Data processing method and device

Abstract

A data processing method, relating to the field of artificial intelligence and comprising: acquiring a skeleton pose of a first character; on the basis of the skeleton pose and by means of a first encoder, obtaining a first feature representation, the first feature representation comprising a preset number of skeleton nodes and a feature representation corresponding to each skeleton node; and on the basis of the first feature representation and by means of a first decoder, obtaining a first reconstructed skeleton pose corresponding to the skeleton pose. According to the present application, feature representations of different skeleton structures are expressed in a same skeleton structure, that is, actions of different skeletons can be projected to a unified main skeleton, such that when encoders and decoders for different skeleton structures are trained, a model corresponding to the current skeleton structure is trained by means of data of other skeleton structures, thereby improving the action generation quality.

Description

A data processing method and device thereof

[0001] This application claims priority to the Chinese patent application filed with the State Intellectual Property Office on August 8, 2023, with application number 202310996509.5 and application name “A data processing method and device thereof”, the entire contents of which are incorporated by reference into this application. Technical Field

[0002] The present application relates to the field of artificial intelligence, and in particular to a data processing method and device thereof. Background Art

[0003] The rhythm of speech describes the speed at which the speaker speaks each syllable, which can be understood as the speaker's speaking habits (different speakers have different pause positions and pause durations when speaking). Tone refers to the melodic pattern of speech, which conveys different expressions (for example, surprise, anger, or happiness), and can also serve a grammatical function. At the sentence level, incorrect stress placement can change the meaning of the entire sentence. Gestures are directly and strongly related to speech. The content of speech describes the content information of the current speech and is also the speech text. Gestures are directly related to text. For example, the gestures made when saying "up", "big", "open", "applause", etc. should be very different from the gestures made when saying "down", "small", "close", "still", etc.

[0004] Body language plays an important role in communication. As shown in Figure 7, body movement generation is an important part of the general process framework of intelligent-driven virtual digital humans.

[0005] In the existing art, action generation is performed using rule-based action matching. The actions in the database are derived from the performances of motion-capture actors, and animators can communicate with the actors about details such as the speed and presentation of the actions. However, this requires the manual construction of a large number of speech-to-action mapping rules and an action library, which results in high computational overhead and memory usage. Different skeletons require different action libraries, and given the limited training samples, the model's action generation quality is poor.

[0006] Summary of the Invention

[0007] This application provides a data processing method that can improve the quality of action generation.

[0008] In a first aspect, the present application provides a data processing method, comprising: obtaining a skeleton posture of a first person; obtaining a first feature representation based on the skeleton posture through a first encoder, wherein the first feature representation includes a pre-set number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes; obtaining a first reconstructed skeleton posture corresponding to the skeleton posture through a first decoder based on the first feature representation; the first reconstructed skeleton posture is used to update the first encoder and the first decoder.

[0009] In an embodiment of the present application, the features of different skeletal structures can be expressed in the same skeletal structure, that is, the actions of different skeletons can be projected into a unified main skeleton, so that when training encoders and decoders of different skeletal structures (optionally, a neural network for aligning audio and posture features) the data of other skeletal structures can be used to train the model corresponding to the current skeletal structure, thereby improving the quality of action generation.

[0010] In a possible implementation, the method further includes: obtaining audio data of the first person, the skeleton posture being the skeleton posture of the first person when the audio data is collected; aligning the audio data and the skeleton posture through a neural network based on the audio data and the first feature representation to obtain an aligned first feature representation; obtaining a first reconstructed skeleton posture corresponding to the skeleton posture through a first decoder based on the first feature representation, including: obtaining a first reconstructed skeleton posture corresponding to the skeleton posture through a first decoder based on the aligned first feature representation.

[0011] Among them, the neural network can be a diffusion model.

[0012] In a possible implementation, the first feature representation includes a preset number of skeleton nodes and a feature representation in which the number of channels corresponding to each skeleton node is a preset number.

[0013] In a possible implementation, the feature representation corresponding to each of the skeleton nodes is related to the rotation feature of the joint.

[0014] In a possible implementation, the method further includes: obtaining a second feature representation through a second encoder based on the skeleton posture, wherein the second feature representation is related to the displacement feature of the joint; obtaining a first reconstructed skeleton posture corresponding to the skeleton posture through a first decoder based on the first feature representation, including: obtaining a first reconstructed skeleton posture corresponding to the skeleton posture through a first decoder based on the first feature representation and the second feature representation.

[0015] In a possible implementation, the method further includes: determining a first loss based on the first reconstructed skeleton pose and the skeleton pose, and updating the first encoder and the first decoder based on the first loss.

[0016] In one possible implementation, the method further includes: obtaining a skeleton posture of a second person; obtaining a third feature representation through a third encoder based on the skeleton posture of the second person, and the third feature representation is related to the displacement feature of the joint; obtaining a second reconstructed skeleton posture through a second decoder based on the first feature representation and the third feature representation; obtaining a fourth feature representation through a fourth encoder based on the second reconstructed skeleton posture; the fourth feature representation includes a pre-set number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes; the feature representation corresponding to each of the skeleton nodes in the fourth feature representation is related to the rotation feature of the joint; determining a second loss based on the first feature representation and the fourth feature representation, and updating the first encoder and the first decoder based on the second loss.

[0017] In one possible implementation, the method further includes: obtaining a fifth feature representation through a fourth encoder based on the skeletal posture of the second character, wherein the fifth feature representation is related to the rotation feature of the joint; obtaining a first discrimination result through a first discriminator based on the fourth feature representation and the third feature representation; obtaining a second discrimination result through a second discriminator based on the fifth feature representation and the third feature representation; determining a third loss based on the first discrimination result and the second discrimination result, and updating the first encoder and the first decoder based on the third loss.

[0018] In one possible implementation, the first reconstructed skeleton pose includes a first velocity of the end nodes of the human body, and the skeleton pose of the first character includes a second velocity of the end nodes of the human body; the method also includes: determining a fourth loss based on the normalized first velocity and the normalized second velocity, and updating the first encoder and the first decoder based on the fourth loss.

[0019] In a possible implementation, the method further includes: determining an alignment effect of the aligned first feature representation through a reward function based on the aligned first feature representation; and updating the neural network based on the alignment effect.

[0020] In a second aspect, the present application provides a data processing device, the device comprising:

[0021] An acquisition module, for acquiring a skeleton posture of a first character;

[0022] a processing module, configured to obtain a first feature representation based on the skeleton posture through a first encoder, wherein the first feature representation includes a preset number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes;

[0023] According to the first feature representation, a first reconstructed skeleton pose corresponding to the skeleton pose is obtained through a first decoder; the first reconstructed skeleton pose is used to update the first encoder and the first decoder.

[0024] In a possible implementation, the acquisition module is further configured to:

[0025] Acquire audio data of the first person, wherein the skeleton posture is the skeleton posture of the first person when the audio data is collected;

[0026] The processing module is further configured to align the audio data with the skeleton posture through a neural network based on the audio data and the first feature representation to obtain an aligned first feature representation;

[0027] The processing module is specifically used to:

[0028] According to the aligned first feature representation, a first reconstructed skeleton pose corresponding to the skeleton pose is obtained through a first decoder.

[0029] In a possible implementation, the first feature representation includes a preset number of skeleton nodes and a feature representation in which the number of channels corresponding to each skeleton node is a preset number.

[0030] In a possible implementation, the feature representation corresponding to each of the skeleton nodes is related to the rotation feature of the joint.

[0031] In a possible implementation, the processing module is further configured to:

[0032] Obtaining a second feature representation according to the skeleton posture through a second encoder, wherein the second feature representation is related to a displacement feature of a joint;

[0033] The processing module is specifically used to:

[0034] According to the first feature representation and the second feature representation, a first reconstructed skeleton pose corresponding to the skeleton pose is obtained through a first decoder.

[0035] In a possible implementation, the processing module is further configured to:

[0036] A first loss is determined based on the first reconstructed skeleton pose and the skeleton pose, and the first encoder and the first decoder are updated based on the first loss.

[0037] In a possible implementation, the processing module is further configured to:

[0038] Get the skeleton pose of the second character;

[0039] Obtaining, through a third encoder, a third feature representation based on the skeletal posture of the second person, wherein the third feature representation is related to a displacement feature of a joint;

[0040] Obtaining a second reconstructed skeleton pose through a second decoder according to the first feature representation and the third feature representation;

[0041] Obtaining a fourth feature representation based on the second reconstructed skeleton posture through a fourth encoder; the fourth feature representation includes a predetermined number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes; the feature representation corresponding to each of the skeleton nodes in the fourth feature representation is related to a rotation feature of a joint;

[0042] A second loss is determined based on the first feature representation and the fourth feature representation, and the first encoder and the first decoder are updated based on the second loss.

[0043] In a possible implementation, the processing module is further configured to:

[0044] Obtaining, through a fourth encoder, a fifth feature representation based on the skeletal posture of the second person, wherein the fifth feature representation is related to a rotation feature of a joint;

[0045] Obtaining a first discrimination result through a first discriminator according to the fourth feature representation and the third feature representation;

[0046] Obtaining a second discrimination result through a second discriminator according to the fifth feature representation and the third feature representation;

[0047] A third loss is determined according to the first discrimination result and the second discrimination result, and the first encoder and the first decoder are updated according to the third loss.

