constitute a limitation to the present disclosure. in:

FIG. 1 is a schematic diagram of an exemplary system architecture to which the avatar generation method and apparatus can be applied according to an embodiment of the present disclosure;

2 is a flowchart of a method for generating an avatar according to an embodiment of the present disclosure;

3 is a flowchart of a method for generating an avatar according to another embodiment of the present disclosure;

4 is a schematic diagram of a method for generating an avatar according to an embodiment of the present disclosure;

5 is a schematic diagram of a method for generating an avatar according to an embodiment of the present disclosure;

FIG. 6 is a block diagram of an avatar generating apparatus according to an embodiment of the present disclosure; and

FIG. 7 is a block diagram of an electronic device to which the avatar generation method can be applied, according to one embodiment of the present disclosure.

Detailed ways

Exemplary embodiments of the present disclosure are described below with reference to the accompanying drawings, which include various details of the embodiments of the present disclosure to facilitate understanding and should be considered as exemplary only. Accordingly, those of ordinary skill in the art will recognize that various changes and modifications of the embodiments described herein can be made without departing from the scope and spirit of the present disclosure. Also, descriptions of well-known functions and constructions are omitted from the following description for clarity and conciseness.

The avatar may include a virtual body. A virtual image can be designed, generated and optimized in a manual way, but it requires high time and labor costs. The interpretation of styles by different designers is subjective, resulting in multiple avatars that may be completely different based on the same style. Also, there may be a mismatch between the face and hair of the artificially generated avatar.

FIG. 1 is a schematic diagram of an exemplary system architecture to which a method and apparatus for generating an avatar can be applied according to an embodiment of the present disclosure. It should be noted that FIG. 1 is only an example of a system architecture to which the embodiments of the present disclosure can be applied, so as to help those skilled in the art to understand the technical content of the present disclosure, but it does not mean that the embodiments of the present disclosure cannot be used for other A device, system, environment or scene.

As shown in FIG. 1 , the system architecture 100 according to this embodiment may include terminal devices 101 , 102 , and 103 , a network 104 and a server 105 . The network 104 is a medium used to provide a communication link between the terminal devices 101 , 102 , 103 and the server 105 . The network 104 may include various connection types, such as wired and/or wireless communication links, and the like.

The user can use the terminal devices 101, 102, 103 to interact with the server 105 through the network 104 to receive or send messages and the like. The terminal devices 101, 102, 103 may be various electronic devices having a display screen and supporting web browsing, including but not limited to smart phones, tablet computers, laptop computers, desktop computers, and the like.

The server 105 may be a server that provides various services, such as a background management server (just an example) that provides support for websites browsed by users using the terminal devices 101, 102, 103. The background management server can analyze and process the received user requests and other data, and feed back the processing results (such as web pages, information, or data obtained or generated according to user requests) to the terminal device.

It should be noted that, the virtual image generation method provided by the embodiment of the present disclosure may generally be executed by the server 105 . Correspondingly, the virtual image generating apparatus provided by the embodiment of the present disclosure may generally be provided in the server 105 . The avatar generation method provided by the embodiment of the present disclosure may also be executed by a server or server cluster different from the server 105 and capable of communicating with the terminal devices 101 , 102 , 103 and/or the server 105 . Correspondingly, the avatar generating apparatus provided by the embodiment of the present disclosure may also be set in a server or server cluster that is different from the server 105 and can communicate with the terminal devices 101 , 102 , 103 and/or the server 105 .

FIG. 2 is a flowchart of a method for generating an avatar according to an embodiment of the present disclosure.

As shown in FIG. 2, the method 200 may include operations S210 to S230.

In operation S210, a style feature map of the target object in the target image is determined.

For example, the target image can contain the target object and the image background. The target image can be processed to remove the background of the image to obtain the processed target image. In one example, a semantic segmentation model can be used to determine the region where the image background is located, and then the image background can be removed.

For another example, multiple deep learning models may be used to process the processed target image to obtain a style feature map. In one example, the plurality of deep learning models may include, for example, a U-Net (U-shaped network) model and a StyleGAN2 (Styled Adversarial Generative Network) model.

