WO2024046473A1 - 一种数据处理方法及其装置 - Google Patents

一种数据处理方法及其装置 Download PDF

Info

Publication number
WO2024046473A1
WO2024046473A1 PCT/CN2023/116552 CN2023116552W WO2024046473A1 WO 2024046473 A1 WO2024046473 A1 WO 2024046473A1 CN 2023116552 W CN2023116552 W CN 2023116552W WO 2024046473 A1 WO2024046473 A1 WO 2024046473A1
Authority
WO
WIPO (PCT)
Prior art keywords
data
action
network
network layers
feature representation
Prior art date
Application number
PCT/CN2023/116552
Other languages
English (en)
French (fr)
Inventor
敖腾隆
刘利斌
娄宇珂
陈宝权
张镇嵩
许松岑
吴小飞
Original Assignee
华为技术有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 华为技术有限公司 filed Critical 华为技术有限公司
Publication of WO2024046473A1 publication Critical patent/WO2024046473A1/zh

Links

Classifications

    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/04—Segmentation; Word boundary detection
    • G—PHYSICS
    • G10—MUSICAL INSTRUMENTS; ACOUSTICS
    • G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
    • G10L15/00—Speech recognition
    • G10L15/08—Speech classification or search
    • G10L15/16—Speech classification or search using artificial neural networks
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/233—Processing of audio elementary streams
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
    • H04N21/23—Processing of content or additional data; Elementary server operations; Server middleware
    • H04N21/234—Processing of video elementary streams, e.g. splicing of video streams or manipulating encoded video stream scene graphs
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/439—Processing of audio elementary streams
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
    • H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
    • H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
    • H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs

Definitions

  • This application relates to the field of artificial intelligence, and in particular, to a data processing method and its device.
  • body language can increase the rhythm of the speech, making it more vivid and persuasive.
  • Body movements can express intentions and convey emotions more accurately, complementing the information conveyed by voice; secondly, body movements can help users focus more on the content communicated with digital people; and body movements can enhance the persuasiveness and credibility of digital people. and realism; body movements can reflect the speaker's intentions and personality. Lack of body language or rigid body movements in communication will lead to the uncanny valley effect.
  • This application provides a data processing method that can make the generated actions have an accurate sense of rhythm, thereby making the actions more suitable for the actions of characters and objects when they actually make speech.
  • this application provides a data processing method, which method includes: acquiring voice data; determining multiple segmentation point positions from the voice data according to the audio characteristics of the speech data, and the segmentation point positions are Corresponding to the predicted rhythm point of the body movement made by the character object when emitting the voice data; according to the voice data and the information indicating the positions of the plurality of segmentation points, a feature representation is obtained through a feature extraction network; according to the Feature representation, through the action generation network, generates action data.
  • the action data can be the joint point rotation angle of the body action or the 3D point cloud information of the action, which is not limited here.
  • the action data of the character object that emits the voice data can be predicted based on the voice data.
  • the rhythm point can be from stillness to movement or from movement to movement when making the body movements. The critical point of stillness. If the rhythm information can be explicitly recognized before feature extraction and used as the input of the model, the subsequently generated actions can have an accurate sense of rhythm, thus making the actions more consistent with the actions of the characters when they actually make speech.
  • the audio characteristics are related to volume and/or pitch.
  • the plurality of dividing point positions include adjacent first positions and second positions, and the time interval between the second position and the first position is within a preset range; the third A position and the second position are determined in one of the following ways: the first position and the second position are onset points in the voice data; the first position is an onset point in the voice data. The second position is not the onset point in the voice data, and the second position is the time point in the voice data when the volume is greater than the threshold.
  • action rhythm points usually appear on the onset of speech (audio features related to volume), but not all onsets are rhythm points of user actions.
  • the distance between adjacent rhythm points of speech There is a range of spacing that is related to the speaker's style. Assuming that the range is [t1, t2], usually the time interval smaller than t1 is due to noise, filler words or unsmooth pronunciation. Started. Intervals beyond t2 usually correspond to pauses or periods of quiet silence. Based on these observations, an episode whose time interval between adjacent episodes is within a preset range can be determined as a rhythm point, and then the rhythm point of the body movement made by the user when emitting the voice data can be identified, and based on the rhythm Click to segment the voice data.
  • a heuristic strategy can be used to identify speech rhythms. Segment the speech, and those less than t1 will not be recognized as rhythm points, while pseudo-rhythm points can be inserted for those exceeding t2. Pseudo-rhythm points can be inserted in the following way: the time from the previous rhythm point exceeds t1, and the volume of the voice exceeds a certain threshold I a . This threshold can be set to the average volume of environmental noise so that adjacent segmentation points The time interval of the positions is within the preset range. If all the volumes in the interval are less than I a , insert the minimum pseudo-rhythm points evenly so that the time interval between adjacent dividing point positions is within the preset range.
  • the voice data and information indicating the locations of the multiple dividing points include: dividing the voice data into multiple voice segments according to the multiple dividing point locations; or, The voice data and characters indicating the positions of the plurality of dividing points.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the method further includes: according to the multiple network layers The feature representation output by one or more network layers close to the output layer in the layer determines the action category of the body movement; the generating action data through the action generation network according to the feature representation includes: according to the plurality of Some or all of the feature representations output by the network layer and the action categories are passed through the action generation network to generate action data.
  • the type of action in addition to inputting the feature representation obtained through the feature extraction network into the action generation network, the type of action can also be identified through a neural network based on the feature representation (such as high-level features in the feature representation). , and input action types and feature representations into the action generation network.
  • the action category itself carries semantics
  • the features input to the action generation network and the subsequently generated body movements have semantic relevance.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the method further includes: according to the multiple network layers The feature representation output by one or more network layers far away from the output layer in the layer is passed through the encoder to obtain the style encoding of the body action; the action data is generated through the action generation network according to the feature representation, including: according to Part or all of the feature representations output by the plurality of network layers and the style encoding are used to generate action data through an action generation network.
  • the distribution of the style codes is uniform distribution.
  • this application provides a data processing method, which method includes:
  • the voice data and first action data of the character object where the first action data is the real action data of the body movements made by the character object when emitting the voice data;
  • a feature representation is obtained through a feature extraction network
  • second action data is generated through an action generation network; the difference between the first action data and the second action data is used to update the feature extraction network and the action generation network.
  • the feature extraction network includes multiple network layers connected in series
  • the feature representation includes feature representations output by the multiple network layers
  • the network is generated through actions according to the feature representation.
  • generate the second action data including:
  • the body is obtained through the encoder.
  • the difference between the first style encoding and the updated style encoding is used to update the encoder.
  • the audio characteristics are related to volume and/or pitch.
  • the plurality of dividing point positions include adjacent first positions and second positions, and the time interval between the second position and the first position is within a preset range; the third A position and the second position are determined in one of the following ways:
  • the first position and the second position are the onset points in the speech data
  • the first position is an onset point in the voice data
  • the second position is not an onset point in the voice data
  • the second position is a time point in the voice data when the volume is greater than a threshold.
  • the voice data and information indicating the locations of the multiple dividing points include:
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the method further includes:
  • Generating second action data through an action generation network based on the feature representation includes:
  • Second action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the action category.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the method further includes:
  • the style encoding of the body movement is obtained through the encoder
  • Generating second action data through an action generation network based on the feature representation includes:
  • Second action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the style encoding.
  • this application provides a data processing device, which includes:
  • Acquisition module used to obtain voice data
  • a processing module configured to determine multiple segmentation point positions from the speech data according to the audio characteristics of the speech data, the segmentation point positions corresponding to predictions of body movements made by the character object when emitting the speech data. rhythm point;
  • a feature representation is obtained through a feature extraction network
  • action data is generated through an action generation network.
  • the audio characteristics are related to volume and/or pitch.
  • the plurality of dividing point positions include adjacent first positions and second positions, and the time interval between the second position and the first position is within a preset range; the third A position and the second position are determined in one of the following ways:
  • the first position and the second position are the onset points in the speech data
  • the first position is an onset point in the voice data
  • the second position is not an onset point in the voice data
  • the The second position is a time point in the voice data when the volume is greater than the threshold.
  • the voice data and information indicating the locations of the multiple dividing points include:
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is also used to:
  • the processing module is specifically used for:
  • Action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the action category.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is also used to:
  • the style encoding of the body movement is obtained through the encoder
  • the processing module is specifically used for:
  • Action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the style encoding.
  • the distribution of the style codes is uniform distribution.
  • this application provides a data processing device, which includes:
  • An acquisition module configured to acquire the voice data and first action data of the character object, where the first action data is the real action data of the body movements made by the character object when the voice data is emitted;
  • a processing module configured to determine multiple segmentation point positions from the speech data according to the audio characteristics of the speech data, the segmentation point positions corresponding to predictions of body movements made by the character object when emitting the speech data. rhythm point;
  • a feature representation is obtained through a feature extraction network
  • second action data is generated through an action generation network; the difference between the first action data and the second action data is used to update the feature extraction network and the action generation network.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is specifically used to:
  • the first style encoding of the body movement is obtained through the encoder
  • the difference between the first style encoding and the updated style encoding is used to update the encoder.
  • the audio characteristics are related to volume and/or pitch.
  • the plurality of dividing point positions include adjacent first positions and second positions, and the time interval between the second position and the first position is within a preset range; the third A position and the second position are determined in one of the following ways:
  • the first position and the second position are the onset points in the speech data
  • the first position is an onset point in the voice data
  • the second position is not an onset point in the voice data
  • the second position is a time point in the voice data when the volume is greater than a threshold.
  • the voice data and information indicating the locations of the multiple dividing points include:
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is also used to:
  • the processing module is specifically used for:
  • Second action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the action category.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is also used to:
  • the style encoding of the body movement is obtained through the encoder
  • the processing module is specifically used for:
  • Second action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the style encoding.
  • embodiments of the present application provide 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 as described above. and any optional method thereof, or the above second aspect and any optional method thereof.
  • embodiments of the present application provide a computer-readable storage medium that stores a computer program that, when run on a computer, causes the computer to execute the first aspect and any of the above.
  • embodiments of the present application provide a computer program that, when run on a computer, causes the computer to execute the above-mentioned first aspect and any optional method thereof, or the above-mentioned second aspect and any optional method thereof. method of selection.
  • this application provides a chip system, which includes a processor for supporting an execution device or a training device to implement the functions involved in the above aspects, for example, sending or processing data involved in the above methods; Or, information.
  • the chip system further includes a memory, and the memory is used to store necessary program instructions and data for executing the device or training the device.
  • the chip system may be composed of chips, or may include chips and other discrete devices.
  • Figure 1A is a schematic structural diagram of the main framework of artificial intelligence
  • FIG. 1B and Figure 2 are schematic diagrams of the application system framework of the present invention.
  • Figure 3 is a schematic diagram of an optional hardware structure of the terminal
  • Figure 4 is a schematic structural diagram of a server
  • Figure 5 is a schematic diagram of a system architecture of this application.
  • Figure 6 shows the process of a cloud service
  • Figure 7 shows the process of a cloud service
  • FIG. 8 is a flowchart of a data processing method provided by an embodiment of the present application.
  • FIG. 9 is a flowchart of the data processing method provided by the embodiment of the present application.
  • FIG. 10 is a flowchart of the data processing method provided by the embodiment of the present application.
  • FIG 11A is a flowchart of the data processing method provided by the embodiment of the present application.
  • FIG. 11B is a flowchart of the data processing method provided by the embodiment of the present application.
  • Figure 12A is a flowchart of the data processing method provided by the embodiment of the present application.
  • Figure 12B is a flowchart of the data processing method provided by the embodiment of the present application.
  • Figure 13 is a schematic structural diagram of a data processing device provided by an embodiment of the present application.
  • Figure 14 is a schematic structural diagram of a data processing device provided by an embodiment of the present application.
  • Figure 15 is a schematic structural diagram of an execution device provided by an embodiment of the present application.
  • Figure 16 is a schematic structural diagram of the training equipment provided by the embodiment of the present application.
  • Figure 17 is a schematic structural diagram of a chip provided by an embodiment of the present application.
  • 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 measurements or values that would be known to one of ordinary skill in the art. The inherent bias in calculated values.
  • the use of “may” when describing embodiments of the present invention refers to “one or more possible embodiments.”
  • the terms “use”, “using”, and “used” may be deemed to be the same as the terms “utilize”, “utilizing”, and “utilize”, respectively. Synonymous with “utilized”.
  • the term “exemplary” is intended to refer to an example or illustration.
  • Figure 1A shows a structural schematic diagram of the artificial intelligence main framework.
  • the following is from the “intelligent information chain” (horizontal axis) and “IT value chain” ( The above artificial intelligence theme framework is elaborated on the two dimensions of vertical axis).
  • the "intelligent information chain” reflects a series of processes from data acquisition to processing. For example, it can be the general process of intelligent information perception, intelligent information representation and formation, intelligent reasoning, intelligent decision-making, intelligent execution and output. In this process, the data has gone through the condensation process of "data-information-knowledge-wisdom".
  • the "IT value chain” reflects the value that artificial intelligence brings to the information technology industry, from the underlying infrastructure of human intelligence and information (providing and processing technology implementation) to the systematic industrial ecological process.
  • Infrastructure provides computing power support for artificial intelligence systems, enables communication with the external world, and supports it through basic platforms.
  • computing power is provided by smart chips (hardware acceleration chips such as CPU, NPU, GPU, ASIC, FPGA, etc.);
  • the basic platform includes distributed computing framework and network and other related platform guarantees and support, which can include cloud storage and Computing, interconnection networks, etc.
  • sensors communicate with the outside world to obtain data, which are provided to smart chips in the distributed computing system provided by the basic platform for calculation.
  • Data from the upper layer of the infrastructure is used to represent data sources in the field of artificial intelligence.
  • the data involves graphics, images, voice, and text, as well as IoT data of traditional devices, including business data of existing systems and sensory data such as force, displacement, liquid level, temperature, and humidity.
  • Data processing usually includes data training, machine learning, deep learning, search, reasoning, decision-making and other methods.
  • machine learning and deep learning can perform symbolic and formal intelligent information modeling, extraction, preprocessing, training, etc. on data.
  • Reasoning refers to the process of simulating human intelligent reasoning in computers or intelligent systems, using formal information to perform machine thinking and problem solving based on reasoning control strategies. Typical functions are search and matching.
  • Decision-making refers to the process of decision-making after intelligent information is reasoned, and usually provides functions such as classification, sorting, and prediction.
  • some general capabilities can be formed based on the results of further data processing, such as algorithms or a general system, such as translation, text analysis, computer vision processing, speech recognition, and image processing. identification, etc.
  • Intelligent products and industry applications refer to the products and applications of artificial intelligence systems in various fields. They are the encapsulation of overall artificial intelligence solutions, productizing intelligent information decision-making and realizing practical applications. Its application fields mainly include: intelligent terminals, intelligent transportation, Smart healthcare, autonomous driving, smart cities, etc.
  • This application can be, but is not limited to, applied to applications that have the function of generating videos containing virtual people (or other objects other than virtual people) based on text or speech (hereinafter may be referred to as virtual people for short).
  • Generating applications) or cloud services provided by cloud-side servers, etc. will be introduced separately below:
  • the product form of the embodiment of the present application may be a virtual human generation application program.
  • Virtual human generation applications can run on terminal devices or cloud-side servers.
  • a virtual human generation application can realize the task of generating a video containing a virtual human (or other objects other than the virtual human) based on text or speech, where, in the scenario of intelligent response, the text Or the voice can be a request, inquiry or other voice or text input by the user.
  • the video of the virtual person generated by the application can include the response voice to the voice or text input by the user and the virtual voice when the corresponding response voice is issued.
  • This virtual person will make human-like movements, that is, it can make body movements that match humans when they respond to speech.
  • body movements may be gestures.
  • the text or voice can be the voice or text input by the user
  • the video of the virtual person generated by the application can include the virtual person corresponding to the voice or text input by the user.
  • the virtual person will make human-like movements, that is, it can make body movements that match the human voice when speaking the input voice or the voice corresponding to the text.
  • the user can open a virtual human generation application installed on the terminal device and input voice or text (which can be active input or passive collection, such as through an audio sensor on the terminal device). collected), the virtual human generation application can generate a virtual human or virtual human action data based on speech or text through the method provided by the embodiment of the present application, and present the virtual human or virtual human action data to the user (The presentation method can be but is not limited to display, save, upload to the cloud, etc.).
  • the user can open a virtual human generation application installed on the terminal device and input voice or text (which can be active input or passive collection, such as through an audio sensor on the terminal device). collected), the virtual human generation application can send the voice or text to the server on the cloud side, and the server on the cloud side generates the virtual person or the action data of the virtual person based on the voice or text through the method provided by the embodiment of the present application. and transmits the virtual person, or the action data of the virtual person, back to the terminal device.
  • the terminal device can present the virtual person, or the action data of the virtual person, to the user (the presentation method may be, but is not limited to, displaying, saving, or uploading to the cloud). side, etc.).
  • the virtual human generation implemented by the virtual human generation application can be specifically used to enable augmented reality (augmented reality, AR), virtual reality (virtual reality, VR), and mixed reality (mixed reality, MR). ) Virtual character driver in application scenarios such as remote conferencing, sports and health, and the metaverse.
  • augmented reality augmented reality
  • VR virtual reality
  • MR mixed reality
  • Virtual character driver in application scenarios such as remote conferencing, sports and health, and the metaverse.
  • Figure 1B is a schematic diagram of the functional architecture of a virtual human generation application in an embodiment of the present application:
  • the virtual human generation application 102 may receive input parameters 101 (for example, including voice or text of the human body) and generate the action data 103 of the virtual person (or the information of the virtual character restored based on the action data of the virtual person).
  • the avatar generation application 102 is executable, for example, on at least one computer system and includes computer code that, when executed by one or more computers, causes the computers to perform operations as described herein. Describe the data processing methods.
  • Figure 2 is a schematic diagram of the physical architecture of running a virtual human generation application in an embodiment of the present application:
  • FIG. 2 shows a schematic diagram of a system architecture.
  • 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 for illustration in FIG. 2), and the server 200 may provide virtual human generation services for one or more terminals.
  • the terminal 100 can be installed with a virtual human generation application, or open a web page related to virtual human generation.
  • the above application program and web page can provide an interface, and the terminal 100 can receive relevant parameters input by the user on the virtual human generation interface. , and sends 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.
  • the terminal 100 can also complete the action of obtaining data processing results based on the received parameters by itself without requiring the cooperation of the server, which is not limited by the embodiments of this application.
  • the terminal 100 in the embodiment of the present application can be a mobile phone, a tablet computer, a wearable device, a vehicle-mounted device, an augmented reality (AR)/virtual reality (VR) device, a notebook computer, or an ultra mobile personal computer (ultra mobile personal computer).
  • - mobile personal computer UMPC
  • netbook personal digital assistant
  • PDA personal digital assistant
  • FIG. 3 shows an optional hardware structure diagram of the terminal 100.
  • the terminal 100 may include 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), Microphone 162 (optional), processor 170, external interface 180, power supply 190 and other components.
  • a radio frequency unit 110 may include 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), Microphone 162 (optional), processor 170, external interface 180, power supply 190 and other components.
  • Figure 3 is only an example of a terminal or a multi-function device, and does not constitute a limitation to the terminal or multi-function device. It may include more or fewer components than shown in the figure, or some components may be combined. Or different parts.
  • the input unit 130 may be used to receive input numeric or character information and generate key signal input related to user settings and function control of the portable multi-function device.
  • the input unit 130 may include a touch screen 131 (optional) and/or other input devices 132.
  • the touch screen 131 can collect the user's touch operations on or near it (such as the user's operations on or near the touch screen using fingers, knuckles, stylus, or any other suitable objects), and drive the corresponding according to the preset program. Connect the device.
  • the touch screen can detect the user's touch action on the touch screen, convert the touch action into a touch signal and send it to the processor 170, and can receive and execute commands from the processor 170; the touch signal at least includes contact point coordinate information.
  • the touch screen 131 can provide an input interface and an output interface between the terminal 100 and the user.
  • touch screens can be implemented in various types such as resistive, capacitive, infrared, and surface acoustic wave.
  • the input unit 130 may also include other input devices.
  • other input devices 132 may include, but are not limited to, one or more of a physical keyboard, function keys (such as volume control keys, switch keys 133, etc.), trackball, mouse, joystick, etc.
  • other input devices 132 can receive input voice or text, etc.
  • the display unit 140 may be used to display information input by the user or information provided to the user, various menus of the terminal 100, interactive interfaces, file display, and/or playback of any kind of multimedia files.
  • the display unit 140 may be used to display the interface of a virtual human generation application program, a virtual human based on speech or text, etc.