[0048] In a possible implementation, the first reconstructed skeleton pose includes a first velocity of a human end node, and the first character's skeleton pose includes a second velocity of a human end node; the processing module is further configured to:

[0049] A fourth loss is determined according to the normalized first speed and the normalized second speed, and the first encoder and the first decoder are updated according to the fourth loss.

[0050] In a possible implementation, the processing module is further configured to:

[0051] Determining an alignment effect of the aligned first feature representation using a reward function according to the aligned first feature representation;

[0052] The neural network is updated according to the alignment effect.

[0053] In a third aspect, an embodiment of the present application provides a data processing device, which may include a memory, a processor, and a bus system, wherein the memory is used to store programs, and the processor is used to execute the programs in the memory to perform the first aspect and any optional method thereof.

[0054] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored. When the computer-readable storage medium is run on a computer, the computer executes the above-mentioned first aspect and any optional method thereof.

[0055] In a fifth aspect, an embodiment of the present application provides a computer program, which, when executed on a computer, enables the computer to execute the above-mentioned first aspect and any optional method thereof.

[0056] In a sixth aspect, the present application provides a chip system comprising a processor for supporting an execution device or a training device in implementing the functions described in the aforementioned aspects, such as transmitting or processing data or information described in the aforementioned methods. In one possible design, the chip system further comprises a memory for storing program instructions and data necessary for the execution device or the training device. The chip system may consist of a single chip or may include a chip and other discrete components. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] FIG1A is a schematic diagram of a structure of an artificial intelligence main framework;

[0058] 1B and 2 are schematic diagrams of the application system framework of the present invention;

[0059] FIG3 is a schematic diagram of an optional hardware structure of a terminal;

[0060] FIG4 is a schematic diagram of the structure of a server;

[0061] FIG5 is a schematic diagram of a system architecture of the present application;

[0062] Figure 6 shows a cloud service process;

[0063] Figure 7 is a process for generating an action;

[0064] FIG8 is a flowchart of a data processing method provided in an embodiment of the present application;

[0065] FIG9 is a flowchart of a data processing method provided in an embodiment of the present application;

[0066] FIG10 is a flowchart of a data processing method provided in an embodiment of the present application;

[0067] FIG11 is a flowchart of a data processing method provided in an embodiment of the present application;

[0068] FIG12A is a flowchart of a data processing method provided in an embodiment of the present application;

[0069] FIG12B is a flowchart of a data processing method provided in an embodiment of the present application;

[0070] FIG13 is a schematic structural diagram of a data processing device provided in an embodiment of the present application;

[0071] FIG14 is a schematic diagram of a structure of an execution device provided in an embodiment of the present application;

[0072] FIG15 is a schematic diagram of a structure of a training device provided in an embodiment of the present application;

[0073] FIG16 is a schematic diagram of the structure of a chip provided in an embodiment of the present application. DETAILED DESCRIPTION

[0074] The following describes the embodiments of the present invention in conjunction with the accompanying drawings. The terms used in the embodiments of the present invention are only used to explain the specific embodiments of the present invention, and are not intended to limit the present invention.

[0075] The embodiments of the present application are described below in conjunction with the accompanying drawings. Those skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0076] The terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequential order. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances, and this is merely a way of distinguishing the objects of the same attributes when describing them in the embodiments of the present application. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, so that the process, method, system, product or equipment comprising a series of units need not be limited to those units, but may include other units that are not clearly listed or inherent to these processes, methods, products or equipment.

[0077] As used herein, the terms "substantially," "about," and similar terms are used as terms of approximation, not as terms of degree, and are intended to take into account the inherent variations in measurements or calculations that one of ordinary skill in the art would recognize. Furthermore, the use of "may" when describing embodiments of the present invention refers to "one or more possible embodiments." As used herein, the terms "use," "using," and "used" may be considered synonymous with the terms "utilize," "utilizing," and "utilized," respectively. Additionally, the term "exemplary" is intended to refer to an example or illustration.

[0078] First, let's describe the overall workflow of an AI system. See Figure 1A, which shows a schematic diagram of the main AI framework. This AI framework will be explained from two perspectives: the "intelligent information chain" (horizontal axis) and the "IT value chain" (vertical axis). The "intelligent information chain" reflects the entire process from data acquisition to processing. For example, it could be the general process of intelligent information perception, intelligent information representation and formation, intelligent reasoning, intelligent decision-making, and intelligent execution and output. Throughout this process, data undergoes a condensed journey from "data-information-knowledge-wisdom." The "IT value chain," spanning the underlying infrastructure of human intelligence, information (provided and processed by technology), and the system's industrial ecosystem, reflects the value that AI brings to the information technology industry.

[0079] (1) Infrastructure

[0080] Infrastructure provides computing power for AI systems, enabling communication with the outside world and supporting this through a foundational platform. External communication occurs through sensors; computing power is provided by intelligent chips (CPUs, NPUs, GPUs, ASICs, FPGAs, and other hardware accelerators). The foundational platform includes a distributed computing framework and network-related platform guarantees and support, including cloud storage and computing, and interconnected networks. For example, sensors communicate with the outside world to acquire data, which is then fed into the intelligent chips within the distributed computing system provided by the foundational platform for computation.

[0081] (2) Data

[0082] Data above the infrastructure layer represents data sources for AI. This data includes graphics, images, voice, and text, as well as IoT data from traditional devices. This includes business data from existing systems and sensor data such as force, displacement, liquid level, temperature, and humidity.

[0083] (3) Data processing

[0084] Data processing generally includes data training, machine learning, deep learning, search, reasoning, decision-making, etc.

[0085] Among them, machine learning and deep learning can symbolize and formalize data for intelligent information modeling, extraction, preprocessing, and training.

[0086] Reasoning refers to the process of simulating human intelligent reasoning in computers or intelligent systems, using formalized information to perform machine thinking and solve problems based on reasoning control strategies. Typical functions are search and matching.

[0087] Decision-making refers to the process of making decisions after intelligent information is reasoned, and usually provides functions such as classification, sorting, and prediction.

[0088] (4) General ability

[0089] After the data has undergone the data processing mentioned above, some general capabilities can be further formed based on the results of the data processing, such as algorithms or a general system, for example, translation, text analysis, computer vision processing, speech recognition, image recognition, etc.

[0090] (5) Smart products and industry applications

[0091] Smart products and industry applications refer to the products and applications of artificial intelligence systems in various fields. They are the encapsulation of the overall artificial intelligence solution, which productizes intelligent information decision-making and realizes practical application. Its application areas mainly include: smart terminals, smart transportation, smart medical care, autonomous driving, smart cities, etc.

[0092] First, we will introduce the application scenarios of this application. This application can be applied to, but is not limited to, applications that include human motion generation functions based on audio or text (hereinafter referred to as human motion generation applications) or cloud services provided by cloud-side servers. The following are introduced respectively:

[0093] 1. Human motion generation applications

[0094] The product form of the embodiment of the present application can be a human motion generation application. The human motion generation application can be run on a terminal device or a cloud-side server.

[0095] In one possible implementation, a human motion generation application can implement the task of generating human motion based on audio or text, wherein the human motion generation application can perform the task of generating human motion in response to the input audio or text, and obtain 3D posture information corresponding to the human motion. Based on the 3D posture information, a virtual character can be restored, and the posture of the character can be aligned with the audio in rhythm and rhythm.

[0096] In one possible implementation, the user can open a human motion generation application installed on a terminal device and input (which can be active input or passive collection, for example, collected through sensors on the terminal device). The human motion generation application can generate human motions for audio or text using the method provided in the embodiment of the present application, and present the 3D posture information or a virtual character restored based on the 3D posture information to the user (the presentation method can be but is not limited to display, saving, uploading to the cloud side, etc.).

[0097] In one possible implementation, a user can open a human motion generation application installed on a terminal device and input audio or text (which can be active input or passive collection, for example, collected through a camera on the terminal device). The human motion generation application can send the audio or text to a cloud-side server. The cloud-side server generates human motions for the audio or text using the method provided in an embodiment of the present application, and transmits the 3D posture information or the information of a virtual character restored based on the 3D posture information back to the terminal device. The terminal device can present the 3D posture information or the virtual character restored based on the 3D posture information to the user (the presentation method can be, but is not limited to, display, saving, uploading to the cloud side, etc.).

[0098] In one possible implementation, the human motion generation function implemented by the human motion generation application can be specifically used to enable the driving of virtual characters in application scenarios such as augmented reality (AR), virtual reality (VR), and mixed reality (MR) remote conferencing, sports health, and the metaverse.

[0099] Next, the human motion generation application in the embodiment of this application is introduced from the functional architecture and the product architecture that realizes the function.

[0100] Referring to FIG. 1B , FIG. 1B is a schematic diagram of the functional architecture of a human motion generation application in an embodiment of the present application:

[0101] In one possible implementation, as shown in FIG1B , a human motion generation application 102 may receive input parameters 101 (e.g., audio or text) and generate 3D pose information 103 (or information of a virtual character restored based on the 3D pose information). The human motion generation application 102 may be executed on, for example, at least one computer system and may include computer code that, when executed by one or more computers, causes the computers to execute the data processing method described herein.

[0102] Referring to FIG. 2 , FIG. 2 is a schematic diagram of the physical architecture of a human motion generation application program according to an embodiment of the present application:

[0103] Referring to Figure 2, a schematic diagram of a system architecture is shown. The system may include a terminal 100 and a server 200. The server 200 may include one or more servers (one server is used as an example in Figure 2), and the server 200 may provide an action generation service for one or more terminals.