In operation S220, a plurality of neural radiation field vectors are determined according to the style feature map.

For example, a NERF (NEural Radiance Field, neural radiation field) model can be used to process the style feature map to obtain multiple vectors as multiple neural radiation field vectors.

In operation S230, an avatar corresponding to the target object is generated according to the plurality of neural radiation field vectors.

For example, image rendering can be performed according to a plurality of neural radiation field vectors to obtain an avatar corresponding to the target object.

With the embodiments of the present disclosure, an avatar with a corresponding style can be generated according to an image provided by a user.

FIG. 3 is a flowchart of a method for generating an avatar according to another embodiment of the present disclosure.

As shown in FIG. 3 , the method 300 may include operations S311 to S312 , operations S321 to S324 , and operations S331 to S332 .

In this embodiment of the present disclosure, the method 300 may determine the style feature map of the target object in the target image, which will be described in detail below with reference to operations S311 to S312.

In operation S311, style encoding data of the target object is determined.

For example, the target object includes a head.

For example, as described above, the image background in the target image can be removed to obtain the processed target image. Next, the processed target image can be processed again by using the U-Net model to obtain the style encoding data of the target object.

In operation S312, a style feature map is determined according to the style encoding data.

For example, the StyleGAN2 model can be used to process style-encoded data to determine style feature maps.

In the embodiment of the present disclosure, the method 300 may also determine a plurality of neural radiation field vectors according to the style feature map. Operations S321 to S324 will be combined below.

In operation S321, a style image is generated according to the style feature map.

For example, a CNN (Convolutional Neural Network, Convolutional Neural Network) can be used to process the style feature map described above to generate a style image. In one example, corresponding convolution kernels can be selected to generate style images from style feature maps.

In operation S322, reconstruction parameters are determined according to the style image.

For example, 3DMM (3D Morphable Models, 3D Morphable Models) can be used to determine the reconstruction parameters according to the style image described above. In one example, the reconstruction parameters may include, for example, shape parameters, expression parameters, texture parameters, camera intrinsic parameters, camera extrinsic parameters, and the like.

In operation S323, the style feature map is fused with the reconstruction parameters to obtain a fused feature map.

For example, the first FC (Fully Connected Neural Network) model can be used to process the style feature map to obtain a style feature information.

For another example, the reconstruction parameters may be processed by using the second FC model to obtain reconstruction feature information.

Next, the style feature information and the reconstructed feature information can be fused to obtain a fused feature map.

In operation S324, a plurality of neural radiation field vectors are determined according to the fusion feature map.

For example, the fused feature maps can be processed using the NERF model to determine multiple neural radiation field vectors. In one example, the NERF model includes multiple fully connected layers.

In the embodiment of the present disclosure, the method 300 may further perform image rendering according to a plurality of neural radiation field vectors to generate a virtual image corresponding to the target image. A detailed description will be made below in conjunction with operations S331 to S332.

In operation S331, an explicit three-dimensional representation of the target object is obtained according to the plurality of neural radiation field vectors.

For example, based on the MC (Marching Cub) algorithm, an explicit three-dimensional representation of the target object can be obtained according to multiple neural radiation field vectors. In one example, the plurality of neural radiation field vectors determined using the NERF model of an embodiment of the present disclosure can be regarded as an implicit three-dimensional representation. Based on the MC algorithm, an explicit three-dimensional representation can be obtained.

In operation S332, the explicit three-dimensional representation is rendered to generate an avatar corresponding to the target object.

For example, an explicit three-dimensional representation can be rendered using the Pytorch3D renderer described above to obtain an avatar.

Through the embodiments of the present disclosure, the reconstruction parameters conforming to the style of the target object can be accurately determined by using the 3DMM. In addition, by fusing the reconstruction parameters with the style feature map, the information output by the U-Net model and the information output by the 3DMM can be fully utilized.

With the embodiments of the present disclosure, a great degree of freedom can be provided for users in terms of stylization, personalization, and demand differentiation, and the user experience is improved. Embodiments of the present disclosure can generate three-dimensional hair outside of facial skin and skin areas. The three-dimensional hair can be matched to the facial style.

