Pose prediction model training method and device, equipment and medium
By setting the number of neurons in the output layer and input layer, and splitting the sample set into multiple training samples, training the pose prediction model, the problem of low training efficiency in the existing technology is solved, and efficient training to meet different needs is achieved.
Patent Information
- Application Number
- CN202510264810.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-03-06
- Publication Date
- 2025-07-01
AI Technical Summary
Existing pose prediction models need to be retrained when facing different prediction needs, resulting in low training efficiency.
By obtaining the number of prediction steps, sample steps and sample set, set the number of neurons in the output layer and input layer, split the sample set into multiple training samples, and use these training samples to train the pose prediction model to obtain the preset pose prediction model.
A general pose prediction model training method for different prediction steps and sample steps requirements is provided, which improves training efficiency and adaptability.
Smart Images
Figure CN120233877A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the technical field of data processing, and more specifically, to a method, apparatus, device, and medium for training a pose prediction model. Background Art
[0002] Currently, in order to alleviate the problem of limited computing power of head-mounted display devices, remote rendering technology is usually adopted. Specifically, a cloud or a network edge server predicts the pose of a head-mounted display device through a pose prediction model, completes image rendering according to the predicted pose, and transmits the rendered image data back to the head-mounted display device for display by the head-mounted display device.
[0003] However, for a pose prediction model, in a single-step prediction scenario, the pose prediction model is required to perform single-step pose prediction, and in a multi-step prediction scenario, the pose prediction model is required to perform corresponding multi-step pose prediction. On this basis, in the face of different requirements, it is necessary to retrain the pose prediction model, which leads to the problem of low training efficiency of the pose prediction model. Therefore, a training method for a general pose prediction model for different requirements needs to be proposed urgently. Summary of the Invention
[0004] An object of the present application is to provide a new technical solution for training a pose prediction model.
[0005] According to a first aspect of the present application, there is provided a method for training a pose prediction model, including:
[0006] Obtaining a prediction step number, a sample step number, and a sample set, where the sample set includes a plurality of poses arranged in time sequence;
[0007] Setting the number of neurons in the output layer to the prediction step number and the number of neurons in the input layer to the sample step number to obtain a pose prediction model to be trained;
[0008] According to the sample step number and the prediction step number, splitting the sample set into a plurality of training samples, where one training sample includes consecutive sample step number poses as samples and consecutive prediction step number poses as labels;
[0009] Training the pose prediction model to be trained according to the plurality of training samples to obtain a preset pose prediction model.
[0010] Optionally, the splitting the training sample set into a plurality of training samples according to the sample step number and the prediction step number includes:
[0011] Obtaining a sample splitting rule;
[0012] Split the sample set into multiple training samples according to the sample step number, the predicted step number, and the sample splitting rule;
[0013] The sample splitting rule includes: the starting step number of the samples in the training sample is less than or equal to the difference between the total number of poses in the sample set and the cut-off step number of the labels in the training sample.
[0014] Optionally, the obtaining of the predicted step number, the sample step number, and the sample set includes:
[0015] Display a first set input interface;
[0016] Receive the predicted step number, the sample step number, and the sample set input through the first set input interface.
[0017] Optionally, setting the number of neurons in the output layer to the predicted step number and the number of neurons in the input layer to the sample step number to obtain a pose prediction model to be trained includes:
[0018] Set the number of neurons in the output layer to the predicted step number, the number of neurons in the input layer to the sample step number, the number of hidden layers to a preset number, and the number of neurons in any one of the hidden layers to a corresponding preset quantity to obtain a pose prediction model to be trained.
[0019] Optionally, before setting the number of neurons in the output layer to the predicted step number, the number of neurons in the input layer to the sample step number, setting the number of hidden layers to a preset number, and the number of neurons in each hidden layer to a preset quantity to obtain a pose prediction model to be trained, the method further includes:
[0020] Display a second set input interface;
[0021] Receive the preset number of layers and the corresponding preset quantity input through the second set input interface.