[0104] Among them, the terminal 100 can be installed with a human motion generation application, or a web page related to the motion generation function can be opened. The above application and web page can provide an interface. The terminal 100 can receive the relevant parameters entered by the user on the motion generation function interface and send the above parameters to the server 200. The server 200 can obtain the processing results based on the received parameters and return the processing results to the terminal 100.

[0105] It should be understood that in some optional implementations, the terminal 100 can also complete the action of obtaining the data processing result based on the received parameters by itself without the cooperation of the server, and the embodiments of the present application are not limited to this.

[0106] Next, the product form of the terminal 100 in FIG2 is described;

[0107] The terminal 100 in the embodiment of the present application can be a mobile phone, a tablet computer, a wearable device, an in-vehicle device, an augmented reality (AR) / virtual reality (VR) device, a laptop computer, an ultra-mobile personal computer (UMPC), a netbook, a personal digital assistant (PDA), etc., and the embodiment of the present application does not impose any restrictions on this.

[0108] FIG3 shows a schematic diagram of an optional hardware structure of the terminal 100 .

[0109] 3 , the terminal 100 may include components such as a radio frequency unit 110, a memory 120, an input unit 130, a display unit 140, a camera 150 (optional), an audio circuit 160 (optional), a speaker 161 (optional), a microphone 162 (optional), a processor 170, an external interface 180, and a power supply 190. Those skilled in the art will appreciate that FIG3 is merely an example of a terminal or multi-function device and does not limit the terminal or multi-function device. The terminal or multi-function device may include more or fewer components than shown, or may combine certain components or have different components.

[0110] The input unit 130 can be used to receive input digital or character information and generate key signal input related to user settings and function control of the portable multifunction device. Specifically, the input unit 130 may include a touch screen 131 (optional) and / or other input devices 132. The touch screen 131 can detect user touch operations on or near it (for example, operations performed on or near the touch screen using a finger, joint, stylus, or any other suitable object) and drive corresponding connected devices according to pre-set programs. The touch screen can detect user touch actions on the touch screen, convert the touch actions into touch signals and transmit them to the processor 170. It can also receive and execute commands sent by the processor 170; the touch signals include at least touch point coordinate information. The touch screen 131 provides an input interface and an output interface between the terminal 100 and the user. Touch screens can be implemented using various types, including resistive, capacitive, infrared, and surface acoustic wave. In addition to the touch screen 131, the input unit 130 may also include other input devices. Specifically, the other input devices 132 may include, but are not limited to, one or more of a physical keyboard, function keys (such as a volume control button 132 , a switch button 133 , etc.), a trackball, a mouse, a joystick, and the like.

[0111] The input device 132 may receive input audio or text, etc.

[0112] The display unit 140 may be used to display information input by or provided to the user, various menus of the terminal 100, interactive interfaces, file display, and / or playback of any multimedia file. In an embodiment of the present application, the display unit 140 may be used to display the interface of a human motion generation application, a virtual human generated based on audio or text, and the like.

[0113] Memory 120 can be used to store instructions and data. It primarily includes an instruction storage area and a data storage area. The data storage area can store various data, such as multimedia files and text. The instruction storage area can store software units such as the operating system, applications, and instructions required for at least one function, or subsets or extensions thereof. It may also include non-volatile random access memory (RAM). It provides processor 170 with management functions for the hardware, software, and data resources within the computing and processing device, supporting control software and applications. It is also used to store multimedia files and running programs and applications.

[0114] The processor 170 is the control center of the terminal 100. It connects all components of the terminal 100 using various interfaces and circuits. By executing instructions stored in the memory 120 and accessing data stored therein, it executes various functions of the terminal 100 and processes data, thereby providing overall control of the terminal device. Optionally, the processor 170 may include one or more processing units. Preferably, the processor 170 may integrate an application processor and a modem processor, with the application processor primarily processing the operating system, user interface, and application programs, while the modem processor primarily handles wireless communications. It is understood that the modem processor may not be integrated into the processor 170. In some embodiments, the processor and memory may be implemented on a single chip; in other embodiments, they may be implemented on separate chips. The processor 170 may also generate corresponding operational control signals and send them to the corresponding components of the computing and processing device. It may also read and process data in the software, particularly the data and programs in the memory 120, to enable the various functional modules therein to perform their corresponding functions, thereby controlling the corresponding components to operate as instructed.

[0115] Among them, the memory 120 can be used to store software codes related to the data processing method, the processor 170 can execute the steps of the chip's data processing method, and can also schedule other units (such as the above-mentioned input unit 130 and display unit 140) to achieve corresponding functions.

[0116] The RF unit 110 (optional) can be used to send and receive information or receive and send signals during a call. For example, after receiving downlink information from the base station, it is passed to the processor 170 for processing; in addition, it sends the designed uplink data to the base station. Generally, the RF circuit includes but is not limited to an antenna, at least one amplifier, a transceiver, a coupler, a low noise amplifier (LNA), a duplexer, etc. In addition, the RF unit 110 can also communicate with network devices and other devices via wireless communication. This wireless communication can use any communication standard or protocol, including but not limited to Global System of Mobile communication (GSM), General Packet Radio Service (GPRS), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Long Term Evolution (LTE), email, Short Messaging Service (SMS), etc.

[0117] In this embodiment of the present application, the RF unit 110 can send audio or text to the server 200, and receive 3D posture information sent by the server 200 or information of a virtual character restored based on the 3D posture information.

[0118] It should be understood that the radio frequency unit 110 is optional and can be replaced by other communication interfaces, such as a network port.

[0119] The terminal 100 also includes a power supply 190 (such as a battery) for supplying power to various components. Preferably, the power supply can be logically connected to the processor 170 through a power management system, thereby managing functions such as charging, discharging, and power consumption through the power management system.

[0120] The terminal 100 further includes an external interface 180 , which may be a standard Micro USB interface or a multi-pin connector, and may be used to connect the terminal 100 to other devices for communication, or to connect a charger to charge the terminal 100 .

[0121] Although not shown, the terminal 100 may also include a flash, a wireless fidelity (WiFi) module, a Bluetooth module, sensors with different functions, etc., which are not described in detail here. Some or all of the methods described below may be applied to the terminal 100 shown in FIG3 .

[0122] Next, the product form of the server 200 in FIG2 is described;

[0123] FIG4 is a schematic diagram of the structure of a server 200. As shown in FIG4, the server 200 includes a bus 201, a processor 202, a communication interface 203, and a memory 204. The processor 202, the memory 204, and the communication interface 203 communicate with each other via the bus 201.

[0124] Bus 201 may be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus. Buses can be classified as address buses, data buses, control buses, etc. For ease of illustration, FIG4 shows only one thick line, but this does not mean that there is only one bus or only one type of bus.

[0125] The processor 202 may be any one or more of a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP), or a digital signal processor (DSP).

[0126] The memory 204 may include volatile memory, such as random access memory (RAM). The memory 204 may also include non-volatile memory, such as read-only memory (ROM), flash memory, hard drive (HDD), or solid state drive (SSD).

[0127] The memory 204 may be used to store software codes related to the data processing method, and the processor 202 may execute the steps of the data processing method of the chip, and may also schedule other units to implement corresponding functions.

[0128] It should be understood that the above-mentioned terminal 100 and server 200 can be centralized or distributed devices, and the processors in the above-mentioned terminal 100 and server 200 (such as processor 170 and processor 202) can be hardware circuits (such as application specific integrated circuit (ASIC), field-programmable gate array (FPGA), general-purpose processor, digital signal processor (DSP), microprocessor or microcontroller, etc.), or a combination of these hardware circuits. For example, the processor can be a hardware system with an instruction execution function, such as a CPU, DSP, etc., or a hardware system without an instruction execution function, such as an ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without an instruction execution function and hardware systems with an instruction execution function.

[0129] It should be understood that the steps related to the model reasoning process in the embodiments of this application involve AI-related operations. When performing AI operations, the instruction execution architecture of the terminal device and server is not limited to the processor-memory architecture described above. The system architecture provided in the embodiments of this application is described in detail below with reference to Figure 5.

[0130] FIG5 is a schematic diagram of the system architecture provided by an embodiment of the present application. As shown in FIG5 , the system architecture 500 includes an execution device 510 , a training device 520 , a database 530 , a client device 540 , a data storage system 550 , and a data acquisition system 560 .

[0131] The execution device 510 includes a calculation module 511, an I / O interface 512, a pre-processing module 513, and a post-processing module 514. The calculation module 511 may include the target model / rule 501, and the pre-processing module 513 and the post-processing module 514 are optional.

[0132] The execution device 510 may be a terminal device or a server that runs the above-mentioned human body motion generation application.

[0133] The data acquisition device 560 is used to collect training samples. The training samples can be audio or text, as well as annotations of characters in the audio or text (such as skeleton posture information). After collecting the training samples, the data acquisition device 560 stores them in the database 530.

[0134] The training device 520 can train the neural network to be trained (such as the encoder, decoder, neural network, etc. in the embodiment of the present application) based on the training samples maintained in the database 530 to obtain the target model / rule 501.

[0135] It should be noted that, in actual applications, the training samples maintained in the database 530 may not all be collected by the data acquisition device 560, but may also be received from other devices. It should also be noted that the training device 520 may not train the target model / rule 501 entirely based on the training samples maintained in the database 530, but may also obtain training samples from the cloud or other places for model training. The above description should not be used as a limitation on the embodiments of the present application.