It should be noted that the virtual image generation method provided by the embodiment of the present disclosure can be applied to a metaverse scene, and can also be applied to other scenes. Through the embodiments of the present disclosure, a complete end-to-end virtual image generation method can be provided for users.

In some embodiments, the NERF model may include an FC model. According to the style feature map or the fusion feature map, the position information and view direction information of the voxel can be determined as the input vector of the NERF model. The neural radiation field vector output by the NERF model may include, for example, color information and voxel density information. The voxel density information is related to where the voxels are located. The color information is related to the position of the voxel and the viewing direction.

In some embodiments, the MC algorithm may traverse each voxel in the input data field one by one to determine whether the voxel intersects the isosurface. In the case where a voxel intersects an isosurface, calculate the intersections of the isosurface with the edges of a cube, and use these intersections to form triangular patches to approximate the isosurface within the voxel. After all voxels have been traversed, the isosurface of the entire data field can be obtained. In one example, the input data field may be derived from the plurality of neural radiation field vectors described above.

In some embodiments, the 3DMM can use facial shape and facial texture as constraints, while taking into account the influence of mid-face pose and external lighting, can determine high-precision reconstruction parameters. The 3DMM can determine initialization parameters based on the style map in order to determine a 3D initialization avatar. From the initialized avatar, a 2D projection image can be determined. Calculate the difference between the projected image and the style map, and adjust the initialization parameters described above according to the difference. After the difference value is smaller than the preset difference threshold, the reconstruction parameters may be determined.

In some embodiments, the Pytorch3D renderer is a component that provides efficient, reusable functionality for 3D computer vision research. The Pytorch3D renderer can store and manipulate the data structure of the triangle mesh, and can also perform efficient operations (such as projection transformation, graph convolution, etc.) based on the data structure of the triangle mesh. Also, the Pytorch3D renderer is a differentiable mesh renderer.

FIG. 4 is a schematic diagram of a method for generating an avatar according to another embodiment of the present disclosure.

As shown in FIG. 4 , an avatar 402 can be generated according to the target image 401 by using the avatar generation model 400 .

For example, the avatar generation model 400 includes a U-Net model 410 , a StyleGAN2 model 420 , a CNN model 430 , a 3DMM 440 , an FC model 450 , an FC model 460 , a NERF model 470 , and a Pytorch3D renderer 480 .

For example, the target object in the target image 401 includes a head. In one example, the image background in the target image 401 may be removed to obtain the processed target image. Next, the U-Net model 410 can process the processed target image again to obtain the style encoding data D_s of the target object.

The StyleGAN2 model 420 may process the style encoding data D_s to determine the style feature map FM_s.

The CNN model 430 can process the style feature map FM_s to generate the style image Img_s.

The 3DMM 440 may determine the reconstruction parameter Para according to the style image Img_s. In an example, the reconstruction parameter Para may include, for example, a shape parameter, an expression parameter, a texture parameter, an internal camera parameter, an external camera parameter, and the like.

The FC model 450 can process the style feature map FM_s, and the FC model 460 can process the reconstruction parameter Para.

The output of the FC model 450 and the output of the FC model 460 may be fused to obtain a fused feature map FM_m.

The NERF model 470 can process the fused feature maps to determine multiple neural radiation field vectors. In one example, the NERF model includes multiple fully connected layers.

Next, based on the MC (Marching Cub) algorithm, an explicit three-dimensional representation of the target object can be obtained according to multiple neural radiation field vectors.

The Pytorch3D renderer 480 can render the explicit three-dimensional representation, resulting in the avatar 402.

In some embodiments, the method described above further includes: inputting the sample image into the avatar generation model to obtain an avatar corresponding to the sample object in the sample image; training the avatar generation model according to the label of the sample image and the avatar . A detailed description will be given below with reference to FIG. 5 .

FIG. 5 is a schematic diagram of a method for generating an avatar according to another embodiment of the present disclosure.

As shown in FIG. 5 , an avatar 502 can be generated from a sample image 501 using an avatar generation model 500 .