  • the memory 120 can be used to store instructions and data.
  • the memory 120 can mainly include a storage instruction area and a storage data area.
  • the storage data area can store various data, such as multimedia files, text, etc.;
  • the storage instruction area can store operating systems, applications, at least Software units such as instructions required for a function, or their subsets or extensions.
  • Non-volatile random access memory may also be included; providing the processor 170 with management of hardware, software and data resources in the computing processing device and supporting control software and applications. It is also used for storage of multimedia files, as well as storage of running programs and applications.
  • the processor 170 is the control center of the terminal 100. It uses various interfaces and lines to connect various parts of the entire terminal 100, and executes various functions of the terminal 100 by running or executing instructions stored in the memory 120 and calling data stored in the memory 120. functions and process data to provide overall control of the terminal device.
  • the processor 170 may include one or more processing units; preferably, the processor 170 may integrate an application processor and a modem processor, where the application processor mainly processes operating systems, user interfaces, application programs, etc. , the modem processor mainly handles wireless communications. It can be understood that the above modem processor may not be integrated into the processor 170 .
  • the processor and memory can be implemented on a single chip, and in some embodiments, they can also be implemented on separate chips.
  • the processor 170 can also be used to generate corresponding operation control signals, send them to corresponding components of the computing processing device, read and process data in the software, especially read and process the data and programs in the memory 120, so that the Each functional module performs a corresponding function, thereby controlling the corresponding components to act according to the instructions.
  • the memory 120 can be used to store software codes related to the data processing method, and the processor 170 can execute the steps of the data processing method of the chip, and can also schedule other units (such as the above-mentioned input unit 130 and the display unit 140) to implement corresponding functions. .
  • the radio frequency unit 110 (optional) can be used to send and receive information or receive and send signals during calls. For example, after receiving downlink information from the base station, it is processed by the processor 170; in addition, the designed uplink data is sent to the base station.
  • RF circuits include but are not limited to antennas, at least one amplifier, transceivers, couplers, low noise amplifiers (LNA), duplexers, etc.
  • the radio frequency unit 110 can also communicate with network devices and other devices through wireless communication.
  • the 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 (Code Division) Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Long Term Evolution (LTE), email, Short Messaging Service (SMS), etc.
  • GSM Global System of Mobile communication
  • GPRS General Packet Radio Service
  • CDMA Code Division Multiple Access
  • WCDMA Wideband Code Division Multiple Access
  • LTE Long Term Evolution
  • SMS Short Messaging Service
  • the radio frequency unit 110 can send voice or text to the server 200, and receive the action data of the virtual person sent by the server 200 or the information of the virtual character restored based on the action data of the virtual person.
  • radio frequency unit 110 is optional and can be replaced by other communication interfaces, such as a network port.
  • the terminal 100 also includes a power supply 190 (such as a battery) that supplies power to various components.
  • a power supply 190 such as a battery
  • the power supply can be logically connected to the processor 170 through a power management system, so that functions such as charging, discharging, and power consumption management can be implemented through the power management system.
  • the terminal 100 also includes an external interface 180, which can be a standard Micro USB interface or a multi-pin connector, which can be used to connect the terminal 100 to communicate with other devices, or can be used to connect a charger to charge the terminal 100. .
  • an external interface 180 can be a standard Micro USB interface or a multi-pin connector, which can be used to connect the terminal 100 to communicate with other devices, or can be used to connect a charger to charge the terminal 100.
  • the terminal 100 may also include a flash light, a wireless fidelity (WiFi) module, a Bluetooth module, sensors with different functions, etc., which will not be described again here. Some or all of the methods described below may be applied in the terminal 100 shown in FIG. 3 .
  • WiFi wireless fidelity
  • Bluetooth Bluetooth
  • FIG 4 provides a schematic structural diagram of a server 200.
  • 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 through the bus 201.
  • the bus 201 may be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc.
  • PCI peripheral component interconnect
  • EISA extended industry standard architecture
  • the bus can be divided into address bus, data bus, control bus, etc. For ease of presentation, only one thick line is used in Figure 4, but it does not mean that there is only one bus or one type of bus.
  • the processor 202 may be a central processing unit (CPU), a graphics processing unit (GPU), a microprocessor (MP) or a digital signal processor (DSP). any one or more of them.
  • CPU central processing unit
  • GPU graphics processing unit
  • MP microprocessor
  • DSP digital signal processor
  • Memory 204 may include volatile memory, such as random access memory (RAM).
  • RAM random access memory
  • the memory 204 may also include non-volatile memory (non-volatile memory), such as read-only memory (ROM), flash memory, mechanical hard drive (hard drive drive, HDD) or solid state drive (solid state drive). , SSD).
  • ROM read-only memory
  • HDD hard drive drive
  • SSD solid state drive
  • the memory 204 can be used to store software codes related to the data processing method, and the processor 202 can execute the steps of the data processing method of the chip, and can also schedule other units to implement corresponding functions.
  • the terminal 100 and the server 200 may be centralized or distributed devices, and the processors (such as the processor 170 and the processor 202) in the terminal 100 and the server 200 may be hardware circuits (such as application specific integrated circuits) application specific integrated circuit (ASIC), field-programmable gate array (FPGA), general-purpose processor, digital signal processing (DSP), microprocessor or microcontroller, etc.), Or a combination of these hardware circuits.
  • the processor can be a hardware system with the function of executing instructions, such as CPU, DSP, etc., or a hardware system without the function of executing instructions, such as ASIC, FPGA, etc., or the above-mentioned processor without the function of executing instructions.
  • the steps related to the model reasoning process in the embodiments of this application involve AI-related operations.
  • the instruction execution architecture of the terminal device and server is not limited to the processor-memory architecture introduced above.
  • the system architecture provided by the embodiment of the present application will be introduced in detail below with reference to Figure 5 .
  • FIG. 5 is a schematic diagram of the system architecture provided by the embodiment of the present application.
  • 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 collection system 560.
  • the execution device 510 includes a computing module 511, an I/O interface 512, a preprocessing module 513 and a preprocessing module 514.
  • the target model/rule 501 may be included in the calculation module 511, and the preprocessing module 513 and the preprocessing module 514 are optional.
  • the execution device 510 may be a terminal device or a server that runs the character virtual human generation application.
  • Data collection device 560 is used to collect training samples.
  • Training samples can be speech or text, as well as annotations of characters in speech or text (such as real action data of characters), etc.
  • the data collection device 560 stores the training samples into the database 530 .
  • the training device 520 can maintain the training samples in the database 530 and the neural network to be trained (such as the feature extraction network, action generation network, encoder, etc. in the embodiment of the present application) to obtain the target model/rule 501.
  • the neural network to be trained such as the feature extraction network, action generation network, encoder, etc. in the embodiment of the present application
  • the training samples maintained in the database 530 are not necessarily collected from the data collection device 560, and may also be received from other devices.
  • the training device 520 may not necessarily train the target model/rules 501 based entirely on the training samples maintained by the database 530. It may also obtain training samples from the cloud or other places for model training. The above description should not be used as a guarantee for this application. Limitations of Examples.
  • the target model/rules 501 trained according to the training device 520 can be applied to different systems or devices, such as the execution device 510 shown in Figure 5.
  • the execution device 510 can be a terminal, such as a mobile phone terminal, a tablet computer, and a notebook.
  • AR augmented reality
  • VR virtual reality
  • the training device 520 can transfer the trained model to the execution device 510 .
  • 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 to the I/O interface 512 through the client device 540 (for example, this Voice or text, etc. in the application embodiment).
  • the preprocessing module 513 and the preprocessing module 514 are used to perform preprocessing according to the input data received by the I/O interface 512. It should be understood that there may be no preprocessing module 513 and 514 or only one preprocessing module. When the preprocessing module 513 and the preprocessing module 514 do not exist, the computing module 511 can be directly used to process the input data.
  • the execution device 510 When the execution device 510 preprocesses input data, or when the calculation module 511 of the execution device 510 performs calculations and other related processes, the execution device 510 can call data, codes, etc. in the data storage system 550 for corresponding processing. , the data, instructions, etc. obtained by corresponding processing can also be stored in the data storage system 550.
  • the I/O interface 512 provides the processing results (such as the action data of the virtual person or the information of the virtual person restored based on the action data of the virtual person) to the client device 540, thereby providing it to the user.
  • the user can manually set the input data, and the "manually set input data" can be operated through the interface provided by the I/O interface 512 .
  • the client device 540 can automatically send input data to the I/O interface 512. If requiring the client device 540 to automatically send the input data requires the user's authorization, the user can set corresponding permissions in the client device 540. The user can view the results output by the execution device 510 on the client device 540, and the specific presentation form may be display, sound, action, etc.
  • the client device 540 can also be used as a data collection terminal to collect the input data of the input I/O interface 512 and the output results of the output I/O interface 512 as new sample data, and store them in the database 530.
  • the I/O interface 512 directly inputs the input data into the I/O interface 512 and outputs the output result of the I/O interface 512 as shown in the figure. Stored in database 530 as new sample data.
  • Figure 5 is only a schematic diagram of a system architecture provided by an embodiment of the present application.
  • the positional relationship between the devices, devices, modules, etc. shown in the figure does not constitute any limitation.
  • the data The 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 above execution device 510 may be deployed in the client device 540.
  • 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 inference process in the embodiment of this application.
  • the computing module 511 of the execution device 520 may include hardware circuits (such as application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), general-purpose processors, digital signal processing (DSP), microprocessor or microcontroller, etc.), or a combination of these hardware circuits.
  • the training device 520 can be a hardware system with the function of executing instructions, such as a CPU, DSP, etc. , or it is a hardware system that does not have the function of executing instructions, such as ASIC, FPGA, etc., or it is a combination of the above-mentioned hardware system that does not have the function of executing instructions and a hardware system that has the function of executing instructions.
  • the calculation module 511 of the execution device 520 can be a hardware system with the function of executing instructions.
  • the steps related to the model reasoning process provided by the embodiment of the present application can be software codes stored in the memory.
  • the calculation module 511 of the execution device 520 The software code can be obtained from the memory, and the obtained software code can be executed to implement the steps related to the model inference process provided by the embodiments of the present application.
  • 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 inference process provided by the embodiments of the present application can also be implemented by The computing module 511 of the execution device 520 is implemented by a hardware system that does not have the function of executing instructions, which is not limited here.
  • the above-mentioned training device 520 can obtain the code stored in the memory (not shown in Figure 5, which can be integrated with the training device 520 or deployed separately from the training device 520) to implement the model training in the embodiment of the present application. Related steps.
  • the training device 520 may include hardware circuits (such as application specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), general-purpose processors, digital signal processors (digital signal processing, DSP, microprocessor or microcontroller, etc.), or a combination of these hardware circuits.
  • ASICs application specific integrated circuits
  • FPGAs field-programmable gate arrays
  • DSP digital signal processors
  • the training device 520 can be a hardware system with the function of executing instructions, such as a CPU, DSP, etc., or a combination of other hardware circuits.
  • a hardware system with the function of executing instructions such as ASIC, FPGA, etc., or a combination of the above-mentioned hardware systems without the function of executing instructions and a hardware system with the function of executing instructions.
  • 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 by the embodiments of the present application can also be implemented by the training device 520 that does not have the function of executing instructions. It is implemented by the hardware system that executes the instruction function, which is not limited here.
  • the server can provide virtual human-generated services to the client side through an application programming interface (API).
  • API application programming interface
  • the terminal device can send relevant parameters (such as voice or text) to the server through the API provided by the cloud.
  • the server can obtain the processing results (such as virtual human action data or virtual human-based action data restoration based on the received parameters). the obtained avatar information, etc.), and return the processing results to the terminal.
  • Figure 6 shows the process of using a virtual human generation cloud service provided by a cloud platform.
  • SDK software development kit
  • the cloud platform provides multiple development versions of SDK for users to choose according to the needs of the development environment, such as JAVA version of SDK and python version. of SDK, PHP version of SDK, Android version of SDK, etc.
  • the local development environment can also develop other functions, forming a collection of virtual Application of human generative abilities.
  • the API call for virtual human generation can be triggered.
  • the application triggers the virtual human generation function, it initiates an API request to the running instance of the virtual human generation service in the cloud environment.
  • the API request carries voice or text, and the running instance in the cloud environment processes the voice or text.
  • Obtain the processing results (such as the action data of the virtual person or the information of the virtual character restored based on the action data of the virtual person, etc.).
  • the cloud environment returns the processing results to the application, thus completing a virtual human generation service call.
  • the server can provide a virtual human-generated model adapted to the character object (or a more generalized character) in the speech or text based on the speech or text provided by the client.
  • the server can provide image information restoration services for the client side through an application programming interface (API).
  • API application programming interface
  • the terminal device can send relevant parameters (such as voice or text) to the server through the API provided by the cloud.
  • the server can obtain the processing results based on the received parameters and adapt the processing results (such as adapting to the voice or text) to the server.
  • the character object (or a more generalized character), a virtual human-generated model, etc.) is returned to the terminal.
  • Figure 7 shows the process of using a model training cloud service provided by a cloud platform.
  • the neural network can be composed of neural units.
  • the neural unit can refer to an arithmetic unit that takes xs and intercept 1 as input.
  • the output of the arithmetic unit can be:
  • s 1, 2,...n, n is a natural number greater than 1
  • Ws is the weight of xs
  • 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 an output signal. The output signal of this 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.
  • Deep Neural Network can be understood as a neural network with many hidden layers. There is no special metric for "many” here. The essence of what we often call multi-layer neural networks and deep neural networks is It's the same thing. From the division of DNN according to the position of different layers, the neural network inside 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 hidden layers. The layers are fully connected, that is to say, any neuron in the i-th layer must be connected to any neuron in the i+1-th layer. Although DNN looks very complicated, the work of each layer is actually not complicated.
  • 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 number of layers where the coefficient W is located, and the subscript corresponds to the output third layer index 2 and the input second layer index 4.
  • the coefficient from the k-th neuron in layer L-1 to the j-th neuron in layer L is defined as Note that the input layer has no W parameter.
  • more hidden layers make the network more capable of describing complex situations in the real world. Theoretically, the more parameters a model has, the higher its complexity and “capacity” will be. The larger it is, which means it can complete more complex learning tasks.
  • Convolutional neural network (Convosutionas Neuras Network, CNN) is a deep neural network with a convolutional structure.
  • the convolutional neural network contains a feature extractor composed of convolutional layers and subsampling layers.
  • the feature extractor can be regarded as a filter, and the convolution process can be regarded as using a trainable filter to convolve with an input image or convolution feature plane (feature map).
  • the convolutional layer refers to the neuron layer in the convolutional neural network that convolves the input signal.
  • a neuron can be connected to only some of the neighboring layer neurons.
  • a convolutional layer usually contains several feature planes, and each feature plane can be composed of some rectangularly arranged neural units.
  • Neural units in the same feature plane share weights, and the shared weights here are convolution kernels.
  • Shared weights can be understood as a way to extract image information independent of position. The underlying principle is that the statistical information of one part of the image is the same as that of other parts. This means that the image information learned in one part can also be used in another part. So for all positions on the image, we can use the same learned image information.
  • multiple convolution kernels can be used to extract different image information. Generally, the greater the number of convolution kernels, the richer the image information reflected by the convolution operation.
  • the convolution kernel can be initialized in the form of a random-sized matrix. During the training process of the convolutional neural network, the convolution kernel can obtain reasonable weights through learning. In addition, the direct benefit of sharing weights is to reduce the connections between the layers of the convolutional neural network, while reducing the risk of overfitting.
  • the convolutional neural network can use the error back propagation (BP) algorithm to modify the size of the parameters in the initial super-resolution model during the training process, so that the reconstruction error loss of the super-resolution model becomes smaller and smaller.
  • BP error back propagation
  • forward propagation of the input signal until the output will produce an error loss
  • the parameters in the initial super-resolution model are updated by back-propagating the error loss information, so that the error loss converges.
  • the backpropagation algorithm is a backpropagation movement dominated by error loss, aiming to obtain the optimal parameters of the super-resolution model, such as the weight matrix.
  • the neural network can use the error back propagation (BP) algorithm to modify the size of the parameters in the initial neural network model during the training process, so that the reconstruction error loss of the neural network model becomes smaller and smaller. Specifically, forward propagation of the input signal until the output will produce an error loss, and the parameters in the initial neural network model are updated by backpropagating the error loss information, so that the error loss converges.
  • the backpropagation algorithm is a backpropagation movement dominated by error loss, aiming to obtain the optimal parameters of the neural network model, such as the weight matrix.
  • body language can increase the rhythm of the speech, making it more vivid and persuasive.
  • body movements can more accurately express intentions and convey emotions, complementing the information conveyed by voice;
  • body movements can help users focus more on the content communicated with digital people;
  • they can improve the persuasiveness and credibility of digital people. degree and realism;
  • it can reflect the speaker’s intention and personality. Lack of body language or rigid body movements in communication will lead to the uncanny valley effect.
  • FIG 8 is a flow diagram of a data processing method provided by an embodiment of the present application.
  • a data processing method provided by an embodiment of the present application may include steps 801 to 804. The following are respectively These steps are described in detail.
  • the user's voice data can be obtained and corresponding action data can be generated based on the voice data.
  • This motion data can be used to generate the corresponding virtual human.
  • the user's text data can be obtained, corresponding voice data can be generated based on the text data, and corresponding action data can be generated based on the voice data (optionally, text data can also be included).
  • This motion data can be used to generate the corresponding virtual human.
  • the user's voice data can be obtained, the reply text is determined based on the text in the voice data, and the corresponding voice data is generated based on the reply text, and then the reply can be based on the voice data (optionally, it can also include text) to generate corresponding action data.
  • This motion data can be used to generate the corresponding virtual human.
  • the action data can be gesture actions, etc.
  • the segmentation point positions correspond to predicted rhythm points of body movements made by the character object when emitting the speech data.
  • the action data of the character object that emits the voice data can be predicted based on the voice data.
  • the user makes a sound for example, when giving a speech
  • the rhythm point is the critical point from stillness to movement or from movement to stillness when making the body movements. point. If the rhythm information can be explicitly recognized before feature extraction and used as the input of the model, the subsequently generated actions can have an accurate sense of rhythm, thus making the actions more consistent with the actions of the characters when they actually make speech.
  • multiple segmentation point positions may be determined from the voice data according to audio features of the voice data, where the audio features are related to at least one of volume and pitch.
  • rhythm points usually appear on the onset of speech (audio features related to volume), but not all onsets are rhythm points of user actions.
  • the intervals between adjacent rhythm points of speech There is a range, and that range is related to the speaker's style. Assuming that the range is [t1, t2], usually time intervals smaller than t1 are caused by noise, filler words, or unsmooth pronunciation. Intervals beyond t2 usually correspond to pauses or periods of quiet silence. Based on these observations, an episode whose time interval between adjacent episodes is within a preset range can be determined as a rhythm point, and then the rhythm point of the body movement made by the user when emitting the voice data can be identified, and based on the rhythm Click to segment the voice data.
  • the first position and the second position are adjacent dividing point positions among multiple dividing point positions.
  • the first position and the second position are determined in the following manner: the first position and the second position are onset points in the speech data, and the The time interval between onset points is within a preset range.
  • the preset range can be [t1, t2] introduced above.
  • a heuristic strategy can be used to identify speech rhythms. Segment the speech, and those less than t1 will not be recognized as rhythm points, while pseudo-rhythm points can be inserted for those exceeding t2. Pseudo-rhythm points can be inserted in the following way: the time from the previous rhythm point exceeds t1, and the volume of the voice exceeds a certain threshold I a . This threshold can be set to the average volume of environmental noise so that adjacent segmentation points The time interval of the positions is within the preset range. If all the volumes in the interval are less than I a , insert the minimum pseudo-rhythm points evenly so that the time interval between adjacent dividing point positions is within the preset range.
  • the plurality of dividing point positions include adjacent first positions and second positions; the first position and the second position may be determined in the following manner: the first The position is the onset point in the voice data, the second position is not the onset point in the voice data, and the second position is the time point in the voice data when the volume is greater than the threshold.
  • the voice data and the information indicating the positions of the multiple segmentation points obtain a feature representation through a feature extraction network.
  • the speech data can be segmented according to the position of the segmentation point to obtain multiple speech segments, and the multiple speech segments can be input into the feature network.