[0136] The target model / rule 501 obtained through training with the training device 520 can be applied to different systems or devices, such as the execution device 510 shown in FIG5 . The execution device 510 can be a terminal, such as a mobile phone terminal, a tablet computer, a laptop computer, an augmented reality (AR) / virtual reality (VR) device, an in-vehicle terminal, etc., or a server, etc.

[0137] Specifically, the training device 520 may transfer the trained model to the execution device 510 .

[0138] In Figure 5, the execution device 510 is configured with an input / output (I / O) interface 512 for data interaction with external devices. The user can input data (such as audio or text in the embodiment of the present application) to the I / O interface 512 through the client device 540.

[0139] Preprocessing module 513 and preprocessing module 514 are used to preprocess the input data received by I / O interface 512. It should be understood that preprocessing module 513 and preprocessing module 514 may be absent or only one preprocessing module may be present. If preprocessing module 513 and preprocessing module 514 are absent, computing module 511 may be used directly to process the input data.

[0140] When the execution device 510 preprocesses the input data, or when the computing module 511 of the execution device 510 performs calculations and other related processing, the execution device 510 can call the data, code, etc. in the data storage system 550 for corresponding processing, and can also store the data, instructions, etc. obtained from the corresponding processing in the data storage system 550.

[0141] Finally, the I / O interface 512 provides the processing results (such as the skeleton posture of the character or the information of the virtual character restored based on the skeleton posture of the character, etc.) to the client device 540, and then provides it to the user.

[0142] In the scenario shown in FIG5 , the user can manually input data, and this "manual input data" can be operated through the interface provided by I / O interface 512. In another scenario, client device 540 can automatically send input data to I / O interface 512. If user authorization is required for client device 540 to automatically send input data, the user can set the corresponding permissions in client device 540. The user can view the results output by execution device 510 on client device 540, and the specific presentation form can be a display, sound, action, or other specific method. Client device 540 can also serve as a data acquisition terminal, collecting input data input into I / O interface 512 and output results from I / O interface 512 as new sample data and storing them in database 530. Of course, collection can also be performed without client device 540, and instead the I / O interface 512 directly stores the input data input into I / O interface 512 and output results from I / O interface 512 as new sample data in database 530.

[0143] It is worth noting that FIG5 is merely a schematic diagram of a system architecture provided by an embodiment of the present application, and the positional relationships between the devices, components, modules, etc. shown in the figure do not constitute any limitation. For example, in FIG5 , the data storage system 550 is an external memory relative to the execution device 510. In other cases, the data storage system 550 can also be placed in the execution device 510. It should be understood that the execution device 510 can be deployed in the client device 540.

[0144] From the inference side of the model:

[0145] In the embodiment of the present application, the computing module 511 of the above-mentioned execution device 520 can obtain the code stored in the data storage system 550 to implement the steps related to the model reasoning process in the embodiment of the present application.

[0146] In an embodiment of the present application, the computing module 511 of the execution device 520 may include a hardware circuit (such as an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a general-purpose processor, a digital signal processor (DSP), a microprocessor or a microcontroller, etc.), or a combination of these hardware circuits. For example, the training device 520 can be a hardware system with an instruction execution function, such as a CPU, DSP, etc., or a hardware system without an instruction execution function, such as an ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without an instruction execution function and hardware systems with an instruction execution function.

[0147] Specifically, the computing module 511 of the execution device 520 can be a hardware system with the function of executing instructions, and the steps related to the model reasoning process provided in the embodiment of the present application can be software codes stored in the memory. The computing module 511 of the execution device 520 can obtain the software code from the memory and execute the obtained software code to implement the steps related to the model reasoning process provided in the embodiment of the present application.

[0148] It should be understood that the computing module 511 of the execution device 520 can be a combination of a hardware system that does not have the function of executing instructions and a hardware system that has the function of executing instructions. Some of the steps related to the model reasoning process provided in the embodiment of the present application can also be implemented by the hardware system that does not have the function of executing instructions in the computing module 511 of the execution device 520, which is not limited here.

[0149] From the training side of the model:

[0150] In an embodiment of the present application, the above-mentioned training device 520 can obtain the code stored in the memory (not shown in Figure 5, which can be integrated into the training device 520 or deployed separately from the training device 520) to implement the steps related to model training in the embodiment of the present application.

[0151] In an embodiment of the present application, the training device 520 may include a hardware circuit (such as an application specific integrated circuit (ASIC), a field-programmable gate array (FPGA), a general-purpose processor, a digital signal processor (DSP), a microprocessor or a microcontroller, etc.), or a combination of these hardware circuits. For example, the training device 520 may be a hardware system with an instruction execution function, such as a CPU, DSP, etc., or a hardware system without an instruction execution function, such as an ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without an instruction execution function and hardware systems with an instruction execution function.

[0152] It should be understood that the training device 520 can be a combination of a hardware system that does not have the function of executing instructions and a hardware system that has the function of executing instructions. Some of the steps related to model training provided in the embodiments of the present application can also be implemented by the hardware system in the training device 520 that does not have the function of executing instructions, which is not limited here.

[0153] 2. Action generation cloud services provided by the server:

[0154] In a possible implementation, the server may provide the terminal with an action generation function service through an application programming interface (API).

[0155] Among them, the terminal device can send relevant parameters (such as audio or text containing the human body) to the server through the API provided by the cloud. The server can obtain processing results (such as the skeleton posture of the character or the information of the virtual character restored based on the skeleton posture of the character, etc.) based on the received parameters, and return the processing results to the terminal.

[0156] The description of the terminal and the server can be the same as that of the above embodiments, and will not be repeated here.

[0157] FIG6 shows a process of using an action generation function cloud service provided by a cloud platform.

[0158] 1. Activate and purchase content review services.

[0159] 2. Users can download the software development kit (SDK) corresponding to the content review service. Usually, the cloud platform provides multiple development versions of the SDK for users to choose according to the requirements of the development environment, such as JAVA version SDK, Python version SDK, PHP version SDK, Android version SDK, etc.

[0160] 3. After the user downloads the corresponding version of the SDK to the local computer as needed, import the SDK project into the local development environment, configure and debug it in the local development environment. The local development environment can also be used to develop other functions, forming an application that integrates action generation functional capabilities.

[0161] 4. When an action generation application is in use and needs to perform an action generation function, it can trigger an API call for the action generation function. When the application triggers the action generation function, it initiates an API request to the running instance of the action generation service in the cloud environment. The API request carries audio or text, and the running instance in the cloud environment processes the audio or text to obtain the processing results (such as the skeletal posture of a character or information about a virtual character restored based on the skeletal posture of the character).

[0162] 5. The cloud environment returns the processing results to the application, thereby completing an action generation function service call.

[0163] Since the embodiments of the present application involve the application of a large number of neural networks, in order to facilitate understanding, the relevant terms and related concepts such as neural networks involved in the embodiments of the present application are first introduced below.

[0164] (1) Neural Network

[0165] A neural network can be composed of neural units. A neural unit can refer to an operation unit with xs and intercept 1 as input. The output of the operation unit can be:

[0166] Where s = 1, 2, ... n, n is a natural number greater than 1, Ws is the weight of xs, and b is the bias of the neural unit. f is the activation function of the neural unit, which is used to introduce nonlinear characteristics into the neural network to convert the input signal in the neural unit into the output signal. The output signal of the activation function can be used as the input of the next convolutional layer. The activation function can be a sigmoid function. A neural network is a network formed by connecting many of the above-mentioned single neural units together, that is, the output of one neural unit can be the input of another neural unit. The input of each neural unit can be connected to the local receptive field of the previous layer to extract the features of the local receptive field. The local receptive field can be an area composed of several neural units.

[0167] (2) Deep Neural Networks

[0168] A deep neural network (DNN) can be understood as a neural network with many hidden layers. There is no special metric for "many" here. The multi-layer neural networks and deep neural networks we often talk about are essentially the same thing. According to the position of different layers of DNN, the neural network inside the DNN can be divided into three categories: input layer, hidden layer, and output layer. Generally speaking, the first layer is the input layer, the last layer is the output layer, and the layers in between are all hidden layers. The layers are fully connected, that is, any neuron in the i-th layer must be connected to any neuron in the i+1-th layer. Although DNN looks complicated, the work of each layer is actually not complicated. Simply put, it is the following linear relationship expression: in, is the input vector, is the output vector, is the offset vector, W is the weight matrix (also called coefficient), and α() is the activation function. Each layer is just an input vector After such a simple operation, the output vector Since there are many DNN layers, the coefficient W and the offset vector So how are the specific parameters defined in DNN? First, let's look at the definition of coefficient W. Take a three-layer DNN as an example, for example: the linear coefficient from the 4th neuron in the second layer to the 2nd neuron in the third layer is defined as The superscript 3 represents the layer number of the coefficient W, while the subscripts correspond to the output of the third layer index 2 and the input of the second layer index 4. In summary, the coefficients from the kth neuron in the L-1th layer to the jth neuron in the Lth layer are defined as Note that the input layer has no W parameter. In deep neural networks, more hidden layers allow the network to better capture complex real-world situations. Theoretically, a model with more parameters is more complex and has greater "capacity," meaning it can handle more complex learning tasks.