[0022] Optionally, after training the pose prediction model to be trained according to the multiple training samples to obtain a preset pose prediction model, the method further includes:
[0023] Obtain the number of sample steps of poses for pose prediction;
[0024] According to the number of sample steps of poses for pose prediction and the preset pose prediction model, obtain the number of predicted poses of the predicted step number.
[0025] According to a second aspect of the present application, there is provided a training device for a pose prediction model, including:
[0026] An obtaining module, configured to obtain a predicted step number, a sample step number, and a sample set, where the sample set includes a plurality of poses arranged in time sequence;
[0027] A setting module, configured to set the number of neurons in the output layer to the prediction steps and the number of neurons in the input layer to the sample steps, so as to obtain a pose prediction model to be trained;
[0028] A splitting module, configured to split the sample set into multiple training samples according to the sample steps and the prediction steps, where one training sample includes consecutive sample steps of poses as samples and consecutive prediction steps of poses as labels;
[0029] A training module, configured to train the pose prediction model to be trained according to the multiple training samples to obtain a preset pose prediction model.
[0030] Optionally, the splitting module is specifically configured to:
[0031] Obtain a sample splitting rule;
[0032] Split the sample set into multiple training samples according to the sample steps, the prediction steps and the sample splitting rule;
[0033] The sample splitting rule includes: the starting step of the sample in the training sample is less than or equal to the difference between the total number of poses in the sample set and the cut-off step of the label in the training sample.
[0034] According to a third aspect of the present application, there is provided an electronic device, where the electronic device includes the device according to any one of the second aspect;
[0035] Alternatively, the electronic device includes a memory and a processor, the memory is used to store computer instructions, and the processor is used to call the computer instructions from the memory to execute the method according to any one of the first aspect.
[0036] According to a fourth aspect of the present application, there is provided a computer-readable storage medium, on which a computer program is stored, and the computer program, when executed by a processor, implements the method according to any one of the first aspect.
[0037] The present application provides a method for training a pose prediction model, including: obtaining a prediction step number, a sample step number, and a sample set, where the sample set includes a plurality of poses arranged in time sequence; setting the number of neurons in the output layer to the prediction step number and the number of neurons in the input layer to the sample step number to obtain a pose prediction model to be trained; splitting the sample set into a plurality of training samples according to the sample step number and the prediction step number, where a training sample includes a continuous sample step number of poses as samples and a continuous prediction step number of poses as labels; training the pose prediction model to be trained according to the plurality of training samples to obtain a preset pose prediction model. This method provides a general method for training a pose prediction model for different prediction step number and sample step number requirements.
[0038] Other features and advantages of the present application will become clear through the following detailed description of the exemplary embodiments of the present application with reference to the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0039] The drawings incorporated in the specification and constituting a part of the specification illustrate embodiments of the present application and, together with the description, are used to explain the principles of the present application.
[0040] Figure 1 is a block diagram of the hardware configuration of an electronic device for implementing a method for training a pose prediction model according to an embodiment of the present application Figure 1 ;
[0041] Figure 2 is a schematic flowchart of a method for implementing a training method of a pose prediction model according to an embodiment of the present application;
[0042] Figure 3 is a schematic structural diagram of a model according to an embodiment of the present application;
[0043] Figure 4 is a schematic structural diagram of a device for implementing a training method of a pose prediction model according to an embodiment of the present application;
[0044] Figure 5 is a block diagram of the hardware configuration of an electronic device for implementing a method for training a pose prediction model according to an embodiment of the present application Figure 2 。 DETAILED DESCRIPTION OF THE EMBODIMENTS
[0045] Various exemplary embodiments of the present application will now be described in detail with reference to the accompanying drawings. It should be noted that: unless otherwise specifically stated, the relative arrangements of components and steps, numerical expressions, and numerical values set forth in these embodiments do not limit the scope of the present application.
[0046] The following description of at least one exemplary embodiment is merely illustrative in nature and is in no way intended to limit the present application, its application, or its use.