For example, avatar generation model 500 includes U-Net model 510, StyleGAN2 model 520, CNN model 530, 3DMM 540, FC model 550, FC model 560, NERF model 570, and Pytorch3D renderer 580. It can be understood that the above detailed descriptions of the U-Net model 410, the StyleGAN2 model 420, the CNN model 430, the 3DMM 440, the FC model 450, the FC model 460, the NERF model 470 and the Pytorch3D renderer 480 are also applicable to this embodiment. The disclosure will not be repeated again.

After the Pytorch3D renderer 580 generates the avatar 502 corresponding to the sample object in the sample image 501 , the loss value 504 can be determined according to the avatar 502 and the label 503 . The parameters of the avatar generation model 500 may be adjusted according to the loss value 504 . For example, the parameters of one or more of the U-Net model 510, StyleGAN2 model 520, CNN model 530, 3DMM 540, FC model 550, FC model 560, NERF model 570 and Pytorch3D renderer 580 may be adjusted according to the loss value 504, Make the loss value 504 less than or equal to the preset loss threshold. In one example, after the loss value is less than or equal to a preset loss threshold, a trained avatar generation model can be obtained as the avatar generation model 400 described above.

For example, the label of the sample image 501 may be manually determined. Multiple alternative avatars may be designed by multiple different avatar designers according to the sample objects in the sample image. Then, multiple designers evaluate multiple candidate avatars respectively, and use the avatar with the highest evaluation as the label of the sample image 501 .

For example, the loss value 504 may be determined using an L1 loss function or an L2 loss function.

FIG. 6 is a block diagram of an avatar generating apparatus according to an embodiment of the present disclosure.

As shown in FIG. 6 , the apparatus 600 may include a first determining module 610 , a second determining module 620 and a generating module 630 .

The first determining module 610 is configured to determine the style feature map of the target object in the target image. In an example, the first determining module 610 may be configured to perform operation S210, which will not be repeated in the present disclosure.

The second determining module 620 is configured to determine a plurality of neural radiation field vectors according to the style feature map. In one example, the second determining module 620 may be configured to perform operation S220, which will not be repeated in the present disclosure.

The generating module 630 is configured to generate a virtual image corresponding to the target object according to the plurality of neural radiation field vectors. In one example, the generating module 630 may be configured to perform operation S230, which will not be repeated in the present disclosure.

In some embodiments, the first determining module includes: a first determining unit for determining style encoding data of the target object; and a second determining unit for determining the style according to the style encoding data feature map. In one example, the first determination unit may be configured to perform operation S311. In one example, the second determining unit may be configured to perform operation S312, which will not be repeated in the present disclosure.

In some embodiments, the second determining module includes: a first generating unit for generating a style image according to the style feature map; a third determining unit for determining reconstruction parameters according to the style image; fusion a unit for merging the style feature map with the reconstruction parameters to obtain a fused feature map; and a fourth determining unit for determining a plurality of neural radiation field vectors according to the fused feature map. In one example, the first generating unit may be configured to perform operation S321. In one example, the third determination unit may be configured to perform operation S322. In one example, the fusion unit may be used to perform operation S323. In one example, the fourth determining unit may be configured to perform operation S324, which will not be repeated in the present disclosure.

In some embodiments, the generating module includes: an obtaining unit for obtaining an explicit three-dimensional representation of the target object according to the plurality of neural radiation field vectors; and a second generating unit for converting the explicit three-dimensional representation of the target object. The three-dimensional expression is rendered in the form of a three-dimensional representation, and an avatar corresponding to the target object is generated. In one example, the obtaining unit may be used to perform operation S331. In one example, the second generating unit may be configured to perform operation S332, which will not be repeated in this disclosure.

In some embodiments, the target object includes a head.

In the technical solutions of the present disclosure, the collection, storage, use, processing, transmission, provision, and disclosure of the user's personal information involved are all in compliance with relevant laws and regulations, and do not violate public order and good customs.

According to embodiments of the present disclosure, the present disclosure also provides an electronic device, a readable storage medium, and a computer program product.

In an embodiment of the present disclosure, an electronic device is provided, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores data executable by the at least one processor The instructions are executed by the at least one processor to enable the at least one processor to perform the methods provided by the present disclosure.

In an embodiment of the present disclosure, a non-transitory computer-readable storage medium storing computer instructions is provided, wherein the computer instructions are used to cause the computer to perform the method provided by the present disclosure.