[0169] (3) Convolutional Neural Network (CNN) is a deep neural network with a convolutional structure. A convolutional neural network contains a feature extractor consisting of a convolution layer and a subsampling layer. The feature extractor can be regarded as a filter, and the convolution process can be regarded as using a trainable filter to convolve an input image or convolution feature plane (feature map). The convolution layer refers to the neuron layer in the convolutional neural network that performs convolution processing on the input signal. In the convolution layer of the convolutional neural network, a neuron can only be connected to some neurons in the adjacent layer. A convolution layer usually contains several feature planes, and each feature plane can be composed of some rectangularly arranged neural units. The neural units in the same feature plane share weights, and the shared weights here are the convolution kernels. Shared weights can be understood as the way to extract image information is independent of position. The implicit principle is that the statistical information of a part of the image is the same as that of other parts. This means that the image information learned in a part can also be used in another part. Therefore, for all positions on the image, we can use the same learned image information. In the same convolutional layer, multiple convolution kernels can be used to extract different image information. Generally speaking, the more convolution kernels there are, the richer the image information reflected by the convolution operation.

[0170] Convolution kernels can be initialized as matrices of random size, and during the training process of the convolutional neural network, the convolution kernels can be learned to obtain reasonable weights. In addition, the direct benefit of shared weights is that they reduce the number of connections between the layers of the convolutional neural network, while also reducing the risk of overfitting.

[0171] (4) Backpropagation algorithm

[0172] Convolutional neural networks can use the back propagation (BP) algorithm to correct the size of the parameters in the initial super-resolution model during training, reducing the reconstruction error loss of the super-resolution model. Specifically, the forward propagation of the input signal to the output generates an error loss. This error loss information is then backpropagated to update the parameters of the initial super-resolution model, thereby converging the error loss. The BP algorithm is a backward propagation movement dominated by the error loss, aiming to obtain the optimal super-resolution model parameters, such as the weight matrix.

[0173] (5) Loss function

[0174] During the training of a deep neural network, because we want the output of the deep neural network to be as close as possible to the desired predicted value, we can compare the current network's predicted value with the desired target value and then update the weight vector of each layer of the neural network based on the difference between the two. (Of course, there is usually an initialization process before the first update, which is to pre-configure the parameters for each layer in the deep neural network.) For example, if the network's predicted value is too high, the weight vector is adjusted to make it predict a lower value. This adjustment is continued until the deep neural network can predict the desired target value or a value very close to the desired target value. Therefore, it is necessary to predefine "how to compare the difference between the predicted value and the target value." This is the loss function (or objective function), which is an important equation used to measure the difference between the predicted value and the target value. For example, the loss function output value (loss) indicates a greater difference, so training a deep neural network becomes a process of minimizing this loss as much as possible.

[0175] (6) Backpropagation algorithm

[0176] Neural networks can use the back propagation (BP) algorithm to correct the size of the parameters in the initial neural network model during training, reducing the reconstruction error loss of the neural network model. Specifically, the forward propagation of the input signal to the output generates error loss. This error loss information is then backpropagated to update the parameters in the initial neural network model, thereby converging the error loss. The BP algorithm is a backward propagation movement dominated by error loss, aiming to obtain the optimal parameters of the neural network model, such as the weight matrix.

[0177] (7) Quantization: The process of constraining the input from a continuous or other large set of values ​​(such as real numbers) to a discrete set of values ​​(such as integers).

[0178] (8) Codebook: Discrete motion representations can be learned based on codebooks, and similar motion sequences can be clustered into a discrete representation. A codebook is similar to a table, a dictionary, or the principal component vector in principal component analysis. Existing research shows that codebooks help reduce motion freezing during motion generation and preserve motion details.

[0179] (9) Speech-driven gesture: Modeling the weak correlation between speech (audio and text) and gestures, thereby directly generating gesture sequences based on speech.

[0180] The rhythm of speech describes the speed at which the speaker speaks each syllable, which can be understood as the speaker's speaking habits (different speakers have different pause positions and pause durations when speaking). Tone refers to the melodic pattern of speech, which conveys different expressions (for example, surprise, anger, or happiness), and can also serve a grammatical function. At the sentence level, incorrect stress placement can change the meaning of the entire sentence. Gestures are directly and strongly related to speech. The content of speech describes the content information of the current speech and is also the speech text. Gestures are directly related to text. For example, the gestures made when saying "up", "big", "open", "applause", etc. should be very different from the gestures made when saying "down", "small", "close", "still", etc.

[0181] Body language plays an important role in communication. As shown in Figure 7, body movement generation is an important part of the general process framework of intelligent-driven virtual digital humans.

[0182] In the existing art, action generation is performed using rule-based action matching. The actions in the database are derived from the performances of motion-capture actors, and animators can communicate with the actors about details such as the speed and presentation of the actions. This requires manually constructing a large number of speech-to-action mapping rules and an action library, which results in high computational overhead and memory usage. Different skeletons require different action libraries, which is costly.

[0183] In order to solve the above problems, the present invention provides a data processing method. The data processing method of the present invention is described in detail below with reference to the accompanying drawings.

[0184] Referring to Figure 8, Figure 8 is a flow chart of a data processing method provided in an embodiment of the present application. As shown in Figure 8, a data processing method provided in an embodiment of the present application may include steps 801 to 803, and these steps are described in detail below.

[0185] 801. Obtain the skeleton posture of the first character;

[0186] In one possible implementation, when training a model for generating character movements, it is necessary to obtain training samples, which may include audio data (or text corresponding to the audio) and corresponding annotations, wherein the audio data may be the audio of the first character when speaking, and the annotations may include the skeleton posture of the character when speaking, and the skeleton posture may include posture information of multiple key nodes of the character (which may be simply referred to as nodes in this embodiment of the application), and the posture may include static and dynamic, the static may be the displacement information of the joint, and the dynamic may be the rotation information of the joint.

[0187] Exemplarily, the joints may include the head, left hand, right hand, left foot, right foot, etc.

[0188] 802. Obtain a first feature representation based on the skeleton posture through a first encoder, where the first feature representation includes a preset number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes.

[0189] In an embodiment of the present application, for different skeleton structures (for example, skeleton structures including different numbers of skeleton nodes, etc.), feature extraction can be performed through an encoder corresponding to the skeleton result, and the different skeleton structures can be unified into the same skeleton structure to obtain corresponding feature representations. For example, the encoder corresponding to the skeleton structure of the first character is the first encoder (it may also include a second encoder, and the first encoder and the second encoder are used to extract dynamic features and static features, respectively), and the encoder corresponding to the skeleton structure of the second character is the third encoder (it may also include a fourth encoder, and the third encoder and the fourth encoder are used to extract dynamic features and static features, respectively).

[0190] The so-called same skeleton structure may be understood as including the same number of nodes and the same connection relationship between the nodes, etc. Furthermore, the feature representation may include feature representations of a preset number of skeleton nodes.

[0191] Among them, the number of channels of the feature representation of each skeleton node obtained by the encoder can also be unified, that is, in one possible implementation, the first feature representation includes a preset number of skeleton nodes and a feature representation in which the number of channels corresponding to each skeleton node is a preset number.

[0192] The features fixed to the unified skeleton may be dynamic features of the skeleton nodes (such as the rotation features of the nodes described above), for example, the direction of rotation, the angular rate of rotation, and the like.

[0193] In one possible implementation, static features can also be extracted through another encoder (such as the second encoder in the embodiment of the present application). For example, a second feature representation can be obtained based on the skeleton posture through the second encoder, and the second feature representation is related to the displacement characteristics of the joint.

[0194] For example, an action of a character can be represented as a static component (e.g., joint displacements) and dynamic components (e.g., joint rotations), where T is the number of time frames and J is the number of joints. Given the motion of skeleton A, it can be expressed as and Static encoder (such as the first encoder in this application) and dynamic encoder (For example, the second encoder in this application) encodes the action into a latent variable:

[0195] in T′ = T / d, where d is the downsampling rate and H is the number of main skeleton joints. In one implementation, the number of main skeleton joints can be set to 7 (including 5 end nodes of the limbs and head, and 2 middle nodes of the body). C′ and C are the number of static and dynamic latent channels, respectively. Repeat represents splicing in the time dimension.

[0196] In an embodiment of the present application, the features of different skeletal structures can be expressed in the same skeletal structure, that is, the actions of different skeletons can be projected into a unified main skeleton, so that when training encoders and decoders of different skeletal structures (optionally, a neural network for aligning audio and posture features) the data of other skeletal structures can be used to train the model corresponding to the current skeletal structure, thereby improving the quality of action generation.

[0197] In a possible implementation, a neural network (such as a diffusion model) can be used to learn the mapping relationship between audio and action on the main skeleton, that is, to align the audio data and the skeleton posture.

[0198] In one possible implementation, audio data of the first person can be obtained, and the skeleton posture is the skeleton posture of the first person when the audio data is collected; based on the audio data and the first feature representation, the audio data and the skeleton posture are aligned through a neural network to obtain the aligned first feature representation.

[0199] For ease of description, the embodiment of the present application may also refer to the above-mentioned neural network for alignment as an action generation module. For example, as shown in FIG9 , the action generation module may include two submodules: a denoising module and a sampling module.

[0200] The input of the denoising module may include: 1) the number of denoising steps t; 2) the noisy action expression 3) Control condition c (including seed action d, style s and audio a). Based on these inputs, the denoised action expression can be estimated:

[0201] During training, conditional generation c1 = [d, s, a] and unconditional generation are combined The denoising result:

[0202] The denoising module is trained by optimizing the Huber loss.

[0203] The sampling module L at each time t t , predicted by the denoising process Then add noise to time t-1 to get L t-1 , repeat this step until t=0.

[0204] In this way, the diffusion model method can well learn the correlation between speech and action.