  • the text data corresponding to the voice data can be divided according to the position of the dividing point to obtain multiple text segments, and each text segment can have a one-to-one correspondence with multiple voice segments.
  • the input speech can be segmented according to the above-mentioned identified rhythm points, and the segmented speech segments are normalized to the length of t2.
  • the action features and text features corresponding to this speech are also resampled to the length of t2.
  • the voice data and the characters indicating the positions of the multiple segmentation points can be input into a feature extraction network to obtain a feature representation, and based on the feature representation, the action generation network is used to generate action data .
  • the type of action in addition to inputting the feature representation obtained through the feature extraction network into the action generation network, the type of action can also be identified through a neural network based on the feature representation (such as high-level features in the feature representation). , and input action types and feature representations into the action generation network.
  • all actions in the data set can be divided into equal-length action segments.
  • the categories corresponding to these actions can be obtained through the neural network, and these action categories can be considered to be the high-level features of the actions.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes the feature representation output by the multiple network layers.
  • Feature representations that is, high-level features
  • generating action data through an action generation network according to the feature representations includes: according to the multiple Part or all of the feature representations output by each network layer, as well as the action categories, are passed through the action generation network to generate action data.
  • the action category itself carries semantics
  • the features input to the action generation network and the subsequently generated body movements have semantic relevance.
  • the VQ-VAE encoder can know that its corresponding high-level feature gesture vocabulary (action category) is Si, and the style code Z i corresponding to the action segment can be decomposed.
  • the method is as follows : As shown in Figure 10, the action Mi -1 of the previous segment is passed through the action encoder to obtain its corresponding action features.
  • the low-level features of the speech in the current period Compared with the low-level speech features of the upper and lower periods and Get the audio features of the current period through the audio encoder Will S i is fused with the learnable gesture style code Z i and input into the action generation network to generate the final action.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes the feature representation output by the multiple network layers, which can be based on the output layer of the multiple network layers.
  • the feature representation output by one or more network layers is passed through the encoder to obtain the style encoding of the body movement;
  • the action generation network is used to generate action data according to the feature representation, which specifically includes: according to the plurality of Part or all of the feature representation output by the network layer and the style encoding are passed through the action generation network to generate action data.
  • the distribution of the style codes is uniform distribution.
  • the decoupling diagram of the high-level features and low-level features introduced above can be referred to as shown in Figure 9: where, given audio and its corresponding text, a language pre-training model is used as a supervision signal for comparative learning. , so that the high-level features of the pre-trained model of speech are close to the high-level features of text, and the low-level features of the pre-trained model of speech are close to the low-level features of text.
  • the action data can be the joint point rotation angle of the body action or the 3D point cloud information of the action, which is not limited here.
  • the method provided in the embodiment of this application can segment the speech segment into normalized feature blocks. Then generate corresponding actions in sequence
  • the overall process is shown in Figure 11A, where,
  • gesture vocabulary and gesture style code They cannot be obtained directly from the input speech, they are obtained by inference from the action category prediction module and the action style prediction module.
  • text data can also be used as input.
  • the text solution can be simplified.
  • audio input without its corresponding text or speaker ID, its corresponding action can also be inferred.
  • the flow chart is shown in Figure 10. In this embodiment, the input is only audio.
  • Figure 11B is a flowchart of a data processing method provided by an embodiment of the present application. The method includes:
  • the voice data and the first action data of the character object are the real action data of the body movements made by the character object when the voice data is emitted;
  • the first action data may be the true value of the action data corresponding to the voice data.
  • the segmentation point positions correspond to the predicted rhythm points of the body movements made by the character object when emitting the speech data;
  • step 1102 For a detailed description of step 1102, please refer to the introduction of step 802 in the above embodiment, and will not be described again here.
  • step 1103 For a detailed description of step 1103, please refer to the introduction of step 803 in the above embodiment, and will not be described again here.
  • the feature representation generate second action data through the action generation network; the difference between the first action data and the second action data is used to update the feature extraction network and the action generation network .
  • step 1104 For a detailed description of step 1104, please refer to the introduction of step 804 in the above embodiment, and will not be described again here.
  • L gen w rec L rec +w perc L perc +w lexeme L lexeme +w z L z ...(1)
  • N J is a random subset of all motion segments in the training set
  • N J is the size of J.
  • N J can be adjusted based on the size of the data set.
  • ⁇ z and are the mean and variance of the gesture style codes in the training mini-batch respectively.
  • the feature extraction network includes multiple network layers connected in series
  • the feature representation includes feature representations output by the multiple network layers
  • the network is generated through actions according to the feature representation.
  • generating second action data including: generating second action data through an action generation network according to the feature representation and initialized style coding; and outputting the second action data according to one or more network layers far away from the output layer among the plurality of network layers.
  • Characteristic representation through the encoder, obtain the first style code of the body movement; through the preset loss function, update the initialized style code to obtain the updated style code;
  • the difference between the first style encoding and the updated style encoding is used to update the encoder.
  • the audio characteristics are related to volume and/or pitch.
  • the plurality of dividing point positions include adjacent first positions and second positions, and the time interval between the second position and the first position is within a preset range; the third A position and the second position are determined in one of the following ways:
  • the first position and the second position are the onset points in the speech data
  • the first position is an onset point in the voice data
  • the second position is not an onset point in the voice data
  • the second position is a time point in the voice data when the volume is greater than a threshold.
  • the voice data and information indicating the locations of the multiple dividing points include:
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the method further includes: according to the multiple network layers The feature representation output by one or more network layers close to the output layer in the layer determines the action category of the body action;
  • Generating second action data through an action generation network based on the feature representation includes:
  • Second action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the action category.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the method further includes: according to the multiple network layers The feature representation output by one or more network layers far away from the output layer in the layer is passed through the encoder to obtain the style encoding of the body action; the second action data is generated through the action generation network according to the feature representation, including : Generate second action data through an action generation network based on part or all of the feature representations output by the plurality of network layers and the style encoding.
  • FIG. 12A is a flow diagram for generating action data using text data.
  • FIG. 12B is a flow diagram for generating action data without using text data.
  • Figure 13 is a schematic structural diagram of a data processing device provided by an embodiment of the present application.
  • a data processing device 1300 provided by an embodiment of the present application includes:
  • step 801 for a specific description of the acquisition module 1301, please refer to the introduction of step 801 in the above embodiment, and will not be described again here.
  • the processing module 1302 is configured to determine multiple segmentation point positions from the speech data according to the audio characteristics of the speech data.
  • the segmentation point positions correspond to the body movements made by the character object when emitting the speech data. Predict rhythm points;
  • a feature representation is obtained through a feature extraction network
  • action data is generated through an action generation network.
  • processing module 1302 For a specific description of the processing module 1302, reference may be made to the introduction of steps 802 to 804 in the above embodiment, and will not be described again here.
  • the audio characteristics are related to volume and/or pitch.
  • the plurality of dividing point positions include adjacent first positions and second positions, and the time interval between the second position and the first position is within a preset range; the third A position and the second position are determined in one of the following ways:
  • the first position and the second position are the onset points in the speech data
  • the first position is an onset point in the voice data
  • the second position is not an onset point in the voice data
  • the second position is a time point in the voice data when the volume is greater than a threshold.
  • the voice data and information indicating the locations of the multiple dividing points include:
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is also used to:
  • the processing module is specifically used for:
  • Action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the action category.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is also used to:
  • the style encoding of the body movement is obtained through the encoder
  • the processing module is specifically used for:
  • Action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the style encoding.
  • the distribution of the style codes is uniform distribution.
  • Figure 14 is a schematic structural diagram of a data processing device provided by an embodiment of the present application.
  • a data processing device 1400 provided by an embodiment of the present application includes:
  • the acquisition module 1401 is used to obtain the voice data and the first action data of the character object.
  • the first action data is the real action data of the body movements made by the character object when the voice data is emitted;
  • step 1101 for a specific description of the acquisition module 1401, please refer to the introduction of step 1101 in the above embodiment, and will not be described again here.
  • the processing module 1402 is configured to determine multiple segmentation point positions from the speech data according to the audio characteristics of the speech data.
  • the segmentation point positions correspond to the body movements made by the character object when emitting the speech data. Predict rhythm points;
  • a feature representation is obtained through a feature extraction network
  • second action data is generated through an action generation network; the difference between the first action data and the second action data is used to update the feature extraction network and the action generation network.
  • processing module 1402 For a specific description of the processing module 1402, please refer to the introduction of steps 1102 to 1104 in the above embodiment, and will not be described again here.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is specifically used to:
  • the first style encoding of the body movement is obtained through the encoder
  • the difference between the first style encoding and the updated style encoding is used to update the encoder.
  • the audio characteristics are related to volume and/or pitch.
  • the plurality of dividing point positions include adjacent first positions and second positions, and the time interval between the second position and the first position is within a preset range; the third A position and the second position are determined in one of the following ways:
  • the first position and the second position are the onset points in the speech data
  • the first position is an onset point in the voice data
  • the second position is not an onset point in the voice data
  • the second position is a time point in the voice data when the volume is greater than a threshold.
  • the voice data and information indicating the locations of the multiple dividing points include:
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is also used to:
  • the processing module is specifically used for:
  • Second action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the action category.
  • the feature extraction network includes multiple network layers connected in series, and the feature representation includes feature representations output by the multiple network layers.
  • the processing module is also used to:
  • the style encoding of the body movement is obtained through the encoder
  • the processing module is specifically used for:
  • Second action data is generated through an action generation network based on part or all of the feature representations output by the plurality of network layers and the style encoding.
  • FIG. 15 is a schematic structural diagram of an execution device provided by an embodiment of the present application.
  • the execution device 1500 can be embodied as a virtual reality VR device, a mobile phone, Tablets, laptops, smart wearable devices, monitoring data processing equipment or servers, etc. are not limited here.
  • the execution device 1500 includes: a receiver 1501, a transmitter 1502, a processor 1503 and a memory 1504 (the number of processors 1503 in the execution device 1500 can be one or more, one processor is taken as an example in Figure 15) , wherein the processor 1503 may include an application processor 15031 and a communication processor 15032.
  • the receiver 1501, the transmitter 1502, the processor 1503, and the memory 1504 may be connected through a bus or other means.
  • Memory 1504 may include read-only memory and random access memory and provides instructions and data to processor 1503 .
  • a portion of memory 1504 may also include non-volatile random access memory (NVRAM).
  • NVRAM non-volatile random access memory
  • the memory 1504 stores processor and operating instructions, executable modules or data structures, or a subset thereof, or an extended set thereof, where the operating instructions may include various operating instructions for implementing various operations.
  • the processor 1503 controls the execution of operations of the device.
  • various components of the execution device are coupled together through a bus system.
  • the bus system may also include a power bus, a control bus, a status signal bus, etc.
  • various buses are called bus systems in the figure.
  • the methods disclosed in the above embodiments of the present application can be applied to the processor 1503 or implemented by the processor 1503.
  • the processor 1503 may be an integrated circuit chip with signal processing capabilities. During the implementation process, each step of the above method can be completed by instructions in the form of hardware integrated logic circuits or software in the processor 1503 .
  • the above-mentioned processor 1503 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 (field-programmable gate array, FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components.
  • DSP digital signal processor
  • ASIC application specific integrated circuit
  • FPGA field-programmable gate array
  • the processor 1503 can implement or execute each method, step and logical block diagram disclosed in the embodiment of this application.
  • a general-purpose processor may be a microprocessor or the processor may be any conventional processor, etc.
  • the steps of the method disclosed in conjunction with the embodiments of the present application can be directly implemented by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor.
  • the software module can be located in random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, registers and other mature storage media in this field.
  • the storage medium is located in the memory 1504.
  • the processor 1503 reads the information in the memory 1504 and completes the steps related to the model inference process in the above method in combination with its hardware.
  • the receiver 1501 may be configured to receive input numeric or character information and generate signal inputs related to performing relevant settings and functional controls of the device.
  • the transmitter 1502 can be used to output numeric or character information through the first interface; the transmitter 1502 can also be used to send instructions to the disk group through the first interface to modify the data in the disk group; the transmitter 1502 can also include a display device such as a display screen .
  • FIG. 16 is a schematic structural diagram of the training device provided by the embodiment of the present application.
  • the training device 1600 is implemented by one or more servers.
  • the training device 1600 There may be relatively large differences due to different configurations or performance, which may include one or more central processing units (CPU) 1616 (for example, one or more processors) and memory 1632, one or more storage applications Storage medium 1630 for program 1642 or data 1644 (eg, one or more mass storage devices).
  • the memory 1632 and the storage medium 1630 may be short-term storage or persistent storage.
  • the program stored in the storage medium 1630 may include one or more modules (not shown in the figure), and each module may include Includes a series of command operations in the training device. Furthermore, the central processor 1616 may be configured to communicate with the storage medium 1630 and execute a series of instruction operations in the storage medium 1630 on the training device 1600 .
  • the training device 1600 may also include one or more power supplies 1626, one or more wired or wireless network interfaces 1650, one or more input and output interfaces 1658; or, one or more operating systems 1641, such as Windows ServerTM, Mac OS XTM , UnixTM, LinuxTM, FreeBSDTM and so on.
  • operating systems 1641 such as Windows ServerTM, Mac OS XTM , UnixTM, LinuxTM, FreeBSDTM and so on.
  • the central processor 1616 is used to perform actions related to model training in the above embodiment.
  • An embodiment of the present application also provides a computer program product that, when run on a computer, causes the computer to perform the steps performed by the foregoing execution device, or causes the computer to perform the steps performed by the foregoing training device.
  • Embodiments of the present application also provide a computer-readable storage medium.
  • the computer-readable storage medium stores a program for performing signal processing.
  • the program When the program is run on a computer, it causes the computer to perform the steps performed by the aforementioned execution device. , or, causing the computer to perform the steps performed by the aforementioned training device.
  • the execution device, training device or terminal device provided by the embodiment of the present application may specifically be a chip.
  • the chip includes: a processing unit and a communication unit.
  • the processing unit may be, for example, a processor.
  • the communication unit may be, for example, an input/output interface. Pins or circuits, 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.
  • the storage unit is a storage unit within the chip, such as a register, cache, etc.
  • the storage unit may also be a storage unit located outside the chip in the wireless access device, such as Read-only memory (ROM) or other types of static storage devices that can store static information and instructions, random access memory (random access memory, RAM), etc.
  • ROM Read-only memory
  • RAM random access memory
  • Figure 17 is a structural schematic diagram of a chip provided by an embodiment of the present application.
  • the chip can be represented as a neural network processor NPU 1700.
  • the NPU 1700 serves as a co-processor and is mounted to the main CPU (Host). CPU), tasks are allocated by the Host CPU.
  • the core part of the NPU is the arithmetic circuit 1703.
  • the arithmetic circuit 1703 is controlled by the controller 1704 to extract the matrix data in the memory and perform multiplication operations.
  • the computing circuit 1703 internally includes multiple processing units (Process Engine, PE).
  • arithmetic circuit 1703 is a two-dimensional systolic array.
  • the arithmetic circuit 1703 may also be a one-dimensional systolic array or other electronic circuit capable of performing mathematical operations such as multiplication and addition.
  • arithmetic circuit 1703 is a general-purpose matrix processor.
  • the arithmetic circuit obtains the corresponding data of matrix B from the weight memory 1702 and caches it on each PE in the arithmetic circuit.
  • the operation circuit takes matrix A data and matrix B from the input memory 1701 to perform matrix operations, and the partial result or final result of the matrix is stored in an accumulator (accumulator) 1708 .
  • the unified memory 1706 is used to store input data and output data.
  • the weight data directly passes through the storage unit access controller (Direct Memory Access Controller, DMAC) 1705, and the DMAC is transferred to the weight memory 1702.
  • Input data is also transferred to unified memory 1706 via DMAC.
  • DMAC Direct Memory Access Controller
  • BIU is the Bus Interface Unit, that is, the bus interface unit 1710, which is used for the interaction between the AXI bus and the DMAC and the Instruction Fetch Buffer (IFB) 1709.
  • IFB Instruction Fetch Buffer
  • the bus interface unit 1710 (Bus Interface Unit, BIU for short) is used to fetch the memory 1709 to obtain instructions from the external memory, and is also used for the storage unit access controller 1705 to obtain the original data of the input matrix A or the weight matrix B from the external memory.
  • BIU Bus Interface Unit
  • DMAC is mainly used to transfer the input data in the external memory DDR to the unified memory 1706 or the weight data to the weight memory 1702 or the input data to the input memory 1701 .
  • the vector calculation unit 1707 includes multiple arithmetic processing units. If necessary, the output of the arithmetic circuit 1703 is further processed, such as vector multiplication, vector addition, exponential operation, logarithmic operation, size comparison, etc. Mainly used for non-convolutional/fully connected layer network calculations in neural networks, such as Batch Normalization, pixel-level summation, upsampling of feature planes, etc.
  • vector calculation unit 1707 can store the processed output vectors to unified memory 1706 .
  • the vector calculation unit 1707 can apply a linear function; or a nonlinear function to the output of the operation circuit 1703, such as linear interpolation on the feature plane extracted by the convolution layer, or a vector of accumulated values, to generate an activation value.
  • vector computation unit 1707 generates into normalized values, pixel-wise summed values, or both.
  • the processed output vector can be used as an activation input to the arithmetic circuit 1703, such as for use in a subsequent layer in a neural network.
  • the instruction fetch buffer 1709 connected to the controller 1704 is used to store instructions used by the controller 1704;
  • the unified memory 1706, the input memory 1701, the weight memory 1702 and the fetch memory 1709 are all On-Chip memories. External memory is private to the NPU hardware architecture.
  • the processor mentioned in any of the above places can be a general central processing unit, a microprocessor, an ASIC, or one or more integrated circuits used to control the execution of the above programs.
  • the device embodiments described above are only illustrative.
  • the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physically separate.
  • the physical unit can be located in one place, or it can be distributed across multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment.
  • the connection relationship between modules indicates that there are communication connections between them, which can be specifically implemented as one or more communication buses or signal lines.
  • the present application can be implemented by software plus necessary general hardware. Of course, it can also be implemented by dedicated hardware including dedicated integrated circuits, dedicated CPUs, dedicated memories, Special components, etc. to achieve. In general, all functions performed by computer programs can be easily implemented with corresponding hardware. Moreover, the specific hardware structures used to implement the same function can also be diverse, such as analog circuits, digital circuits or special-purpose circuits. circuit etc. However, for this application, software program implementation is a better implementation in most cases. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence or that contributes to the existing technology.
  • the computer software product is stored in a readable storage medium, such as a computer floppy disk. , U disk, mobile hard disk, ROM, RAM, magnetic disk or optical disk, etc., including several instructions to cause a computer device (which can be a personal computer, training device, or network device, etc.) to execute the steps described in various embodiments of the present application. method.
  • a computer device which can be a personal computer, training device, or network device, etc.
  • the computer program product includes one or more computer instructions.
  • the computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device.
  • the computer instructions may be stored in or transmitted from one computer-readable storage medium to another, for example, the computer instructions may be transferred from a website, computer, training device, or data
  • the center transmits to another website site, computer, training equipment or data center through wired (such as coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (such as infrared, wireless, microwave, etc.) means.
  • wired such as coaxial cable, optical fiber, digital subscriber line (DSL)
  • wireless such as infrared, wireless, microwave, etc.
  • the computer-readable storage medium may be any available medium that a computer can store, or a data storage device such as a training device or a data center integrated with one or more available media.
  • the available media may be magnetic media (eg, floppy disk, hard disk, magnetic tape), optical media (eg, DVD), or semiconductor media (eg, solid state disk (Solid State Disk, SSD)), etc.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Computational Linguistics (AREA)
  • Health & Medical Sciences (AREA)
  • Audiology, Speech & Language Pathology (AREA)
  • Human Computer Interaction (AREA)
  • Physics & Mathematics (AREA)
  • Acoustics & Sound (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • User Interface Of Digital Computer (AREA)