In an embodiment of the present disclosure, a computer program product is provided, including a computer program, and the computer program implements the method provided by the present disclosure when executed by a processor.

FIG. 7 shows a schematic block diagram of an example electronic device 700 that may be used to implement embodiments of the present disclosure. Electronic devices are intended to represent various forms of digital computers, such as laptops, desktops, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. Electronic devices may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smart phones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are by way of example only, and are not intended to limit implementations of the disclosure described and/or claimed herein.

As shown in FIG. 7 , the device 700 includes a computing unit 701 that can be executed according to a computer program stored in a read only memory (ROM) 702 or loaded into a random access memory (RAM) 703 from a storage unit 708 Various appropriate actions and handling. In the RAM 703, various programs and data necessary for the operation of the device 700 can also be stored. The computing unit 701, the ROM 702, and the RAM 703 are connected to each other through a bus 704. An input/output (I/O) interface 705 is also connected to bus 704 .

Various components in the device 700 are connected to the I/O interface 705, including: an input unit 706, such as a keyboard, mouse, etc.; an output unit 707, such as various types of displays, speakers, etc.; a storage unit 708, such as a magnetic disk, an optical disk, etc. ; and a communication unit 709, such as a network card, a modem, a wireless communication transceiver, and the like. The communication unit 709 allows the device 700 to exchange information/data with other devices through a computer network such as the Internet and/or various telecommunication networks.

Computing unit 701 may be various general-purpose and/or special-purpose processing components with processing and computing capabilities. Some examples of computing units 701 include, but are not limited to, central processing units (CPUs), graphics processing units (GPUs), various specialized artificial intelligence (AI) computing chips, various computing units that run machine learning model algorithms, digital signal processing processor (DSP), and any suitable processor, controller, microcontroller, etc. The computing unit 701 executes the various methods and processes described above, such as the avatar generation method. For example, in some embodiments, the avatar generation method may be implemented as a computer software program tangibly embodied on a machine-readable medium, such as storage unit 708 . In some embodiments, part or all of the computer program may be loaded and/or installed on device 700 via ROM 702 and/or communication unit 709 . When the computer program is loaded into RAM 703 and executed by computing unit 701, one or more steps of the avatar generation method described above may be performed. Alternatively, in other embodiments, the computing unit 701 may be configured to perform the avatar generation method by any other suitable means (eg, by means of firmware).

Various implementations of the systems and techniques described herein above may be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips system (SOC), complex programmable logic device (CPLD), computer hardware, firmware, software, and/or combinations thereof. These various embodiments may include being implemented in one or more computer programs executable and/or interpretable on a programmable system including at least one programmable processor The processor, which may be a special purpose or general-purpose programmable processor, may receive data and instructions from a storage system, at least one input device, and at least one output device, and transmit data and instructions to the storage system, the at least one input device, and the at least one output device an output device.

Program code for implementing the methods of the present disclosure may be written in any combination of one or more programming languages. These program codes may be provided to a processor or controller of a general purpose computer, special purpose computer or other programmable data processing apparatus, such that the program code, when executed by the processor or controller, performs the functions/functions specified in the flowcharts and/or block diagrams. Action is implemented. The program code may execute entirely on the machine, partly on the machine, partly on the machine and partly on a remote machine as a stand-alone software package or entirely on the remote machine or server.

In the context of the present disclosure, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in connection with the instruction execution system, apparatus or device. The machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. Machine-readable media may include, but are not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or devices, or any suitable combination of the foregoing. More specific examples of machine-readable storage media would include one or more wire-based electrical connections, portable computer disks, hard disks, random access memory (RAM), read only memory (ROM), erasable programmable read only memory (EPROM or flash memory), fiber optics, compact disk read only memory (CD-ROM), optical storage, magnetic storage, or any suitable combination of the foregoing.

To provide interaction with a user, the systems and techniques described herein may be implemented on a computer having a display device (eg, a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user ); and a keyboard and pointing device (eg, a mouse or trackball) through which a user can provide input to the computer. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (eg, visual feedback, auditory feedback, or tactile feedback); and can be in any form (including audio input, voice input, or tactile input) to receive input from the user.