[0205] In order to further align the speech and action, a reward function for quantifying the alignment effect can be set, and the neural network for aligning speech and action can be updated based on the value of the reward function.

[0206] In a possible implementation, the alignment effect of the aligned first feature representation can be determined by a reward function based on the aligned first feature representation; and the neural network is updated based on the alignment effect.

[0207] Taking the action optimization module as an example, in order to further align speech and action, the present application uses the action optimization module to fine-tune the action generation model. Exemplarily, the action optimization module can include two submodules: 1) a quantization submodule; and 2) a reinforcement learning module.

[0208] As shown in Figure 10, since speech gestures are primarily associated with upper body movements, VQVAE quantization primarily targets the main skeletal motions of the upper body. As shown in Figure 11, a speech-gesture matching pair τ0 is sampled through VQVAE. The gesture encodings within this speech-gesture pair are then randomly replaced. This sampling method yields K candidate action sequences. It is assumed that the more replacements a gesture undergoes, the less it matches the original speech. The reward model is trained using these constructed gestures. After the reward model is trained, the action generation model is further optimized based on the reward model to generate actions that better match the speech.

[0209] Furthermore, to prevent the generated foot from drifting, physics guidance can be incorporated. For example, when detecting a sudden acceleration in the body's root node motion, it can be assumed that the foot is in contact with the ground. Based on this assumption and constraint, IK can be used to optimize the touchdown motion.

[0210] 803. Obtain a first reconstructed skeleton pose corresponding to the skeleton pose through a first decoder according to the first feature representation; the first reconstructed skeleton pose is used to update the first encoder and the first decoder.

[0211] In a possible implementation, a first reconstructed skeleton pose corresponding to the skeleton pose may be obtained through a first decoder according to the first feature representation.

[0212] In a possible implementation, a first reconstructed skeleton posture corresponding to the skeleton posture can be obtained through a first decoder based on the first feature representation and the second feature representation (that is, the feature representation containing dynamic information and static information of the action).

[0213] In a possible implementation, a first reconstructed skeleton pose corresponding to the skeleton pose may be obtained through a first decoder based on the aligned first feature representation.

[0214] Among them, the decoder D A Will (Second Characteristic Representation) and L A (Second feature representation) is decoded into corresponding static and dynamic components:

[0215] in represents the static component of reconstruction, Indicates L A By D A Decoded dynamic components.

[0216] In one possible implementation, the encoder and decoder (eg, the first encoder and first decoder, the second encoder, etc. in the embodiment of the present application) can be updated by constructing a loss.

[0217] In a possible implementation, a first loss may be determined based on the first reconstructed skeleton pose and the skeleton pose, and the first encoder and the first decoder may be updated based on the first loss.

[0218] During the training phase, D A Trying to reconstruct the action, the decoder can be trained by minimizing the following reconstruction loss:

[0219] Where FK is the forward kinematics feedback.

[0220] In one possible implementation, the skeletal posture of the second character can also be obtained; the skeletal structure corresponding to the second character can be different from the skeletal structure of the first character. When generating actions for the skeletal structure corresponding to the second character, a third encoder and a fourth encoder can be used, wherein the third encoder is used to extract static features, and the fourth encoder can be used to extract dynamic features.

[0221] In one possible implementation, based on the skeletal posture of the second person, a third feature representation is obtained through a third encoder, and the third feature representation is related to the displacement feature of the joint; based on the first feature representation and the third feature representation, a second reconstructed skeletal posture is obtained through a second decoder; that is, the dynamic features of the first person and the static features of the second person are decoded through the decoder corresponding to the skeletal structure of the second person, thereby obtaining a generated action. The generated action is then re-encoded through the encoder corresponding to the skeletal structure of the second person, and the result obtained can be used to determine the second loss together with the first feature representation. Specifically, based on the second reconstructed skeletal posture, a fourth feature representation can be obtained through a fourth encoder; the fourth feature representation includes a pre-set number of skeletal nodes and a feature representation corresponding to each of the skeletal nodes; the feature representation corresponding to each of the skeletal nodes in the fourth feature representation is related to the rotation feature of the joint; the second loss is determined based on the first feature representation and the fourth feature representation, and the first encoder and the first decoder are updated based on the second loss.

[0222] The second loss can also be called latent variable consistency loss, which is defined as follows:

[0223] in Indicates L A By D B Decoded dynamic components.

[0224] In one possible implementation, to measure the effectiveness of redirection, an adversarial loss can be defined.

[0225] Specifically, a fifth feature representation can be obtained through a fourth encoder based on the skeletal posture of the second character, and the fifth feature representation is related to the rotation feature of the joint; a first discrimination result can be obtained through a first discriminator based on the fourth feature representation and the third feature representation; a second discrimination result can be obtained through a second discriminator based on the fifth feature representation and the third feature representation; a third loss is determined based on the first discrimination result and the second discrimination result, and the first encoder and the first decoder are updated based on the third loss.

[0226] For example, to measure the effect of redirection, the adversarial loss can be defined as follows:

[0227] In one possible implementation, since different skeletons can have the same set of end nodes, typically the head, left hand, right hand, left foot, and right foot, the ends of the original skeleton and the retargeted skeleton should have the same normalized velocity to avoid problems caused by redirection, such as slipping feet. The first reconstructed skeleton pose may include a first velocity of the end nodes of the human body, and the skeleton pose of the first character includes a second velocity of the end nodes of the human body; a fourth loss may be determined based on the normalized first velocity and the normalized second velocity, and the first encoder and the first decoder may be updated based on the fourth loss.

[0228] For example, the formula for the fourth loss is:

[0229] in and are the velocities of the e-th end nodes of skeleton A and skeleton B, ε is the set of end nodes, h A and h B are the heights of skeleton A and skeleton B respectively.

[0230] For example, the overall loss can be:

[0231] Exemplarily, the overall process of the present invention can be shown in Figure 12A. First, the different skeletons are unified into the main skeleton through the skeleton unification module, and then the correspondence between the audio and the main skeleton movements is learned through the diffusion model. Finally, we optimize the generated results through an optimization module and decode the main skeleton movements to the target skeleton.

[0232] FIG12B shows the overall process of the embodiment 1. In the training phase, for skeletons A and B that are different in topology or length, the skeletons are firstly trained through the skeleton redirection network E.A and E B , encoding all actions to the same main skeleton. A diffusion model then learns the correspondence between the main skeleton's motion and the audio. Finally, the generated actions are used to train a reinforcement learning module, and the learned reinforcement learning module is used to fine-tune the previous action generation module. During the inference phase, the action generation module generates actions corresponding to the main skeleton directly based on the audio. The action decoder then decodes the generated actions to the target skeleton.

[0233] 13 , which is a schematic diagram of the structure of a data processing device provided in an embodiment of the present application. As shown in FIG13 , a data processing device 1300 provided in an embodiment of the present application includes:

[0234] An acquisition module 1301 is configured to acquire a skeleton posture of a first person;

[0235] The specific description of the acquisition module 1301 can refer to the introduction of step 801 in the above embodiment, which will not be repeated here.

[0236] A processing module 1302 is configured to obtain a first feature representation based on the skeleton posture through a first encoder, where the first feature representation includes a predetermined number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes;

[0237] According to the first feature representation, a first reconstructed skeleton pose corresponding to the skeleton pose is obtained through a first decoder; the first reconstructed skeleton pose is used to update the first encoder and the first decoder.

[0238] The specific description of the processing module 1302 can refer to the introduction of steps 802 and 803 in the above embodiment, which will not be repeated here.

[0239] In a possible implementation, the obtaining module 1301 is further configured to:

[0240] Acquire audio data of the first person, wherein the skeleton posture is the skeleton posture of the first person when the audio data is collected;

[0241] The processing module 1302 is further configured to align the audio data with the skeleton posture through a neural network based on the audio data and the first feature representation to obtain an aligned first feature representation;

[0242] The processing module 1302 is specifically configured to:

[0243] According to the aligned first feature representation, a first reconstructed skeleton pose corresponding to the skeleton pose is obtained through a first decoder.

[0244] In a possible implementation, the first feature representation includes a preset number of skeleton nodes and a feature representation in which the number of channels corresponding to each skeleton node is a preset number.

[0245] In a possible implementation, the feature representation corresponding to each of the skeleton nodes is related to the rotation feature of the joint.

[0246] In a possible implementation, the processing module 1302 is further configured to:

[0247] Obtaining a second feature representation according to the skeleton posture through a second encoder, wherein the second feature representation is related to a displacement feature of a joint;

[0248] The processing module 1302 is specifically configured to:

[0249] According to the first feature representation and the second feature representation, a first reconstructed skeleton pose corresponding to the skeleton pose is obtained through a first decoder.

[0250] In a possible implementation, the processing module 1302 is further configured to:

[0251] A first loss is determined based on the first reconstructed skeleton pose and the skeleton pose, and the first encoder and the first decoder are updated based on the first loss.

[0252] In a possible implementation, the processing module 1302 is further configured to:

[0253] Get the skeleton pose of the second character;

[0254] Obtaining, through a third encoder, a third feature representation based on the skeletal posture of the second person, wherein the third feature representation is related to a displacement feature of a joint;

[0255] Obtaining a second reconstructed skeleton pose through a second decoder according to the first feature representation and the third feature representation;

[0256] Obtaining a fourth feature representation based on the second reconstructed skeleton posture through a fourth encoder; the fourth feature representation includes a predetermined number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes; the feature representation corresponding to each of the skeleton nodes in the fourth feature representation is related to a rotation feature of a joint;

[0257] A second loss is determined based on the first feature representation and the fourth feature representation, and the first encoder and the first decoder are updated based on the second loss.