Abstract

一种数据处理方法，应用于基于语音或文本实现的动作生成，所述方法包括：获取语音数据；根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；根据所述特征表示，通过动作生成网络，生成动作数据。本申请可以使得生成的动作具有准确的节奏感，进而使得动作更贴合人物对象在真实发出语音时的动作。

Description

一种数据处理方法及其装置
本申请要求于2022年9月2日提交中国专利局、申请号为202211071943.4、发明名称为“一种数据处理方法及其装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及人工智能领域，尤其涉及一种数据处理方法及其装置。
背景技术
在公共演讲中，肢体语言可以增加演讲的节奏感，让演讲更生动，更具有说服力。有研究表明，肢体语言在沟通交流中扮演重要角色。肢体动作可以更加准确地表达意图和传递情感，与语音传达的信息互补；其次，肢体动作能够帮助用户更专注于与数字人沟通的内容；且肢体动作能够提升数字人的说服力，可信度及真实感；肢体动作能够反应说话人的意图及个性。沟通中缺少肢体语言或者肢体动作僵化，会陷入恐怖谷效应。
人们期待虚拟数字人(或者可以简称为数字人)的行为更像人。虚拟数字人可以跟人一样，在演讲的时候，可以配合语音做出节奏性，强调性的动作。对于一些特殊的语音语义可以有特定的身体动作，例如，人们说“好”或者“ok”的时候，倾向于比OK的手势，说到“第一”或者“首先”的时候，倾向于微微停顿，比个1的手型。
因此，如何根据语音生成对应的手势等身体动作是一个亟待解决的问题。
发明内容
本申请提供了一种数据处理方法，可以使得生成的动作具有准确的节奏感，进而使得动作更贴合人物对象在真实发出语音时的动作。
第一方面，本申请提供了一种数据处理方法，所述方法包括：获取语音数据；根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；根据所述特征表示，通过动作生成网络，生成动作数据。
在一种可能的实现中，动作数据可以为身体动作的关节点旋转角度或者动作的3D点云信息，这里并不限定。
在一种可能的实现中，可以根据语音数据来预测发出该语音数据的人物对象的动作数据。其中，用户在发声时(例如进行演讲时)，会做出一定的动作，且该动作存在一定的节奏点，例如，节奏点可以为做出所述身体动作时由静止到运动或者由运动到静止的临界点。若能够在特征提取之前就显性的将节奏信息识别出来并作为模型的输入，可以使得后续生成的动作具有准确的节奏感，进而使得动作更贴合人物对象在真实发出语音时的动作。
在一种可能的实现中，所述音频特征与音量和/或音高有关。
在一种可能的实现中，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：所述第一位置和所述第二位置为所述语音数据中的onset点；所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
在一种可能的实现中，动作节奏点通常出现在语音的onset(和音量相关的音频特征)上，但却不是全部的onset都是用户动作的节奏点，语音的相邻节奏点之间的间隔是有一个范围的，该范围与讲者风格相关。假设该范围为[t1，t2]，通常小于t1的时间间隔是由于噪声、填充词或者不顺畅的发音引 起的。超过t2的时间间隔通常是对应着停顿或者是不说话的安静阶段。基于这些观察，可以将相邻onset的时间间隔在预设范围内的onset确定为是节奏点，进而可以识别出用户在发出所述语音数据时做出的身体动作的节奏点，并基于该节奏点对语音数据进行切分。
在一种可能的实现中，可以使用一个启发性的策略来识别语音节奏。将语音进行切分，小于t1的不识别为节奏点，而针对于超过t2的可以插入伪节奏点。可以用以下的方式插入伪节奏点：距离上一个节奏点的时间超过t1，并且语音的音量超过一定的阈值Ia，可以将该阈值设置为环境噪声的平均音量，以使得相邻的分割点位置的时间间隔在预设范围内。如果区间内所有的音量都小于Ia，就均匀地插入最少的伪节奏点，以使得相邻的分割点位置的时间间隔在预设范围内。
在一种可能的实现中，所述语音数据以及指示所述多个分割点位置的信息，包括：根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，所述语音数据以及指示所述多个分割点位置的字符。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；所述根据所述特征表示，通过动作生成网络，生成动作数据，包括：根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成动作数据。
在一种可能的实现中，除了将通过特征提取网络得到的特征表示输入到动作生成网络之外，还可以根据特征表示(例如特征表示中的高层次特征)通过神经网络来识别出动作的类型，并将动作类型和特征表示输入到动作生成网络。
由于动作类别本身携带语义，使得输入到动作生成网络的特征以及后续生成的身体动作具备语义相关性。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；所述根据所述特征表示，通过动作生成网络，生成动作数据，包括：根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成动作数据。
在一种可能的实现中，所述风格编码的分布为均匀分布。
第二方面，本申请提供了一种数据处理方法，所述方法包括：
获取人物对象的语音数据以及第一动作数据，所述第一动作数据为所述人物对象在发出所述语音数据时做出的身体动作的真实动作数据；
根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
根据所述特征表示，通过动作生成网络，生成第二动作数据；所述第一动作数据和所述第二动作数据之间的差异用于更新所述特征提取网络以及所述动作生成网络。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：
根据所述特征表示以及初始化的风格编码，通过动作生成网络，生成第二动作数据；
根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身 体动作的第一风格编码；
通过预设的损失函数，更新所述初始化的风格编码，以得到更新后的风格编码；
所述第一风格编码和所述更新后的风格编码之间的差异用于更新所述编码器。
在一种可能的实现中，所述音频特征与音量和/或音高有关。
在一种可能的实现中，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
所述第一位置和所述第二位置为所述语音数据中的onset点；
所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
在一种可能的实现中，所述语音数据以及指示所述多个分割点位置的信息，包括：
根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
所述语音数据以及指示所述多个分割点位置的字符。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：
根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成第二动作数据。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：
根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成第二动作数据。
第三方面，本申请提供了一种数据处理装置，所述装置包括：
获取模块，用于获取语音数据；
处理模块，用于根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
根据所述特征表示，通过动作生成网络，生成动作数据。
在一种可能的实现中，所述音频特征与音量和/或音高有关。
在一种可能的实现中，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
所述第一位置和所述第二位置为所述语音数据中的onset点；
所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所 述第二位置为所述语音数据中音量大于阈值的时间点。
在一种可能的实现中，所述语音数据以及指示所述多个分割点位置的信息，包括：
根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
所述语音数据以及指示所述多个分割点位置的字符。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
所述处理模块，具体用于：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成动作数据。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
所述处理模块，具体用于：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成动作数据。
在一种可能的实现中，所述风格编码的分布为均匀分布。
第四方面，本申请提供了一种数据处理装置，所述装置包括：
获取模块，用于获取人物对象的语音数据以及第一动作数据，所述第一动作数据为所述人物对象在发出所述语音数据时做出的身体动作的真实动作数据；
处理模块，用于根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
根据所述特征表示，通过动作生成网络，生成第二动作数据；所述第一动作数据和所述第二动作数据之间的差异用于更新所述特征提取网络以及所述动作生成网络。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，具体用于：
根据所述特征表示以及初始化的风格编码，通过动作生成网络，生成第二动作数据；
根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的第一风格编码；
通过预设的损失函数，更新所述初始化的风格编码，以得到更新后的风格编码；
所述第一风格编码和所述更新后的风格编码之间的差异用于更新所述编码器。
在一种可能的实现中，所述音频特征与音量和/或音高有关。
在一种可能的实现中，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
所述第一位置和所述第二位置为所述语音数据中的onset点；
所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
在一种可能的实现中，所述语音数据以及指示所述多个分割点位置的信息，包括：
根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
所述语音数据以及指示所述多个分割点位置的字符。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
所述处理模块，具体用于：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成第二动作数据。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
所述处理模块，具体用于：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成第二动作数据。
第五方面，本申请实施例提供了一种数据处理装置，可以包括存储器、处理器以及总线系统，其中，存储器用于存储程序，处理器用于执行存储器中的程序，以执行如上述第一方面及其任一可选的方法、或者如上述第二方面及其任一可选的方法。
第六方面，本申请实施例提供了一种计算机可读存储介质，所述计算机可读存储介质中存储有计算机程序，当其在计算机上运行时，使得计算机执行上述第一方面及其任一可选的方法、或者如上述第二方面及其任一可选的方法。
第七方面，本申请实施例提供了一种计算机程序，当其在计算机上运行时，使得计算机执行上述第一方面及其任一可选的方法、或者如上述第二方面及其任一可选的方法。
第八方面，本申请提供了一种芯片系统，该芯片系统包括处理器，用于支持执行设备或训练设备实现上述方面中所涉及的功能，例如，发送或处理上述方法中所涉及的数据；或，信息。在一种可能的设计中，所述芯片系统还包括存储器，所述存储器，用于保存执行设备或训练设备必要的程序指令和数据。该芯片系统，可以由芯片构成，也可以包括芯片和其他分立器件。
附图说明
图1A为人工智能主体框架的一种结构示意图；
图1B和图2为本发明的应用系统框架示意；
图3为终端的一种可选的硬件结构示意图；
图4为一种服务器的结构示意图；
图5为本申请的一种系统架构示意；
图6为一种云服务的流程；
图7为一种云服务的流程；
图8为本申请实施例提供的一种数据处理方法的流程示意；
图9为本申请实施例提供的数据处理方法的流程示意；
图10为本申请实施例提供的数据处理方法的流程示意；
图11A为本申请实施例提供的数据处理方法的流程示意；
图11B为本申请实施例提供的数据处理方法的流程示意；
图12A为本申请实施例提供的数据处理方法的流程示意；
图12B为本申请实施例提供的数据处理方法的流程示意；
图13为本申请实施例提供的数据处理装置的一种结构示意图；
图14为本申请实施例提供的数据处理装置的一种结构示意图；
图15为本申请实施例提供的执行设备的一种结构示意图；
图16为本申请实施例提供的训练设备一种结构示意图；
图17为本申请实施例提供的芯片的一种结构示意图。
具体实施方式
下面结合本发明实施例中的附图对本发明实施例进行描述。本发明的实施方式部分使用的术语仅用于对本发明的具体实施例进行解释，而非旨在限定本发明。
下面结合附图，对本申请的实施例进行描述。本领域普通技术人员可知，随着技术的发展和新场景的出现，本申请实施例提供的技术方案对于类似的技术问题，同样适用。
本申请的说明书和权利要求书及上述附图中的术语“第一”、“第二”等是用于区别类似的对象，而不必用于描述特定的顺序或先后次序。应该理解这样使用的术语在适当情况下可以互换，这仅仅是描述本申请的实施例中对相同属性的对象在描述时所采用的区分方式。此外，术语“包括”和“具有”以及他们的任何变形，意图在于覆盖不排他的包含，以便包含一系列单元的过程、方法、系统、产品或设备不必限于那些单元，而是可包括没有清楚地列出的或对于这些过程、方法、产品或设备固有的其它单元。
本文中所用用语“基本(substantially)”、“大约(about)”及类似用语用作近似用语、而并非用作程度用语，且旨在考虑到所属领域中的普通技术人员将知的测量值或计算值的固有偏差。此外，在阐述本发明实施例时使用“可(may)”是指“可能的一个或多个实施例”。本文中所用用语“使用(use)”、“正使用(using)”、及“被使用(used)”可被视为分别与用语“利用(utilize)”、“正利用(utilizing)”、及“被利用(utilized)”同义。另外，用语“示例性(exemplary)”旨在指代实例或例示。
首先对人工智能系统总体工作流程进行描述，请参见图1A，图1A示出的为人工智能主体框架的一种结构示意图，下面从“智能信息链”(水平轴)和“IT价值链”(垂直轴)两个维度对上述人工智能主题框架进行阐述。其中，“智能信息链”反映从数据的获取到处理的一列过程。举例来说，可以是智能信息感知、智能信息表示与形成、智能推理、智能决策、智能执行与输出的一般过程。在这个过程中，数据经历了“数据—信息—知识—智慧”的凝练过程。“IT价值链”从人智能的底层基础设施、信息(提供和处理技术实现)到系统的产业生态过程，反映人工智能为信息技术产业带来的价值。
(1)基础设施
基础设施为人工智能系统提供计算能力支持，实现与外部世界的沟通，并通过基础平台实现支撑。通过传感器与外部沟通；计算能力由智能芯片(CPU、NPU、GPU、ASIC、FPGA等硬件加速芯片)提供；基础平台包括分布式计算框架及网络等相关的平台保障和支持，可以包括云存储和计算、互联互通网络等。举例来说，传感器和外部沟通获取数据，这些数据提供给基础平台提供的分布式计算系统中的智能芯片进行计算。
(2)数据
基础设施的上一层的数据用于表示人工智能领域的数据来源。数据涉及到图形、图像、语音、文本，还涉及到传统设备的物联网数据，包括已有系统的业务数据以及力、位移、液位、温度、湿度等感知数据。
(3)数据处理
数据处理通常包括数据训练，机器学习，深度学习，搜索，推理，决策等方式。
其中，机器学习和深度学习可以对数据进行符号化和形式化的智能信息建模、抽取、预处理、训练等。
推理是指在计算机或智能系统中，模拟人类的智能推理方式，依据推理控制策略，利用形式化的信息进行机器思维和求解问题的过程，典型的功能是搜索与匹配。
决策是指智能信息经过推理后进行决策的过程，通常提供分类、排序、预测等功能。
(4)通用能力
对数据经过上面提到的数据处理后，进一步基于数据处理的结果可以形成一些通用的能力，比如可以是算法或者一个通用系统，例如，翻译，文本的分析，计算机视觉的处理，语音识别，图像的识别等等。
(5)智能产品及行业应用
智能产品及行业应用指人工智能系统在各领域的产品和应用，是对人工智能整体解决方案的封装，将智能信息决策产品化、实现落地应用，其应用领域主要包括：智能终端、智能交通、智能医疗、自动驾驶、智慧城市等。
首先介绍本申请的应用场景，本申请可以但不限于应用在具备基于文字或语音生成包含虚拟人(或者除虚拟人之外的其他对象)的视频的功能的应用程序(以下可以简称为虚拟人生成类应用程序)或者云侧服务器提供的云服务等，接下来分别进行介绍：
一、虚拟人生成类应用程序
本申请实施例的产品形态可以为虚拟人生成类应用程序。虚拟人生成类应用程序可以运行在终端设备或者云侧的服务器上。
在一种可能的实现中，虚拟人生成类应用程序可以实现基于文字或语音生成包含虚拟人(或者除虚拟人之外的其他对象)的视频的任务，其中，在智能应答的场景中，文字或语音可以为用户输入的请求、问询或者其他请求交互的语音或者文字，应用程序所生成的虚拟人的视频可以包含针对于用户输入的语音或者文字的应答语音以及相应发出应答语音时的虚拟人，该虚拟人会做出仿人类的动作，也就是能够做出人类在发出应答语音时相匹配的身体动作。示例性的，身体动作可以为手势。
其中，在动作生成类的场景中，文字或语音可以为用户输入的语音或者文字，应用程序所生成的虚拟人的视频可以包含相应发出用户输入的语音或者文字对应的语音时的虚拟人，该虚拟人会做出仿人类的动作，也就是能够做出人类在发出输入的语音或者文字对应的语音时相匹配的身体动作。
在一种可能的实现中，用户可以打开终端设备上安装的虚拟人生成类应用程序，并输入语音或者文字(可以是主动的输入、也可以是被动的采集，例如通过终端设备上的音频传感器采集得到)，虚拟人生成类应用程序可以通过本申请实施例提供的方法基于语音或者文字生成虚拟人、或者是虚拟人的动作数据，并将虚拟人、或者是虚拟人的动作数据呈现给用户(呈现方式可以但不限于是显示、保存、上传到云侧等)。
在一种可能的实现中，用户可以打开终端设备上安装的虚拟人生成类应用程序，并输入语音或文字(可以是主动的输入、也可以是被动的采集，例如通过终端设备上的音频传感器采集得到)，虚拟人生成类应用程序可以将语音或文字发送至云侧的服务器，云侧的服务器通过本申请实施例提供的方法基于语音或者文字生成虚拟人、或者是虚拟人的动作数据，并将虚拟人、或者是虚拟人的动作数据回传至终端设备，终端设备可以将虚拟人、或者是虚拟人的动作数据呈现给用户(呈现方式可以但不限于是显示、保存、上传到云侧等)。
在一种可能的实现中，虚拟人生成类应用程序实现的虚拟人生成可以具体用于使能增强现实(augmented reality，AR)、虚拟现实(virtual reality，VR)、混合现实(mixed reality，MR)远程会议、运动健康、元宇宙等应用场景中的虚拟人物驱动等。
接下来分别从功能架构以及实现功能的产品架构介绍本申请实施例中的虚拟人生成类应用程序。
参照图1B，图1B为本申请实施例中虚拟人生成类应用程序的功能架构示意：
在一种可能的实现中，如图1B所示，虚拟人生成类应用程序102可接收输入的参数101(例如包含 人体的语音或文字)且产生虚拟人的动作数据103(或者基于虚拟人的动作数据还原得到的虚拟人物的信息)。虚拟人生成类应用程序102可在(举例来说)至少一个计算机系统上执行，且包括计算机代码，所述计算机代码在由一或多个计算机执行时致使所述计算机执行用于执行本文中所描述的数据处理方法。
参照图2，图2为本申请实施例中运行虚拟人生成类应用程序的实体架构示意：
参见图2，图2示出了一种系统架构示意图。该系统可以包括终端100、以及服务器200。其中，服务器200可以包括一个或者多个服务器(图2中以包括一个服务器作为示例进行说明)，服务器200可以为一个或者多个终端提供虚拟人生成服务。
其中，终端100上可以安装有虚拟人生成类应用程序，或者打开与虚拟人生成相关的网页，上述应用程序和网页可以提供一个界面，终端100可以接收用户在虚拟人生成界面上输入的相关参数，并将上述参数发送至服务器200，服务器200可以基于接收到的参数，得到处理结果，并将处理结果返回至至终端100。
应理解，在一些可选的实现中，终端100也可以由自身完成基于接收到的参数，得到数据处理结果的动作，而不需要服务器配合实现，本申请实施例并不限定。
接下来描述图2中终端100的产品形态；
本申请实施例中的终端100可以为手机、平板电脑、可穿戴设备、车载设备、增强现实(augmented reality，AR)/虚拟现实(virtual reality，VR)设备、笔记本电脑、超级移动个人计算机(ultra-mobile personal computer，UMPC)、上网本、个人数字助理(personal digital assistant，PDA)等，本申请实施例对此不作任何限制。
图3示出了终端100的一种可选的硬件结构示意图。
参考图3所示，终端100可以包括射频单元110、存储器120、输入单元130、显示单元140、摄像头150(可选的)、音频电路160(可选的)、扬声器161(可选的)、麦克风162(可选的)、处理器170、外部接口180、电源190等部件。本领域技术人员可以理解，图3仅仅是终端或多功能设备的举例，并不构成对终端或多功能设备的限定，可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件。
输入单元130可用于接收输入的数字或字符信息，以及产生与该便携式多功能装置的用户设置以及功能控制有关的键信号输入。具体地，输入单元130可包括触摸屏131(可选的)和/或其他输入设备132。该触摸屏131可收集用户在其上或附近的触摸操作(比如用户使用手指、关节、触笔等任何适合的物体在触摸屏上或在触摸屏附近的操作)，并根据预先设定的程序驱动相应的连接装置。触摸屏可以检测用户对触摸屏的触摸动作，将该触摸动作转换为触摸信号发送给该处理器170，并能接收该处理器170发来的命令并加以执行；该触摸信号至少包括触点坐标信息。该触摸屏131可以提供该终端100和用户之间的输入界面和输出界面。此外，可以采用电阻式、电容式、红外线以及表面声波等多种类型实现触摸屏。除了触摸屏131，输入单元130还可以包括其他输入设备。具体地，其他输入设备132可以包括但不限于物理键盘、功能键(比如音量控制按键、开关按键133等)、轨迹球、鼠标、操作杆等中的一种或多种。
其中，其他输入设备132可以接收到输入的语音或文字等等。
该显示单元140可用于显示由用户输入的信息或提供给用户的信息、终端100的各种菜单、交互界面、文件显示和/或任意一种多媒体文件的播放。在本申请实施例中，显示单元140可用于显示虚拟人生成类应用程序的界面、基于语音或文字得到的虚拟人等。
该存储器120可用于存储指令和数据，存储器120可主要包括存储指令区和存储数据区，存储数据区可存储各种数据，如多媒体文件、文本等；存储指令区可存储操作系统、应用、至少一个功能所需的指令等软件单元，或者他们的子集、扩展集。还可以包括非易失性随机存储器；向处理器170提供包括管理计算处理设备中的硬件、软件以及数据资源，支持控制软件和应用。还用于多媒体文件的存储，以及运行程序和应用的存储。