The systems and techniques described herein may be implemented on a computing system that includes back-end components (eg, as a data server), or a computing system that includes middleware components (eg, an application server), or a computing system that includes front-end components (eg, a user's computer having a graphical user interface or web browser through which a user may interact with implementations of the systems and techniques described herein), or including such backend components, middleware components, Or any combination of front-end components in a computing system. The components of the system may be interconnected by any form or medium of digital data communication (eg, a communication network). Examples of communication networks include: Local Area Networks (LANs), Wide Area Networks (WANs), and the Internet.

A computer system can include clients and servers. Clients and servers are generally remote from each other and usually interact through a communication network. The relationship of client and server arises by computer programs running on the respective computers and having a client-server relationship to each other.

It should be understood that steps may be reordered, added or deleted using the various forms of flow shown above. For example, the steps described in the present disclosure can be executed in parallel, sequentially, or in different orders. As long as the desired results of the technical solutions disclosed in the present disclosure can be achieved, no limitation is imposed herein.

The above-mentioned specific embodiments do not constitute a limitation on the protection scope of the present disclosure. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present disclosure should be included within the protection scope of the present disclosure.

Claims (13)

1. An avatar generation method, comprising:

determining a style characteristic diagram of a target object in a target image;

determining a plurality of nerve radiation field vectors according to the style characteristic diagram; and

and generating an avatar corresponding to the target object according to the plurality of nerve radiation field vectors.

2. The method of claim 1, wherein the determining a style profile of a target object in a target image comprises:

determining style encoding data of the target object; and

and determining the style characteristic graph according to the style encoding data.

3. The method of claim 1, wherein the determining a plurality of nerve radiation field vectors from the style feature map comprises:

generating a style image according to the style characteristic diagram;

determining reconstruction parameters according to the style image;

fusing the style characteristic diagram with the reconstruction parameters to obtain a fused characteristic diagram; and

and determining a plurality of nerve radiation field vectors according to the fused feature map.

4. The method of claim 3, wherein the generating an avatar corresponding to the target object from the plurality of nerve radiation field vectors comprises:

obtaining an explicit three-dimensional expression of the target object according to the plurality of nerve radiation field vectors; and

rendering the explicit three-dimensional representation to generate an avatar corresponding to the target object.

5. The method of any of claims 1 to 4, wherein the target object comprises a head.

6. An avatar generation apparatus comprising:

the first determining module is used for determining a style characteristic diagram of a target object in a target image;

the second determining module is used for determining a plurality of nerve radiation field vectors according to the style characteristic diagram; and

and the generating module is used for generating an avatar corresponding to the target object according to the plurality of nerve radiation field vectors.

7. The apparatus of claim 6, wherein the first determining means comprises:

a first determining unit for determining style encoding data of the target object; and

and the second determining unit is used for determining the style characteristic diagram according to the style encoding data.

8. The apparatus of claim 6, wherein the second determining means comprises:

the first generating unit is used for generating a style image according to the style characteristic diagram;

a third determining unit, configured to determine a reconstruction parameter according to the style image;

the fusion unit is used for fusing the style characteristic diagram with the reconstruction parameters to obtain a fusion characteristic diagram; and

and the fourth determining unit is used for determining a plurality of nerve radiation field vectors according to the fusion characteristic map.

9. The apparatus of claim 8, wherein the generating means comprises:

the obtaining unit is used for obtaining the explicit three-dimensional expression of the target object according to the plurality of nerve radiation field vectors; and

and the second generation unit is used for rendering the explicit three-dimensional expression and generating an avatar corresponding to the target object.

10. The apparatus of any of claims 6 to 9, wherein the target object comprises a head.

11. An electronic device, comprising:

at least one processor; and

a memory communicatively coupled to the at least one processor; wherein,

the memory stores instructions executable by the at least one processor to enable the at least one processor to perform the method of any one of claims 1 to 5.

12. A non-transitory computer readable storage medium having stored thereon computer instructions for causing the computer to perform the method of any one of claims 1 to 5.

13. A computer program product comprising a computer program which, when executed by a processor, implements the method of any one of claims 1 to 5.

Images