[0258] In a possible implementation, the processing module 1302 is further configured to:

[0259] Obtaining, through a fourth encoder, a fifth feature representation based on the skeletal posture of the second person, wherein the fifth feature representation is related to a rotation feature of a joint;

[0260] Obtaining a first discrimination result through a first discriminator according to the fourth feature representation and the third feature representation;

[0261] Obtaining a second discrimination result through a second discriminator according to the fifth feature representation and the third feature representation;

[0262] A third loss is determined according to the first discrimination result and the second discrimination result, and the first encoder and the first decoder are updated according to the third loss.

[0263] In a possible implementation, the first reconstructed skeleton pose includes a first velocity of a human end node, and the first character's skeleton pose includes a second velocity of a human end node; the processing module 1302 is further configured to:

[0264] A fourth loss is determined according to the normalized first speed and the normalized second speed, and the first encoder and the first decoder are updated according to the fourth loss.

[0265] In a possible implementation, the processing module 1302 is further configured to:

[0266] Determining an alignment effect of the aligned first feature representation using a reward function according to the aligned first feature representation;

[0267] The neural network is updated according to the alignment effect.

[0268] Next, an execution device provided in an embodiment of the present application is introduced. Please refer to Figure 14. Figure 14 is a structural schematic diagram of an execution device provided in an embodiment of the present application. The execution device 1400 can be specifically manifested as a virtual reality VR device, a mobile phone, a tablet, a laptop computer, a smart wearable device, a monitoring data processing device or a server, etc., which is not limited here. Specifically, the execution device 1400 includes: a receiver 1401, a transmitter 1402, a processor 1403 and a memory 1404 (wherein the number of processors 1403 in the execution device 1400 can be one or more, and Figure 14 takes one processor as an example), wherein the processor 1403 may include an application processor 14031 and a communication processor 14032. In some embodiments of the present application, the receiver 1401, the transmitter 1402, the processor 1403 and the memory 1404 may be connected via a bus or other means.

[0269] Memory 1404 may include read-only memory and random access memory, and provides instructions and data to processor 1403. A portion of memory 1404 may also include non-volatile random access memory (NVRAM). Memory 1404 stores processor and operation instructions, executable modules, or data structures, or subsets or extended sets thereof. The operation instructions may include various operation instructions for implementing various operations.

[0270] Processor 1403 controls the operation of the execution device. In specific applications, the various components of the execution device are coupled together via a bus system. In addition to a data bus, the bus system may also include a power bus, a control bus, and a status signal bus. However, for clarity, all bus systems are referred to as a bus system in the figure.

[0271] The methods disclosed in the above embodiments of the present application can be applied to or implemented by processor 1403. Processor 1403 can be an integrated circuit chip with signal processing capabilities. During implementation, each step of the above method can be completed by hardware integrated logic circuits or software instructions in processor 1403. The above processor 1403 can be a general-purpose processor, a digital signal processor (DSP), a microprocessor, or a microcontroller, and can further include an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, discrete gate or transistor logic devices, or discrete hardware components. The processor 1403 can implement or execute the various methods, steps, and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the methods disclosed in conjunction with the embodiments of the present application can be directly implemented as being executed by a hardware decoding processor, or can be executed by a combination of hardware and software modules in the decoding processor. The software module can be located in a storage medium known in the art, such as random access memory, flash memory, read-only memory, programmable read-only memory, electrically erasable programmable memory, or registers. The storage medium is located in memory 1404. Processor 1403 reads information from memory 1404 and, in conjunction with its hardware, completes the steps involved in the model inference process in the above method.

[0272] Receiver 1401 can be used to receive input digital or character information and generate signal input related to executing device-related settings and function control. Transmitter 1402 can be used to output digital or character information through the first interface. Transmitter 1402 can also be used to send instructions to the disk pack through the first interface to modify data in the disk pack. Transmitter 1402 can also include a display device such as a display screen.

[0273] The present application also provides a training device. Please refer to FIG. 15 , which is a schematic diagram of the structure of a training device provided by an embodiment of the present application. Specifically, the training device 1500 is implemented by one or more servers. The training device 1500 may vary significantly due to different configurations or performance. It may include one or more central processing units (CPUs) 1515 (e.g., one or more processors), a memory 1532, and one or more storage media 1530 (e.g., one or more mass storage devices) storing application programs 1542 or data 1544. The memory 1532 and storage medium 1530 may be either ephemeral or persistent storage. The program stored in the storage medium 1530 may include one or more modules (not shown), each module including a series of instruction operations on the training device. Furthermore, the CPU 1515 may be configured to communicate with the storage medium 1530 to execute the series of instruction operations in the storage medium 1530 on the training device 1500.

[0274] The training device 1500 may also include one or more power supplies 1526, one or more wired or wireless network interfaces 1550, one or more input and output interfaces 1558; or, one or more operating systems 1541, such as Windows Server™, Mac OS X™, Unix™, Linux™, FreeBSD™, etc.

[0275] In the embodiment of the present application, the central processing unit 1515 is used to execute actions related to model training in the above embodiment.

[0276] An embodiment of the present application also provides a computer program product, which, when running on a computer, enables the computer to execute the steps executed by the aforementioned execution device, or enables the computer to execute the steps executed by the aforementioned training device.

[0277] A computer-readable storage medium is also provided in an embodiment of the present application, which stores a program for signal processing. When the computer-readable storage medium is run on a computer, it enables the computer to execute the steps executed by the aforementioned execution device, or enables the computer to execute the steps executed by the aforementioned training device.

[0278] The execution device, training device or terminal device provided in the embodiments of the present application can specifically be a chip, and the chip includes: a processing unit and a communication unit, the processing unit can be, for example, a processor, and the communication unit can be, for example, an input / output interface, a pin or a circuit, etc. The processing unit can execute the computer execution instructions stored in the storage unit, so that the chip in the execution device executes the data processing method described in the above embodiment, or so that the chip in the training device executes the data processing method described in the above embodiment. Optionally, the storage unit is a storage unit in the chip, such as a register, a cache, etc. The storage unit can also be a storage unit located outside the chip in the wireless access device, such as a read-only memory (ROM) or other types of static storage devices that can store static information and instructions, a random access memory (RAM), etc.

[0279] Specifically, see Figure 16 , which is a schematic diagram of the structure of a chip provided in an embodiment of the present application. The chip can be represented as a neural network processor (NPU) 1600. NPU 1600 is mounted on a host CPU (host CPU) as a coprocessor, with tasks assigned by the host CPU. The core of the NPU is arithmetic circuit 1603, which is controlled by controller 1604 to extract matrix data from memory and perform multiplication operations.

[0280] In some implementations, arithmetic circuit 1603 includes multiple processing units (PEs). In some implementations, arithmetic circuit 1603 is a two-dimensional systolic array. Arithmetic circuit 1603 can also be a one-dimensional systolic array or other electronic circuitry capable of performing mathematical operations such as multiplication and addition. In some implementations, arithmetic circuit 1603 is a general-purpose matrix processor.

[0281] For example, assume there are input matrix A, weight matrix B, and output matrix C. The arithmetic circuit retrieves the corresponding data of matrix B from weight memory 1602 and caches it on each PE in the arithmetic circuit. The arithmetic circuit retrieves the data of matrix A from input memory 1601 and performs a matrix operation on matrix B. The partial or final matrix result is stored in accumulator 1608.

[0282] Unified memory 1606 is used to store input and output data. Weight data is directly transferred to weight memory 1602 through the Direct Memory Access Controller (DMAC) 1605. Input data is also transferred to unified memory 1606 through the DMAC.

[0283] BIU stands for Bus Interface Unit 1610 , which is used for interaction between the AXI bus, DMAC, and instruction fetch buffer (IFB) 1609 .

[0284] The bus interface unit 1610 (BIU) is used for the instruction fetch memory 1609 to obtain instructions from the external memory, and is also used for the storage unit access controller 1605 to obtain the original data of the input matrix A or the weight matrix B from the external memory.

[0285] DMAC is mainly used to move input data in the external memory DDR to the unified memory 1606 or move weight data to the weight memory 1602 or move input data to the input memory 1601.

[0286] The vector calculation unit 1607 includes multiple operation processing units. When necessary, it further processes the output of the operation circuit 1603, such as vector multiplication, vector addition, exponential operation, logarithmic operation, size comparison, etc. It is mainly used for non-convolutional / fully connected layer network calculations in neural networks, such as batch normalization, pixel-level summation, and upsampling of feature planes.

[0287] In some implementations, vector calculation unit 1607 can store the processed output vector to unified memory 1606. For example, vector calculation unit 1607 can apply a linear function or a nonlinear function to the output of operation circuit 1603, such as linear interpolation of the feature plane extracted by the convolution layer, or accumulate a vector of values ​​to generate an activation value. In some implementations, vector calculation unit 1607 generates a normalized value, a pixel-level summed value, or both. In some implementations, the processed output vector can be used as an activation input to operation circuit 1603, for example, for use in subsequent layers in a neural network.

[0288] An instruction fetch buffer 1609 connected to the controller 1604 is used to store instructions used by the controller 1604;

[0289] Unified memory 1606, input memory 1601, weight memory 1602, and instruction fetch memory 1609 are all on-chip memories. External memories are private to the NPU hardware architecture.