处理器170是终端100的控制中心，利用各种接口和线路连接整个终端100的各个部分，通过运行或执行存储在存储器120内的指令以及调用存储在存储器120内的数据，执行终端100的各种功能和处理数据，从而对终端设备进行整体控制。可选的，处理器170可包括一个或多个处理单元；优选的，处理器170可集成应用处理器和调制解调处理器，其中，应用处理器主要处理操作系统、用户界面和应用程序等，调制解调处理器主要处理无线通信。可以理解的是，上述调制解调处理器也可以不集成到处理器170中。在一些实施例中，处理器、存储器、可以在单一芯片上实现，在一些实施例中，他们也可以在独立的芯片上分别实现。处理器170还可以用于产生相应的操作控制信号，发给计算处理设备相应的部件，读取以及处理软件中的数据，尤其是读取和处理存储器120中的数据和程序，以使其中的各个功能模块执行相应的功能，从而控制相应的部件按指令的要求进行动作。
其中，存储器120可以用于存储数据处理方法相关的软件代码，处理器170可以执行芯片的数据处理方法的步骤，也可以调度其他单元(例如上述输入单元130以及显示单元140)以实现相应的功能。
该射频单元110(可选的)可用于收发信息或通话过程中信号的接收和发送，例如，将基站的下行信息接收后，给处理器170处理；另外，将设计上行的数据发送给基站。通常，RF电路包括但不限于天线、至少一个放大器、收发信机、耦合器、低噪声放大器(Low Noise Amplifier，LNA)、双工器等。此外，射频单元110还可以通过无线通信与网络设备和其他设备通信。该无线通信可以使用任一通信标准或协议，包括但不限于全球移动通讯系统(Global System of Mobile communication，GSM)、通用分组无线服务(General Packet Radio Service，GPRS)、码分多址(Code Division Multiple Access，CDMA)、宽带码分多址(Wideband Code Division Multiple Access，WCDMA)、长期演进(Long Term Evolution，LTE)、电子邮件、短消息服务(Short Messaging Service，SMS)等。
其中，在本申请实施例中，该射频单元110可以将语音或文字发送至服务器200，并接收到服务器200发送的虚拟人的动作数据或者基于虚拟人的动作数据还原得到的虚拟人物的信息。
应理解，该射频单元110为可选的，其可以被替换为其他通信接口，例如可以是网口。
终端100还包括给各个部件供电的电源190(比如电池)，优选的，电源可以通过电源管理系统与处理器170逻辑相连，从而通过电源管理系统实现管理充电、放电、以及功耗管理等功能。
终端100还包括外部接口180，该外部接口可以是标准的Micro USB接口，也可以使多针连接器，可以用于连接终端100与其他装置进行通信，也可以用于连接充电器为终端100充电。
尽管未示出，终端100还可以包括闪光灯、无线保真(wireless fidelity，WiFi)模块、蓝牙模块、不同功能的传感器等，在此不再赘述。下文中描述的部分或全部方法均可以应用在如图3所示的终端100中。
接下来描述图2中服务器200的产品形态；
图4提供了一种服务器200的结构示意图，如图4所示，服务器200包括总线201、处理器202、通信接口203和存储器204。处理器202、存储器204和通信接口203之间通过总线201通信。
总线201可以是外设部件互连标准(peripheral component interconnect，PCI)总线或扩展工业标准结构(extended industry standard architecture，EISA)总线等。总线可以分为地址总线、数据总线、控制总线等。为便于表示，图4中仅用一条粗线表示，但并不表示仅有一根总线或一种类型的总线。
处理器202可以为中央处理器(central processing unit，CPU)、图形处理器(graphics processing unit，GPU)、微处理器(micro processor，MP)或者数字信号处理器(digital signal processor，DSP)等处理器中的任意一种或多种。
存储器204可以包括易失性存储器(volatile memory)，例如随机存取存储器(random access memory，RAM)。存储器204还可以包括非易失性存储器(non-volatile memory)，例如只读存储器(read-only memory，ROM)，快闪存储器，机械硬盘(hard drive drive，HDD)或固态硬盘(solid state drive，SSD)。
其中，存储器204可以用于存储数据处理方法相关的软件代码，处理器202可以执行芯片的数据处理方法的步骤，也可以调度其他单元以实现相应的功能。
应理解，上述终端100和服务器200可以为集中式或者是分布式的设备，上述终端100和服务器200中的处理器(例如处理器170以及处理器202)可以为硬件电路(如专用集成电路(application specific integrated circuit，ASIC)、现场可编程门阵列(field-programmable gate array，FPGA)、通用处理器、数字信号处理器(digital signal processing，DSP)、微处理器或微控制器等等)、或这些硬件电路的组合，例如，处理器可以为具有执行指令功能的硬件系统，如CPU、DSP等，或者为不具有执行指令功能的硬件系统，如ASIC、FPGA等，或者为上述不具有执行指令功能的硬件系统以及具有执行指令功能的硬件系统的组合。
应理解，本申请实施例中的和模型推理过程相关的步骤涉及AI相关的运算，在执行AI运算时，终端设备和服务器的指令执行架构不仅仅局限在上述介绍的处理器结合存储器的架构。下面结合图5对本申请实施例提供的系统架构进行详细的介绍。
图5为本申请实施例提供的系统架构示意图。如图5所示，系统架构500包括执行设备510、训练设备520、数据库530、客户设备540、数据存储系统550以及数据采集系统560。
执行设备510包括计算模块511、I/O接口512、预处理模块513和预处理模块514。计算模块511中可以包括目标模型/规则501，预处理模块513和预处理模块514是可选的。
其中，执行设备510可以为上述运行人物虚拟人生成类应用程序的终端设备或者服务器。
数据采集设备560用于采集训练样本。训练样本可以为语音或文字，以及对语音或文字中的人物的标注(例如人物的真实动作数据)等。在采集到训练样本之后，数据采集设备560将这些训练样本存入数据库530。
训练设备520可以基于数据库530中维护训练样本，对待训练的神经网络(例如本申请实施例中的特征提取网络、动作生成网络、编码器等)，以得到目标模型/规则501。
需要说明的是，在实际应用中，数据库530中维护的训练样本不一定都来自于数据采集设备560的采集，也有可能是从其他设备接收得到的。另外需要说明的是，训练设备520也不一定完全基于数据库530维护的训练样本进行目标模型/规则501的训练，也有可能从云端或其他地方获取训练样本进行模型训练，上述描述不应该作为对本申请实施例的限定。
根据训练设备520训练得到的目标模型/规则501可以应用于不同的系统或设备中，如应用于图5所示的执行设备510，该执行设备510可以是终端，如手机终端，平板电脑，笔记本电脑，增强现实(augmented reality，AR)/虚拟现实(virtual reality，VR)设备，车载终端等，还可以是服务器等。
具体的，训练设备520可以将训练后的模型传递至执行设备510。
在图5中，执行设备510配置输入/输出(input/output，I/O)接口512，用于与外部设备进行数据交互，用户可以通过客户设备540向I/O接口512输入数据(例如本申请实施例中语音或文字等)。
预处理模块513和预处理模块514用于根据I/O接口512接收到的输入数据进行预处理。应理解，可以没有预处理模块513和预处理模块514或者只有的一个预处理模块。当不存在预处理模块513和预处理模块514时，可以直接采用计算模块511对输入数据进行处理。
在执行设备510对输入数据进行预处理，或者在执行设备510的计算模块511执行计算等相关的处理过程中，执行设备510可以调用数据存储系统550中的数据、代码等以用于相应的处理，也可以将相应处理得到的数据、指令等存入数据存储系统550中。
最后，I/O接口512将处理结果(例如虚拟人的动作数据或者基于虚拟人的动作数据还原得到的虚拟人物的信息等)提供给客户设备540，从而提供给用户。
在图5所示情况下，用户可以手动给定输入数据，该“手动给定输入数据”可以通过I/O接口512提供的界面进行操作。另一种情况下，客户设备540可以自动地向I/O接口512发送输入数据，如果要求客户设备540自动发送输入数据需要获得用户的授权，则用户可以在客户设备540中设置相应权限。用户可以在客户设备540查看执行设备510输出的结果，具体的呈现形式可以是显示、声音、动作等具体方式。客户设备540也可以作为数据采集端，采集如图所示输入I/O接口512的输入数据及输出I/O接口512的输出结果作为新的样本数据，并存入数据库530。当然，也可以不经过客户设备540进行采集，而是由I/O接口512直接将如图所示输入I/O接口512的输入数据及输出I/O接口512的输出结果， 作为新的样本数据存入数据库530。
值得注意的是，图5仅是本申请实施例提供的一种系统架构的示意图，图中所示设备、器件、模块等之间的位置关系不构成任何限制，例如，在图5中，数据存储系统550相对执行设备510是外部存储器，在其它情况下，也可以将数据存储系统550置于执行设备510中。应理解，上述执行设备510可以部署于客户设备540中。
从模型的推理侧来说：
本申请实施例中，上述执行设备520的计算模块511可以获取到数据存储系统550中存储的代码来实现本申请实施例中的和模型推理过程相关的步骤。
本申请实施例中，执行设备520的计算模块511可以包括硬件电路(如专用集成电路(application specific integrated circuit，ASIC)、现场可编程门阵列(field-programmable gate array，FPGA)、通用处理器、数字信号处理器(digital signal processing，DSP)、微处理器或微控制器等等)、或这些硬件电路的组合，例如，训练设备520可以为具有执行指令功能的硬件系统，如CPU、DSP等，或者为不具有执行指令功能的硬件系统，如ASIC、FPGA等，或者为上述不具有执行指令功能的硬件系统以及具有执行指令功能的硬件系统的组合。
具体的，执行设备520的计算模块511可以为具有执行指令功能的硬件系统，本申请实施例提供的和模型推理过程相关的步骤可以为存储在存储器中的软件代码，执行设备520的计算模块511可以从存储器中获取到软件代码，并执行获取到的软件代码来实现本申请实施例提供的和模型推理过程相关的步骤。
应理解，执行设备520的计算模块511可以为不具有执行指令功能的硬件系统以及具有执行指令功能的硬件系统的组合，本申请实施例提供的和模型推理过程相关的步骤的部分步骤还可以通过执行设备520的计算模块511中不具有执行指令功能的硬件系统来实现，这里并不限定。
从模型的训练侧来说：
本申请实施例中，上述训练设备520可以获取到存储器(图5中未示出，可以集成于训练设备520或者与训练设备520分离部署)中存储的代码来实现本申请实施例中和模型训练相关的步骤。
本申请实施例中，训练设备520可以包括硬件电路(如专用集成电路(application specific integrated circuit，ASIC)、现场可编程门阵列(field-programmable gate array，FPGA)、通用处理器、数字信号处理器(digital signal processing，DSP)、微处理器或微控制器等等)、或这些硬件电路的组合，例如，训练设备520可以为具有执行指令功能的硬件系统，如CPU、DSP等，或者为不具有执行指令功能的硬件系统，如ASIC、FPGA等，或者为上述不具有执行指令功能的硬件系统以及具有执行指令功能的硬件系统的组合。
应理解，训练设备520可以为不具有执行指令功能的硬件系统以及具有执行指令功能的硬件系统的组合，本申请实施例提供的中和模型训练相关的部分步骤还可以通过训练设备520中不具有执行指令功能的硬件系统来实现，这里并不限定。
二、服务器提供的虚拟人生成类云服务：
在一种可能的实现中，服务器可以通过应用程序编程接口(application programming interface，API)为端侧提供虚拟人生成的服务。
其中，终端设备可以通过云端提供的API，将相关参数(例如语音或文字)发送至服务器，服务器可以基于接收到的参数，得到处理结果(例如虚拟人的动作数据或者基于虚拟人的动作数据还原得到的虚拟人物的信息等)，并将处理结果返回至至终端。
关于终端以及服务器的描述可以上述实施例的描述，这里不再赘述。
如图6示出了使用一项云平台提供的虚拟人生成类云服务的流程。
1.开通并购买内容审核服务。
2.用户可以下载内容审核服务对应的软件开发工具包(software development kit，SDK)，通常云平台提供多个开发版本的SDK，供用户根据开发环境的需求选择，例如JAVA版本的SDK、python版本的 SDK、PHP版本的SDK、Android版本的SDK等。
3.用户根据需求下载对应版本的SDK到本地后，将SDK工程导入至本地开发环境，在本地开发环境中进行配置和调试，本地开发环境还可以进行其他功能的开发，使得形成一个集合了虚拟人生成类能力的应用。
4.虚拟人生成类应用在被使用的过程中，当需要进行虚拟人生成时，可以触发虚拟人生成的API调用。当应用触发虚拟人生成功能时，发起API请求至云环境中的虚拟人生成类服务的运行实例，其中，API请求中携带语音或文字，由云环境中的运行实例对语音或文字进行处理，获得处理结果(例如虚拟人的动作数据或者基于虚拟人的动作数据还原得到的虚拟人物的信息等)。
5.云环境将处理结果返回至应用，由此完成一次的虚拟人生成服务调用。
三、服务器提供的模型训练类云服务：
在一种可能的实现中，服务器可以基于客户提供的语音或文字，来提供一个适配于该语音或文字中人物对象(或者更泛化的人物)的虚拟人生成的模型。
在一种可能的实现中，服务器可以通过应用程序编程接口(application programming interface，API)为端侧提供配图信息还原的服务。
其中，终端设备可以通过云端提供的API，将相关参数(例如语音或文字)发送至服务器，服务器可以基于接收到的参数，得到处理结果，并将处理结果(例如适配于该语音或文字中人物对象(或者更泛化的人物)的虚拟人生成的模型等)返回至终端。
如图7示出了使用一项云平台提供的模型训练类云服务的流程。
关于终端以及服务器的描述可以上述实施例的描述，这里不再赘述。
由于本申请实施例涉及大量神经网络的应用，为了便于理解，下面先对本申请实施例涉及的相关术语及神经网络等相关概念进行介绍。
(1)神经网络
神经网络可以是由神经单元组成的，神经单元可以是指以xs和截距1为输入的运算单元，该运算单元的输出可以为：
其中，s＝1、2、……n，n为大于1的自然数，Ws为xs的权重，b为神经单元的偏置。f为神经单元的激活函数(activation functions)，用于将非线性特性引入神经网络中，来将神经单元中的输入信号转换为输出信号。该激活函数的输出信号可以作为下一层卷积层的输入。激活函数可以是sigmoid函数。神经网络是将许多个上述单一的神经单元联结在一起形成的网络，即一个神经单元的输出可以是另一个神经单元的输入。每个神经单元的输入可以与前一层的局部接受域相连，来提取局部接受域的特征，局部接受域可以是由若干个神经单元组成的区域。
(2)深度神经网络
深度神经网络(Deep Neural Network，DNN)，可以理解为具有很多层隐含层的神经网络，这里的“很多”并没有特别的度量标准，我们常说的多层神经网络和深度神经网络其本质上是同一个东西。从DNN按不同层的位置划分，DNN内部的神经网络可以分为三类：输入层，隐含层，输出层。一般来说第一层是输入层，最后一层是输出层，中间的层数都是隐含层。层与层之间是全连接的，也就是说，第i层的任意一个神经元一定与第i+1层的任意一个神经元相连。虽然DNN看起来很复杂，但是就每一层的工作来说，其实并不复杂，简单来说就是如下线性关系表达式：其中，是输入向量，是输出向量，是偏移向量，W是权重矩阵(也称系数)，α()是激活函数。每一层仅仅是对输入向量经过如此简单的操作得到输出向量由于DNN层数多，则系数W和偏移向量的数量也就是很多了。那么，具体的参数在DNN是如何定义的呢？首先我们来看看系数W的定义。以一个三层的DNN为例，如：第二层的第4个神经元到第三层的第2个神经元的线性系数定义为上标3代表系数W所在的层数，而下标对应的是输出的第三层索引2和输入的第二层索引4。总结下，第L-1层的第k个神经元到第L层的第j个神经元的系数定义为注意，输入层是没有W参数的。在深度神经网络中，更多的隐含层让网络更能够刻画现实世界中的复杂情形。理论上而言，参数越多的模型复杂度越高，“容量” 也就越大，也就意味着它能完成更复杂的学习任务。
(3)卷积神经网络(Convosutionas Neuras Network，CNN)是一种带有卷积结构的深度神经网络。卷积神经网络包含了一个由卷积层和子采样层构成的特征抽取器。该特征抽取器可以看作是滤波器，卷积过程可以看作是使用一个可训练的滤波器与一个输入的图像或者卷积特征平面(feature map)做卷积。卷积层是指卷积神经网络中对输入信号进行卷积处理的神经元层。在卷积神经网络的卷积层中，一个神经元可以只与部分邻层神经元连接。一个卷积层中，通常包含若干个特征平面，每个特征平面可以由一些矩形排列的神经单元组成。同一特征平面的神经单元共享权重，这里共享的权重就是卷积核。共享权重可以理解为提取图像信息的方式与位置无关。这其中隐含的原理是：图像的某一部分的统计信息与其他部分是一样的。即意味着在某一部分学习的图像信息也能用在另一部分上。所以对于图像上的所有位置，我们都能使用同样的学习得到的图像信息。在同一卷积层中，可以使用多个卷积核来提取不同的图像信息，一般地，卷积核数量越多，卷积操作反映的图像信息越丰富。
卷积核可以以随机大小的矩阵的形式初始化，在卷积神经网络的训练过程中卷积核可以通过学习得到合理的权重。另外，共享权重带来的直接好处是减少卷积神经网络各层之间的连接，同时又降低了过拟合的风险。
(4)反向传播算法
卷积神经网络可以采用误差反向传播(back propagation，BP)算法在训练过程中修正初始的超分辨率模型中参数的大小，使得超分辨率模型的重建误差损失越来越小。具体地，前向传递输入信号直至输出会产生误差损失，通过反向传播误差损失信息来更新初始的超分辨率模型中参数，从而使误差损失收敛。反向传播算法是以误差损失为主导的反向传播运动，旨在得到最优的超分辨率模型的参数，例如权重矩阵。
(5)损失函数
在训练深度神经网络的过程中，因为希望深度神经网络的输出尽可能的接近真正想要预测的值，所以可以通过比较当前网络的预测值和真正想要的目标值，再根据两者之间的差异情况来更新每一层神经网络的权重向量(当然，在第一次更新之前通常会有初始化的过程，即为深度神经网络中的各层预先配置参数)，比如，如果网络的预测值高了，就调整权重向量让它预测低一些，不断地调整，直到深度神经网络能够预测出真正想要的目标值或与真正想要的目标值非常接近的值。因此，就需要预先定义“如何比较预测值和目标值之间的差异”，这便是损失函数(loss function)或目标函数(objective function)，它们是用于衡量预测值和目标值的差异的重要方程。其中，以损失函数举例，损失函数的输出值(loss)越高表示差异越大，那么深度神经网络的训练就变成了尽可能缩小这个loss的过程。
(6)反向传播算法
神经网络可以采用误差反向传播(back propagation，BP)算法在训练过程中修正初始的神经网络模型中参数的大小，使得神经网络模型的重建误差损失越来越小。具体地，前向传递输入信号直至输出会产生误差损失，通过反向传播误差损失信息来更新初始的神经网络模型中参数，从而使误差损失收敛。反向传播算法是以误差损失为主导的反向传播运动，旨在得到最优的神经网络模型的参数，例如权重矩阵。
在公共演讲中，肢体语言可以增加演讲的节奏感，让演讲更生动，更具有说服力。有研究表明，肢体语言在沟通交流中扮演重要角色。首先，肢体动作更加准确地表达意图和传递情感，与语音传达的信息互补；其次，肢体动作能够帮助用户更专注于与数字人沟通的内容；第三，能够提升数字人的说服力，可信度及真实感；最后，能够反应说话人的意图及个性。沟通中缺少肢体语言或者肢体动作僵化，会陷入恐怖谷效应。
人们期待虚拟数字人(或者可以简称为数字人)的行为更像人。虚拟数字人可以跟人一样，在演讲的时候，可以配合语音做出节奏性，强调性的动作。对于一些特殊的语音语义可以有特定的动作，例如，人们说“好/ok”的时候，倾向于比OK的手势，说到“第一/首先”的时候，倾向于微微停顿，比个1的手型。
因此，如何根据语音生成对应的手势等身体动作是一个亟待解决的问题。
为了解决上述问题，本申请实施例提供了一种数据处理方法。下面结合附图对本申请实施例的数据处理方法进行详细的介绍。
参照图8，图8为本申请实施例提供的一种数据处理方法的流程示意，如图8所示，本申请实施例提供的一种数据处理方法，可以包括步骤801至804，下面分别对这些步骤进行详细的描述。
801、获取获取语音数据。
在动作生成类应用中，可以获取到用户的语音数据，并基于该语音数据生成对应的动作数据。该动作数据可用于生成对应的虚拟人。
在动作生成类应用中，可以获取到用户的文本数据，并基于该文本数据生成对应的语音数据，进而可以基于该语音数据(可选的，还可以包括文本数据)生成对应的动作数据。该动作数据可用于生成对应的虚拟人。
在交互类应用中，可以获取到用户的语音数据，基于该语音数据中的文本确定答复文本，并根据答复文本生成对应的语音数据，进而可以基于该语音数据(可选的，还可以包括答复文本)生成对应的动作数据。该动作数据可用于生成对应的虚拟人。
在一种可能的实现中，该动作数据可以为手势动作等。
802、根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点。
在一种可能的实现中，可以根据语音数据来预测发出该语音数据的人物对象的动作数据。其中，用户在发声时(例如进行演讲时)，会做出一定的动作，且该动作存在一定的节奏点，节奏点为做出所述身体动作时由静止到运动或者由运动到静止的临界点。若能够在特征提取之前就显性的将节奏信息识别出来并作为模型的输入，可以使得后续生成的动作具有准确的节奏感，进而使得动作更贴合人物对象在真实发出语音时的动作。
接下来介绍如何识别出上述节奏信息。
在一种可能的实现中，可以根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，其中，所述音频特征与音量以及音高中的至少一种有关。
在一种可能的实现中，节奏点通常出现在语音的onset(和音量相关的音频特征)上，但却不是全部的onset都是用户动作的节奏点，语音的相邻节奏点之间的间隔是有一个范围的，该范围与讲者风格相关。假设该范围为[t1，t2]，通常小于t1的时间间隔是由于噪声、填充词或者不顺畅的发音引起的。超过t2的时间间隔通常是对应着停顿或者是不说话的安静阶段。基于这些观察，可以将相邻onset的时间间隔在预设范围内的onset确定为是节奏点，进而可以识别出用户在发出所述语音数据时做出的身体动作的节奏点，并基于该节奏点对语音数据进行切分。
以第一位置和第二位置为例，第一位置和第二位置为多个分割点位置中相邻的分割点位置。
在一种可能的实现中，所述第一位置和所述第二位置为通过如下方式确定的：所述第一位置和所述第二位置为所述语音数据中的onset点，且所述onset点之间的时间间隔在预设范围内。其中，预设范围可以为上述介绍的[t1，t2]。