[0290] The processor mentioned in any of the above places can be a general-purpose central processing unit, a microprocessor, an ASIC, or one or more integrated circuits for controlling the execution of the above program.

[0291] It should also be noted that the device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separate, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed across multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present embodiment. In addition, in the drawings of the device embodiments provided in this application, the connection relationship between the modules indicates that there is a communication connection between them, which can be specifically implemented as one or more communication buses or signal lines.

[0292] Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software plus necessary general hardware, and of course can also be implemented by special hardware including application-specific integrated circuits, special CPUs, special memories, special components, etc. In general, all functions performed by computer programs can be easily implemented with corresponding hardware, and the specific hardware structures used to implement the same function can also be diverse, such as analog circuits, digital circuits or special circuits, etc. However, for the present application, software program implementation is a better implementation method in most cases. Based on this understanding, the technical solution of the present application is essentially or the part that contributes to the prior art can be embodied in the form of a software product, which is stored in a readable storage medium, such as a computer's floppy disk, USB flash drive, mobile hard disk, ROM, RAM, magnetic disk or optical disk, etc., and includes a number of instructions to enable a computer device (which can be a personal computer, training equipment, or network equipment, etc.) to execute the methods described in each embodiment of the present application.

[0293] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware, or any combination thereof. When implemented by software, all or part of the embodiments may be implemented in the form of a computer program product.

[0294] The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices. The computer instructions can be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions can be transmitted from a website, a computer, a training device or a data center by wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) mode to another website, a computer, a training device or a data center. The computer-readable storage medium can be any available medium that a computer can store or a data storage device such as a training device, a data center, etc. that includes one or more available media integrations. The available medium can be a magnetic medium, (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium (e.g., a solid-state drive (SSD)).

Claims

1. A data processing method, characterized in that: The method comprises: Get the skeleton pose of the first character; According to the skeleton posture, a first feature representation is obtained through a first encoder, wherein the first feature representation includes a preset number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes; According to the first feature representation, a first reconstructed skeleton posture corresponding to the skeleton posture is obtained through a first decoder; the first reconstructed skeleton posture is used to update the first encoder and the first decoder.

2. The method according to claim 1, characterized in that The method further comprises: Acquire audio data of the first person, wherein the skeleton posture is the skeleton posture of the first person when the audio data is collected; According to the audio data and the first feature representation, aligning the audio data and the skeleton posture through a neural network to obtain an aligned first feature representation; The step of obtaining, according to the first feature representation, a first reconstructed skeleton posture corresponding to the skeleton posture through a first decoder includes: According to the aligned first feature representation, a first reconstructed skeleton posture corresponding to the skeleton posture is obtained through a first decoder.

3. The method according to claim 1 or 2, characterized in that: The first feature representation includes a preset number of skeleton nodes and a feature representation in which the number of channels corresponding to each of the skeleton nodes is a preset number.

4. The method according to any one of claims 1 to 3, characterized in that: The feature representation corresponding to each of the skeleton nodes is related to the rotation feature of the joint.

5. The method according to claim 4, characterized in that The method further comprises: According to the skeleton posture, a second feature representation is obtained through a second encoder, wherein the second feature representation is related to the displacement feature of the joint; The step of obtaining, according to the first feature representation, a first reconstructed skeleton posture corresponding to the skeleton posture through a first decoder includes: According to the first feature representation and the second feature representation, a first reconstructed skeleton posture corresponding to the skeleton posture is obtained through a first decoder.

6. The method according to any one of claims 1 to 5, characterized in that: The method further comprises: A first loss is determined according to the first reconstructed skeleton pose and the skeleton pose, and the first encoder and the first decoder are updated according to the first loss.

7. The method according to any one of claims 1 to 6, characterized in that: The method further comprises: Get the skeleton pose of the second character; According to the skeleton posture of the second person, a third feature representation is obtained through a third encoder, wherein the third feature representation is related to the displacement feature of the joint; Obtaining a second reconstructed skeleton pose through a second decoder according to the first feature representation and the third feature representation; According to the second reconstructed skeleton posture, a fourth feature representation is obtained through a fourth encoder; the fourth feature representation includes a preset number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes; the feature representation corresponding to each of the skeleton nodes in the fourth feature representation is related to the rotation feature of the joint; A second loss is determined according to the first feature representation and the fourth feature representation, and the first encoder and the first decoder are updated according to the second loss.

8. The method according to claim 7, characterized in that The method further comprises: According to the skeleton posture of the second person, a fifth feature representation is obtained through a fourth encoder, wherein the fifth feature representation is related to the joint Rotational features are related; According to the fourth feature representation and the third feature representation, a first discrimination result is obtained by a first discriminator; According to the fifth feature representation and the third feature representation, a second discrimination result is obtained by a second discriminator; A third loss is determined according to the first discrimination result and the second discrimination result, and the first encoder and the first decoder are updated according to the third loss.

9. The method according to any one of claims 1 to 8, characterized in that: The first reconstructed skeleton posture includes a first velocity of a human end node, and the first character's skeleton posture includes a second velocity of a human end node; the method further includes: A fourth loss is determined according to the normalized first speed and the normalized second speed, and the first encoder and the first decoder are updated according to the fourth loss.

10. The method according to any one of claims 2 to 9, characterized in that: The method further comprises: According to the aligned first feature representation, determining the alignment effect of the aligned first feature representation through a reward function; The neural network is updated according to the alignment effect.

11. A data processing device, characterized in that: The device comprises: An acquisition module, used to acquire the skeleton posture of the first person; A processing module, configured to obtain a first feature representation through a first encoder according to the skeleton posture, wherein the first feature representation includes a preset number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes; According to the first feature representation, a first reconstructed skeleton posture corresponding to the skeleton posture is obtained through a first decoder; the first reconstructed skeleton posture is used to update the first encoder and the first decoder.

12. The device according to claim 11, characterized in that The acquisition module is further used for: Acquire audio data of the first person, wherein the skeleton posture is the skeleton posture of the first person when the audio data is collected; The processing module is further used to align the audio data and the skeleton posture through a neural network according to the audio data and the first feature representation to obtain an aligned first feature representation; The processing module is specifically used for: According to the aligned first feature representation, a first reconstructed skeleton posture corresponding to the skeleton posture is obtained through a first decoder.

13. The device according to claim 11 or 12, characterized in that The first feature representation includes a preset number of skeleton nodes and a feature representation in which the number of channels corresponding to each of the skeleton nodes is a preset number.

14. The device according to any one of claims 11 to 13, characterized in that: The feature representation corresponding to each of the skeleton nodes is related to the rotation feature of the joint.

15. The device according to claim 14, characterized in that The processing module is further used for: According to the skeleton posture, a second feature representation is obtained through a second encoder, wherein the second feature representation is related to the displacement feature of the joint; The processing module is specifically used for: According to the first feature representation and the second feature representation, a first reconstructed skeleton posture corresponding to the skeleton posture is obtained through a first decoder.

16. The device according to any one of claims 11 to 15, characterized in that The processing module is further used for: A first loss is determined according to the first reconstructed skeleton pose and the skeleton pose, and the first encoder and the first decoder are updated according to the first loss.

17. The device according to any one of claims 11 to 16, characterized in that The processing module is further used for: Get the skeleton pose of the second character; According to the skeleton posture of the second person, a third feature representation is obtained through a third encoder, wherein the third feature representation is related to the displacement feature of the joint; Obtaining a second reconstructed skeleton pose through a second decoder according to the first feature representation and the third feature representation; According to the second reconstructed skeleton posture, a fourth feature representation is obtained through a fourth encoder; the fourth feature representation includes a preset number of skeleton nodes and a feature representation corresponding to each of the skeleton nodes; the feature representation corresponding to each of the skeleton nodes in the fourth feature representation is related to the rotation feature of the joint; A second loss is determined according to the first feature representation and the fourth feature representation, and the first encoder and the first decoder are updated according to the second loss.

18. The device according to claim 17, characterized in that The processing module is further used for: According to the skeleton posture of the second person, a fifth feature representation is obtained through a fourth encoder, wherein the fifth feature representation is related to the rotation feature of the joint; According to the fourth feature representation and the third feature representation, a first discrimination result is obtained by a first discriminator; According to the fifth feature representation and the third feature representation, a second discrimination result is obtained by a second discriminator; A third loss is determined according to the first discrimination result and the second discrimination result, and the first encoder and the first decoder are updated according to the third loss.

19. The device according to any one of claims 11 to 18, characterized in that The first reconstructed skeleton posture includes a first velocity of a human end node, and the first character's skeleton posture includes a second velocity of a human end node; the processing module is further used for: A fourth loss is determined according to the normalized first speed and the normalized second speed, and the first encoder and the first decoder are updated according to the fourth loss.

20. The device according to any one of claims 12 to 19, characterized in that The processing module is also used for: According to the aligned first feature representation, determining the alignment effect of the aligned first feature representation through a reward function; The neural network is updated according to the alignment effect.

21. A computer storage medium, characterized in that The computer storage medium stores one or more instructions which, when executed by one or more computers, cause the one or more computers to perform the operations of the method of any one of claims 1 to 10.

22. A computer program product, characterized in that The method comprises computer-readable instructions, and when the computer-readable instructions are executed on a computer device, the computer device is caused to execute the method according to any one of claims 1 to 10.

23. A system comprising at least one processor and at least one memory; the processor and the memory are connected via a communication bus and communicate with each other; The at least one memory is used to store code; The at least one processor is configured to execute the code to perform the method according to any one of claims 1 to 10.

Images