在一种可能的实现中，可以使用一个启发性的策略来识别语音节奏。将语音进行切分，小于t1的不识别为节奏点，而针对于超过t2的可以插入伪节奏点。可以用以下的方式插入伪节奏点：距离上一个节奏点的时间超过t1，并且语音的音量超过一定的阈值Ia，可以将该阈值设置为环境噪声的平均音量，以使得相邻的分割点位置的时间间隔在预设范围内。如果区间内所有的音量都小于Ia，就均匀地插入最少的伪节奏点，以使得相邻的分割点位置的时间间隔在预设范围内。
在一种可能的实现中，所述多个分割点位置包括相邻的第一位置和第二位置；所述第一位置和所述第二位置可以为通过如下方式确定的：所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
803、根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示。
804、根据所述特征表示，通过动作生成网络，生成动作数据。
在一种可能的实现中，可以根据分割点位置将语音数据进行分割，得到多个语音片段，将多个语音片段输入到特征网络中。
在一种可能的实现中，可以根据分割点位置将语音数据对应的文本数据进行分割，得到多个文本片段，每个文本片段可以和多个语音片段之间一一对应。
在一种可能的实现中，可以根据上述识别地节奏点，对输入的语音进行分割，并将分割好的语音片段都归一化到t2的长度。将这段语音对应的动作特征及文本特征也重采样到t2的长度。
在一种可能的实现中，可以将所述语音数据以及指示所述多个分割点位置的字符输入到特征提取网络，得到特征表示，并根据所述特征表示，通过动作生成网络，生成动作数据。
在一种可能的实现中，除了将通过特征提取网络得到的特征表示输入到动作生成网络之外，还可以根据特征表示(例如特征表示中的高层次特征)通过神经网络来识别出动作的类型，并将动作类型和特征表示输入到动作生成网络。
根据上述基于节奏的分割方法，可以将数据集内的所有动作都切分成等长的动作片段。通过神经网络可以得到这些动作对应的类别，可以认为这些动作类别就是动作的高层次特征。
对数据集中的high-level的语音特征进行聚类，可以发现每个类别内对应的文本都很相似，例如“many”“quite a few”“lots of”“much”“more”等；将这些high-level语音特征对应的动作输入到训练好的动作编码器，将动作编码进行可视化，可以发现这些动作集中在某几个类别。
在具体实现中，可以使用vq-vae在动作的隐空间对动作进行聚类，也可以直接用kmeans对动作进行聚类。根据分割的结果，得到一系列动作片段{M1,M2,…,Mi,…Mn}，通过VQ-VAE的方式，可以得到一个手势词典库
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，可以根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示(也就是高层特征)，确定所述身体动作的动作类别；所述根据所述特征表示，通过动作生成网络，生成动作数据，包括：根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成动作数据。
由于动作类别本身携带语义，使得输入到动作生成网络的特征以及后续生成的身体动作具备语义相关性。
此外，对于任意一个动作片段Mi，通过VQ-VAE的编码器可以知道其对应的高层次特征手势词汇(动作类别)为Si，可以分解出该动作片段对应的风格码Zi，方法如图10所示，将上一个片段的动作Mi-1通过动作编码器得到其对应的动作特征将当前时段的语音的低层次特征与其上下两个时段的低层次语音特征及通过音频编码器得到当前时段的音频特征将Si与可学习的手势风格码Zi融合在一起输入到动作生成网络中生成最终的动作。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，可以根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；所述根据所述特征表示，通过动作生成网络，生成动作数据，具体包括：根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成动作数据。
在一种可能的实现中，所述风格编码的分布为均匀分布。
在一种可能的实现中，关于上述介绍的高层特征和底层特征的解耦示意可以参照图9所示：其中，给定音频及其对应的文本，使用语言预训练模型作为对比学习的监督信号，让语音的预训练模型的高层次特征与文本的高层次特征距离接近，让语音的预训练模型的低层次特征与文本的低层次特征距离接近。
在一种可能的实现中，动作数据可以为身体动作的关节点旋转角度或者动作的3D点云信息，这里并不限定。
示例性的，给定一段语音输入，本申请实施例提供的方法可以会将该语音片段分割成归一化的特征块然后按顺序生成对应的动作整体流程如图11A所示，其中，
其中，为前面学到的动作生成器，动作编码器及音频编码器。
在上述公式中，仅有手势词汇及手势风格码是不能直接从输入语音得到的，他们从动作类别预测模块和动作风格预测模块推理获得。
在一种可能的实现中，可以将文本数据也作为输入。此外，文本方案可以简化，在只有音频输入，没有其对应的文本、讲者ID的情况下，也可以推理出其对应的动作。流程图如图10所示，在该实施例中，输入只有音频。
接下来从训练侧对本申请实施例中的数据处理方法进行介绍。
参照图11B，图11B为本申请实施例提供的一种数据处理方法的流程示意，所述方法包括：
1101、获取人物对象的语音数据以及第一动作数据，所述第一动作数据为所述人物对象在发出所述语音数据时做出的身体动作的真实动作数据；
其中，第一动作数据可以为语音数据对应的动作数据的真值。
1102、根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
关于步骤1102的具体描述可以参照上述实施例中步骤802的介绍，这里不再赘述。
1103、根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
关于步骤1103的具体描述可以参照上述实施例中步骤803的介绍，这里不再赘述。
1104、根据所述特征表示，通过动作生成网络，生成第二动作数据；所述第一动作数据和所述第二动作数据之间的差异用于更新所述特征提取网络以及所述动作生成网络。
关于步骤1104的具体描述可以参照上述实施例中步骤804的介绍，这里不再赘述。
除了第一动作数据和所述第二动作数据之间的差异之外，还可以包括其他损失，具体可以为：
Lgen＝wrecLrec+wpercLperc+wlexemeLlexeme+wzLz    …(1)
其中重建损失：
是生成的动作片段和GT动作片段Mi之间的MSE误差。
感知误差损失：
其中前面训练的动作编码器。
假设手势动作的类别是由词汇决定，其他通道的信息只能影响动作的变化性。为了强化这个约束，使用当前的手势词汇及一些随机的其他特征生成一系列动作
其中P为位置编码，手语词损失定义为：
J是训练集中所有运动片段的随机子集，NJ是J的大小。可以根据数据集的大小来调整NJ。
最后使用KL离散损失来归一化风格码：
其中μz和分别为训练小批量中的手势风格码的均值和方差。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：根据所述特征表示以及初始化的风格编码，通过动作生成网络，生成第二动作数据；根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的第一风格编码；通过预设的损失函数，更新所述初始化的风格编码，以得到更新后的风格编码；
所述第一风格编码和所述更新后的风格编码之间的差异用于更新所述编码器。
在一种可能的实现中，所述音频特征与音量和/或音高有关。
在一种可能的实现中，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
所述第一位置和所述第二位置为所述语音数据中的onset点；
所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
在一种可能的实现中，所述语音数据以及指示所述多个分割点位置的信息，包括：
根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
所述语音数据以及指示所述多个分割点位置的字符。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成第二动作数据。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成第二动作数据。
参照图12A，图12A为使用文本数据来生成动作数据的一个流程示意，参照图12B，图12B为不使用文本数据来生成动作数据的一个流程示意。
参照图13，图13为本申请实施例提供的一种数据处理装置的结构示意，如图13所示，本申请实施例提供的一种数据处理装置1300，包括：
获取模块1301，用于获取语音数据；
其中，关于获取模块1301的具体描述可以参照上述实施例中步骤801的介绍，这里不再赘述。
处理模块1302，用于根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
根据所述特征表示，通过动作生成网络，生成动作数据。
其中，关于处理模块1302的具体描述可以参照上述实施例中步骤802至步骤804的介绍，这里不再赘述。
在一种可能的实现中，所述音频特征与音量和/或音高有关。
在一种可能的实现中，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
所述第一位置和所述第二位置为所述语音数据中的onset点；
所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
在一种可能的实现中，所述语音数据以及指示所述多个分割点位置的信息，包括：
根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
所述语音数据以及指示所述多个分割点位置的字符。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作 类别；
所述处理模块，具体用于：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成动作数据。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
所述处理模块，具体用于：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成动作数据。
在一种可能的实现中，所述风格编码的分布为均匀分布。
参照图14，图14为本申请实施例提供的一种数据处理装置的结构示意，如图14所示，本申请实施例提供的一种数据处理装置1400，包括：
获取模块1401，用于获取人物对象的语音数据以及第一动作数据，所述第一动作数据为所述人物对象在发出所述语音数据时做出的身体动作的真实动作数据；
其中，关于获取模块1401的具体描述可以参照上述实施例中步骤1101的介绍，这里不再赘述。
处理模块1402，用于根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
根据所述特征表示，通过动作生成网络，生成第二动作数据；所述第一动作数据和所述第二动作数据之间的差异用于更新所述特征提取网络以及所述动作生成网络。
其中，关于处理模块1402的具体描述可以参照上述实施例中步骤1102至步骤1104的介绍，这里不再赘述。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，具体用于：
根据所述特征表示以及初始化的风格编码，通过动作生成网络，生成第二动作数据；
根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的第一风格编码；
通过预设的损失函数，更新所述初始化的风格编码，以得到更新后的风格编码；
所述第一风格编码和所述更新后的风格编码之间的差异用于更新所述编码器。
在一种可能的实现中，所述音频特征与音量和/或音高有关。
在一种可能的实现中，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
所述第一位置和所述第二位置为所述语音数据中的onset点；
所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
在一种可能的实现中，所述语音数据以及指示所述多个分割点位置的信息，包括：
根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
所述语音数据以及指示所述多个分割点位置的字符。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
所述处理模块，具体用于：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成第二动作数据。
在一种可能的实现中，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
所述处理模块，具体用于：
根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成第二动作数据。
接下来介绍本申请实施例提供的一种执行设备，请参阅图15，图15为本申请实施例提供的执行设备的一种结构示意图，执行设备1500具体可以表现为虚拟现实VR设备、手机、平板、笔记本电脑、智能穿戴设备、监控数据处理设备或服务器等，此处不做限定。具体的，执行设备1500包括：接收器1501、发射器1502、处理器1503和存储器1504(其中执行设备1500中的处理器1503的数量可以一个或多个，图15中以一个处理器为例)，其中，处理器1503可以包括应用处理器15031和通信处理器15032。在本申请的一些实施例中，接收器1501、发射器1502、处理器1503和存储器1504可通过总线或其它方式连接。
存储器1504可以包括只读存储器和随机存取存储器，并向处理器1503提供指令和数据。存储器1504的一部分还可以包括非易失性随机存取存储器(non-volatile random access memory，NVRAM)。存储器1504存储有处理器和操作指令、可执行模块或者数据结构，或者它们的子集，或者它们的扩展集，其中，操作指令可包括各种操作指令，用于实现各种操作。
处理器1503控制执行设备的操作。具体的应用中，执行设备的各个组件通过总线系统耦合在一起，其中总线系统除包括数据总线之外，还可以包括电源总线、控制总线和状态信号总线等。但是为了清楚说明起见，在图中将各种总线都称为总线系统。
上述本申请实施例揭示的方法可以应用于处理器1503中，或者由处理器1503实现。处理器1503可以是一种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法的各步骤可以通过处理器1503中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器1503可以是通用处理器、数字信号处理器(digital signal processing，DSP)、微处理器或微控制器，还可进一步包括专用集成电路(application specific integrated circuit，ASIC)、现场可编程门阵列(field-programmable gate array，FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。该处理器1503可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器1504，处理器1503读取存储器1504中的信息，结合其硬件完成上述方法中涉及模型推理过程的步骤。
接收器1501可用于接收输入的数字或字符信息，以及产生与执行设备的相关设置以及功能控制有关的信号输入。发射器1502可用于通过第一接口输出数字或字符信息；发射器1502还可用于通过第一接口向磁盘组发送指令，以修改磁盘组中的数据；发射器1502还可以包括显示屏等显示设备。
本申请实施例还提供了一种训练设备，请参阅图16，图16是本申请实施例提供的训练设备一种结构示意图，具体的，训练设备1600由一个或多个服务器实现，训练设备1600可因配置或性能不同而产生比较大的差异，可以包括一个或一个以上中央处理器(central processing units，CPU)1616(例如，一个或一个以上处理器)和存储器1632，一个或一个以上存储应用程序1642或数据1644的存储介质1630(例如一个或一个以上海量存储设备)。其中，存储器1632和存储介质1630可以是短暂存储或持久存储。存储在存储介质1630的程序可以包括一个或一个以上模块(图示没标出)，每个模块可以包 括对训练设备中的一系列指令操作。更进一步地，中央处理器1616可以设置为与存储介质1630通信，在训练设备1600上执行存储介质1630中的一系列指令操作。
训练设备1600还可以包括一个或一个以上电源1626，一个或一个以上有线或无线网络接口1650，一个或一个以上输入输出接口1658；或，一个或一个以上操作系统1641，例如Windows ServerTM，Mac OS XTM，UnixTM,LinuxTM，FreeBSDTM等等。
本申请实施例中，中央处理器1616，用于执行上述实施例中和模型训练相关的动作。
本申请实施例中还提供一种包括计算机程序产品，当其在计算机上运行时，使得计算机执行如前述执行设备所执行的步骤，或者，使得计算机执行如前述训练设备所执行的步骤。
本申请实施例中还提供一种计算机可读存储介质，该计算机可读存储介质中存储有用于进行信号处理的程序，当其在计算机上运行时，使得计算机执行如前述执行设备所执行的步骤，或者，使得计算机执行如前述训练设备所执行的步骤。
本申请实施例提供的执行设备、训练设备或终端设备具体可以为芯片，芯片包括：处理单元和通信单元，所述处理单元例如可以是处理器，所述通信单元例如可以是输入/输出接口、管脚或电路等。该处理单元可执行存储单元存储的计算机执行指令，以使执行设备内的芯片执行上述实施例描述的数据处理方法，或者，以使训练设备内的芯片执行上述实施例描述的数据处理方法。可选地，所述存储单元为所述芯片内的存储单元，如寄存器、缓存等，所述存储单元还可以是所述无线接入设备端内的位于所述芯片外部的存储单元，如只读存储器(read-only memory，ROM)或可存储静态信息和指令的其他类型的静态存储设备，随机存取存储器(random access memory，RAM)等。
具体的，请参阅图17，图17为本申请实施例提供的芯片的一种结构示意图，所述芯片可以表现为神经网络处理器NPU 1700，NPU 1700作为协处理器挂载到主CPU(Host CPU)上，由Host CPU分配任务。NPU的核心部分为运算电路1703，通过控制器1704控制运算电路1703提取存储器中的矩阵数据并进行乘法运算。
在一些实现中，运算电路1703内部包括多个处理单元(Process Engine,PE)。在一些实现中，运算电路1703是二维脉动阵列。运算电路1703还可以是一维脉动阵列或者能够执行例如乘法和加法这样的数学运算的其它电子线路。在一些实现中，运算电路1703是通用的矩阵处理器。
举例来说，假设有输入矩阵A，权重矩阵B，输出矩阵C。运算电路从权重存储器1702中取矩阵B相应的数据，并缓存在运算电路中每一个PE上。运算电路从输入存储器1701中取矩阵A数据与矩阵B进行矩阵运算，得到的矩阵的部分结果或最终结果，保存在累加器(accumulator)1708中。
统一存储器1706用于存放输入数据以及输出数据。权重数据直接通过存储单元访问控制器(Direct Memory Access Controller，DMAC)1705，DMAC被搬运到权重存储器1702中。输入数据也通过DMAC被搬运到统一存储器1706中。
BIU为Bus Interface Unit即，总线接口单元1710，用于AXI总线与DMAC和取指存储器(Instruction Fetch Buffer，IFB)1709的交互。
总线接口单元1710(Bus Interface Unit，简称BIU)，用于取指存储器1709从外部存储器获取指令，还用于存储单元访问控制器1705从外部存储器获取输入矩阵A或者权重矩阵B的原数据。
DMAC主要用于将外部存储器DDR中的输入数据搬运到统一存储器1706或将权重数据搬运到权重存储器1702中或将输入数据数据搬运到输入存储器1701中。
向量计算单元1707包括多个运算处理单元，在需要的情况下，对运算电路1703的输出做进一步处理，如向量乘，向量加，指数运算，对数运算，大小比较等等。主要用于神经网络中非卷积/全连接层网络计算，如Batch Normalization(批归一化)，像素级求和，对特征平面进行上采样等。
在一些实现中，向量计算单元1707能将经处理的输出的向量存储到统一存储器1706。例如，向量计算单元1707可以将线性函数；或，非线性函数应用到运算电路1703的输出，例如对卷积层提取的特征平面进行线性插值，再例如累加值的向量，用以生成激活值。在一些实现中，向量计算单元1707生 成归一化的值、像素级求和的值，或二者均有。在一些实现中，处理过的输出的向量能够用作到运算电路1703的激活输入，例如用于在神经网络中的后续层中的使用。
控制器1704连接的取指存储器(instruction fetch buffer)1709，用于存储控制器1704使用的指令；
统一存储器1706，输入存储器1701，权重存储器1702以及取指存储器1709均为On-Chip存储器。外部存储器私有于该NPU硬件架构。
其中，上述任一处提到的处理器，可以是一个通用中央处理器，微处理器，ASIC，或一个或多个用于控制上述程序执行的集成电路。
另外需说明的是，以上所描述的装置实施例仅仅是示意性的，其中所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部模块来实现本实施例方案的目的。另外，本申请提供的装置实施例附图中，模块之间的连接关系表示它们之间具有通信连接，具体可以实现为一条或多条通信总线或信号线。
通过以上的实施方式的描述，所属领域的技术人员可以清楚地了解到本申请可借助软件加必需的通用硬件的方式来实现，当然也可以通过专用硬件包括专用集成电路、专用CPU、专用存储器、专用元器件等来实现。一般情况下，凡由计算机程序完成的功能都可以很容易地用相应的硬件来实现，而且，用来实现同一功能的具体硬件结构也可以是多种多样的，例如模拟电路、数字电路或专用电路等。但是，对本申请而言更多情况下软件程序实现是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以软件产品的形式体现出来，该计算机软件产品存储在可读取的存储介质中，如计算机的软盘、U盘、移动硬盘、ROM、RAM、磁碟或者光盘等，包括若干指令用以使得一台计算机设备(可以是个人计算机，训练设备，或者网络设备等)执行本申请各个实施例所述的方法。
在上述实施例中，可以全部或部分地通过软件、硬件、固件或者其任意组合来实现。当使用软件实现时，可以全部或部分地以计算机程序产品的形式实现。
所述计算机程序产品包括一个或多个计算机指令。在计算机上加载和执行所述计算机程序指令时，全部或部分地产生按照本申请实施例所述的流程或功能。所述计算机可以是通用计算机、专用计算机、计算机网络、或者其他可编程装置。所述计算机指令可以存储在计算机可读存储介质中，或者从一个计算机可读存储介质向另一计算机可读存储介质传输，例如，所述计算机指令可以从一个网站站点、计算机、训练设备或数据中心通过有线(例如同轴电缆、光纤、数字用户线(DSL))或无线(例如红外、无线、微波等)方式向另一个网站站点、计算机、训练设备或数据中心进行传输。所述计算机可读存储介质可以是计算机能够存储的任何可用介质或者是包含一个或多个可用介质集成的训练设备、数据中心等数据存储设备。所述可用介质可以是磁性介质，(例如，软盘、硬盘、磁带)、光介质(例如，DVD)、或者半导体介质(例如固态硬盘(Solid State Disk，SSD))等。

Claims (31)

  1. 一种数据处理方法，其特征在于，所述方法包括：
    获取语音数据；
    根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
    根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
    根据所述特征表示，通过动作生成网络，生成动作数据。
  2. 根据权利要求1所述的方法，其特征在于，所述音频特征与音量和/或音高有关。
  3. 根据权利要求1或2所述的方法，其特征在于，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
    所述第一位置和所述第二位置为所述语音数据中的onset点；
    所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
  4. 根据权利要求1至3任一所述的方法，其特征在于，所述语音数据以及指示所述多个分割点位置的信息，包括：
    根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
    所述语音数据以及指示所述多个分割点位置的字符。
  5. 根据权利要求1至4任一所述的方法，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：
    根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
    所述根据所述特征表示，通过动作生成网络，生成动作数据，包括：
    根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成动作数据。
  6. 根据权利要求1至5任一所述的方法，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：
    根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
    所述根据所述特征表示，通过动作生成网络，生成动作数据，包括：
    根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成动作数据。
  7. 根据权利要求6所述的方法，其特征在于，所述风格编码的分布为均匀分布。
  8. 一种数据处理方法，其特征在于，所述方法包括：
    获取人物对象的语音数据以及第一动作数据，所述第一动作数据为所述人物对象在发出所述语音数据时做出的身体动作的真实动作数据；
    根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
    根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
    根据所述特征表示，通过动作生成网络，生成第二动作数据；所述第一动作数据和所述第二动作数据之间的差异用于更新所述特征提取网络以及所述动作生成网络。
  9. 根据权利要求8所述的方法，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：
    根据所述特征表示以及初始化的风格编码，通过动作生成网络，生成第二动作数据；
    根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的第一风格编码；
    通过预设的损失函数，更新所述初始化的风格编码，以得到更新后的风格编码；
    所述第一风格编码和所述更新后的风格编码之间的差异用于更新所述编码器。
  10. 根据权利要求8或9所述的方法，其特征在于，所述音频特征与音量和/或音高有关。
  11. 根据权利要求8至10任一所述的方法，其特征在于，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
    所述第一位置和所述第二位置为所述语音数据中的onset点；
    所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
  12. 根据权利要求8至11任一所述的方法，其特征在于，所述语音数据以及指示所述多个分割点位置的信息，包括：
    根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
    所述语音数据以及指示所述多个分割点位置的字符。
  13. 根据权利要求8至12任一所述的方法，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：
    根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
    所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：
    根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成第二动作数据。
  14. 根据权利要求8至13任一所述的方法，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述方法还包括：
    根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
    所述根据所述特征表示，通过动作生成网络，生成第二动作数据，包括：
    根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成第二动作数据。
  15. 一种数据处理装置，其特征在于，所述装置包括：
    获取模块，用于获取语音数据；
    处理模块，用于根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于所述语音数据时做出的身体动作的预测节奏点；
    根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
    根据所述特征表示，通过动作生成网络，生成动作数据。
  16. 根据权利要求15所述的装置，其特征在于，所述音频特征与音量和/或音高有关。
  17. 根据权利要求15或16所述的装置，其特征在于，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
    所述第一位置和所述第二位置为所述语音数据中的onset点；
    所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
  18. 根据权利要求15至17任一所述的装置，其特征在于，所述语音数据以及指示所述多个分割点位置的信息，包括：
    根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
    所述语音数据以及指示所述多个分割点位置的字符。
  19. 根据权利要求15至18任一所述的装置，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
    根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
    所述处理模块，具体用于：
    根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成动作数据。
  20. 根据权利要求15至19任一所述的装置，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
    根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
    所述处理模块，具体用于：
    根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成动作数据。
  21. 根据权利要求20所述的装置，其特征在于，所述风格编码的分布为均匀分布。
  22. 一种数据处理装置，其特征在于，所述装置包括：
    获取模块，用于获取人物对象的语音数据以及第一动作数据，所述第一动作数据为所述人物对象在发出所述语音数据时做出的身体动作的真实动作数据；
    处理模块，用于根据所述语音数据的音频特征，从所述语音数据中确定多个分割点位置，所述分割点位置对应于人物对象在发出所述语音数据时做出的身体动作的预测节奏点；
    根据所述语音数据以及指示所述多个分割点位置的信息，通过特征提取网络，得到特征表示；
    根据所述特征表示，通过动作生成网络，生成第二动作数据；所述第一动作数据和所述第二动作数据之间的差异用于更新所述特征提取网络以及所述动作生成网络。
  23. 根据权利要求22所述的装置，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，具体用于：
    根据所述特征表示以及初始化的风格编码，通过动作生成网络，生成第二动作数据；
    根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的第一风格编码；
    通过预设的损失函数，更新所述初始化的风格编码，以得到更新后的风格编码；
    所述第一风格编码和所述更新后的风格编码之间的差异用于更新所述编码器。
  24. 根据权利要求22或23所述的装置，其特征在于，所述音频特征与音量和/或音高有关。
  25. 根据权利要求22至24任一所述的装置，其特征在于，所述多个分割点位置包括相邻的第一位置和第二位置，所述第二位置和所述第一位置的时间间隔在预设范围内；所述第一位置和所述第二位置为通过如下方式的一种确定的：
    所述第一位置和所述第二位置为所述语音数据中的onset点；
    所述第一位置为所述语音数据中的onset点，所述第二位置不为所述语音数据中的onset点，且所述第二位置为所述语音数据中音量大于阈值的时间点。
  26. 根据权利要求22至25任一所述的装置，其特征在于，所述语音数据以及指示所述多个分割点位置的信息，包括：
    根据所述多个分割点位置，将所述语音数据划分成的多个语音片段；或者，
    所述语音数据以及指示所述多个分割点位置的字符。
  27. 根据权利要求22至26任一所述的装置，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
    根据所述多个网络层中靠近输出层的一个或多个网络层输出的特征表示，确定所述身体动作的动作类别；
    所述处理模块，具体用于：
    根据所述多个网络层输出的特征表示中的部分或全部、以及所述动作类别，通过动作生成网络，生成第二动作数据。
  28. 根据权利要求22至27任一所述的装置，其特征在于，所述特征提取网络包括多个串连的网络层，所述特征表示包括所述多个网络层输出的特征表示，所述处理模块，还用于：
    根据所述多个网络层中远离输出层的一个或多个网络层输出的特征表示，通过编码器，得到所述身体动作的风格编码；
    所述处理模块，具体用于：
    根据所述多个网络层输出的特征表示中的部分或全部、以及所述风格编码，通过动作生成网络，生成第二动作数据。
  29. 一种计算机存储介质，其特征在于，所述计算机存储介质存储有一个或多个指令，所述指令在由一个或多个计算机执行时使得所述一个或多个计算机执行权利要求1至14中任一项所述方法的操作。
  30. 一种计算机程序产品，其特征在于，包括计算机可读指令，当所述计算机可读指令在计算机设备上运行时，使得所述计算机设备执行如权利要求1至14任一所述的方法。
  31. 一种系统，包括至少一个处理器，至少一个存储器；所述处理器、所述存储器通过通信总线连接并完成相互间的通信；
    所述至少一个存储器用于存储代码；
    所述至少一个处理器用于执行所述代码，以执行如权利要求1至14任一所述的方法。
PCT/CN2023/116552 2022-09-02 2023-09-01 一种数据处理方法及其装置 WO2024046473A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202211071943.4A CN115866291A (zh) 2022-09-02 2022-09-02 一种数据处理方法及其装置
CN202211071943.4 2022-09-02

Publications (1)

Publication Number Publication Date
WO2024046473A1 true WO2024046473A1 (zh) 2024-03-07

Family

ID=85660751

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2023/116552 WO2024046473A1 (zh) 2022-09-02 2023-09-01 一种数据处理方法及其装置

Country Status (2)

Country Link
CN (1) CN115866291A (zh)
WO (1) WO2024046473A1 (zh)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115866291A (zh) * 2022-09-02 2023-03-28 华为技术有限公司 一种数据处理方法及其装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200265836A1 (en) * 2019-02-15 2020-08-20 Wipro Limited System and method for controlling devices through voice interaction
CN113192161A (zh) * 2021-04-22 2021-07-30 清华珠三角研究院 一种虚拟人形象视频生成方法、系统、装置及存储介质
CN113763532A (zh) * 2021-04-19 2021-12-07 腾讯科技（深圳）有限公司 基于三维虚拟对象的人机交互方法、装置、设备及介质
CN113750523A (zh) * 2021-04-19 2021-12-07 腾讯科技（深圳）有限公司 三维虚拟对象的动作生成方法、装置、设备及存储介质
CN114911973A (zh) * 2022-05-09 2022-08-16 网易（杭州）网络有限公司 动作生成方法、装置、电子设备及存储介质
CN115866291A (zh) * 2022-09-02 2023-03-28 华为技术有限公司 一种数据处理方法及其装置

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20200265836A1 (en) * 2019-02-15 2020-08-20 Wipro Limited System and method for controlling devices through voice interaction
CN113763532A (zh) * 2021-04-19 2021-12-07 腾讯科技（深圳）有限公司 基于三维虚拟对象的人机交互方法、装置、设备及介质
CN113750523A (zh) * 2021-04-19 2021-12-07 腾讯科技（深圳）有限公司 三维虚拟对象的动作生成方法、装置、设备及存储介质
CN113192161A (zh) * 2021-04-22 2021-07-30 清华珠三角研究院 一种虚拟人形象视频生成方法、系统、装置及存储介质
CN114911973A (zh) * 2022-05-09 2022-08-16 网易（杭州）网络有限公司 动作生成方法、装置、电子设备及存储介质
CN115866291A (zh) * 2022-09-02 2023-03-28 华为技术有限公司 一种数据处理方法及其装置

Also Published As

Publication number Publication date
CN115866291A (zh) 2023-03-28

Similar Documents

Publication Publication Date Title
CN111951805B (zh) 一种文本数据处理方法及装置
CN111797893B (zh) 一种神经网络的训练方法、图像分类系统及相关设备
KR102530548B1 (ko) 신경망 프로세싱 유닛
WO2022179492A1 (zh) 一种卷积神经网络的剪枝处理方法、数据处理方法及设备
WO2024041479A1 (zh) 一种数据处理方法及其装置
WO2023221928A1 (zh) 一种推荐方法、训练方法以及装置
CN111950596A (zh) 一种用于神经网络的训练方法以及相关设备
WO2023284716A1 (zh) 一种神经网络搜索方法及相关设备
CN111191136A (zh) 一种信息推荐方法以及相关设备
WO2024083121A1 (zh) 一种数据处理方法及其装置
WO2023231954A1 (zh) 一种数据的去噪方法以及相关设备
WO2024213099A1 (zh) 一种数据处理方法及其装置
WO2024046473A1 (zh) 一种数据处理方法及其装置
WO2023231753A1 (zh) 一种神经网络的训练方法、数据的处理方法以及设备
WO2023051369A1 (zh) 一种神经网络的获取方法、数据处理方法以及相关设备
WO2024067373A1 (zh) 一种数据处理方法及相关装置
CN116861850A (zh) 一种数据处理方法及其装置
WO2024199409A1 (zh) 一种数据处理方法及其装置
CN112529149A (zh) 一种数据处理方法及相关装置
CN116737895A (zh) 一种数据处理方法及相关设备
WO2024017282A1 (zh) 一种数据处理方法及其装置
WO2024055952A1 (zh) 一种数据处理方法及其装置
WO2021218725A1 (zh) 一种图像数据处理方法及相关装置
WO2023197857A1 (zh) 一种模型切分方法及其相关设备
WO2023207823A1 (zh) 类别描述的特征信息的获取方法、图像的处理方法及设备

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 23859500

Country of ref document: EP

Kind code of ref document: A1