[0047] Techniques, methods, and devices known to those of ordinary skill in the relevant art may not be discussed in detail, but where appropriate, such techniques, methods, and devices should be regarded as part of the specification.
[0048] In all examples shown and discussed herein, any specific values should be construed as merely exemplary and not as a limitation. Thus, other examples of the exemplary embodiments may have different values.
[0049] It should be noted that like reference numerals and letters refer to like items in the following figures, and thus, once an item is defined in one figure, further discussion thereof is not required in subsequent figures.
[0050] Figure 1 is a block diagram of the hardware configuration of an electronic device for implementing a training method of a pose prediction model according to an embodiment of the present application Figure 1 .
[0051] The electronic device 1000 may be a terminal or a server. Further, the terminal may be a head-mounted device (such as an AR device, an MR device, and a VR device), a portable computer, a tablet computer, a handheld computer, etc. The server may be a cloud server, etc.
[0052] The electronic device 1000 may include a processor 1100, a memory 1200, an interface device 1300, a communication device 1400, a display device 1500, an input device 1600, a speaker 1700, a microphone 1800, and so on. Among them, the processor 1100 may be a central processing unit CPU, a microprocessor MCU, etc. The memory 1200 includes, for example, a ROM (read-only memory), a RAM (random access memory), a non-volatile memory such as a hard disk, etc. The interface device 1300 includes, for example, a USB interface, a headphone interface, etc. The communication device 1400 can perform wired or wireless communication, for example. The display device 1500 is, for example, a liquid crystal display screen, a touch display screen, etc. The input device 1600 may include, for example, a touch screen, a keyboard, etc. The user can input / output voice information through the speaker 1700 and the microphone 1800.
[0053] Although multiple devices are shown for the electronic device 1000 in Figure 1 , this application may only relate to some of the devices, for example, the electronic device 1000 only relates to the memory 1200 and the processor 1100.
[0054] Applied to the embodiments of the present application, the memory 1200 of the electronic device 1000 is used to store instructions for controlling the processor 1100 to execute the training method of the pose prediction model provided by the embodiments of the present application.
[0055] In the above description, a person skilled in the art can design instructions according to the solution disclosed in the present application. How the instructions control the processor to operate is well known in the art, so it will not be described in detail herein.
[0056] The present application provides a method for training a pose prediction model, which is applied to an electronic device as shown in Figure 1 shown, such as Figure 2 shown, and includes the following steps S2100 to step S2400.
[0057] Step S2100, obtain the prediction step number, the sample step number, and the sample set.
[0058] Among them, the sample set includes a plurality of poses arranged in time sequence.
[0059] In this embodiment, the prediction step number is the number of poses that the pose prediction model to be trained can predict at one time. The sample step number is the number of poses used as samples in a training sample when training the pose prediction model to be trained.
[0060] In an embodiment of the present application, for a single-step prediction scenario, the prediction step number can be set to 1, and for a multi-step prediction scenario, the prediction step number can be set to the corresponding number of steps for multiple steps.
[0061] In addition, the sample set includes a plurality of poses arranged in time sequence, and the poses in the sample set can be sequentially collected by a pose sensor. In one example, the sample set can be expressed as [x1, x2,..., x n , it can be understood that n is the total number of poses in the sample set.
[0062] In this embodiment, the prediction step number, the sample step number, and the sample set in the above step S2100 can be set in advance by the user according to needs. That is to say, the training method of the pose prediction model provided by the present application does not limit the specific values of the prediction step number, the sample step number, and the sample set.
[0063] In an embodiment of the present application, the above step S2100 is specifically implemented by the following steps S2110 and step S2111.
[0064] Step S2110, display a first set input interface.
[0065] Step S2111, receive the prediction step number, the sample step number, and the sample set input through the first set input interface.
[0066] In this embodiment, to enable a user to input a predicted number of steps, a sample number of steps, and a sample set that meet their own needs, an electronic device provides a first set input interface. The user can input the predicted number of steps, the sample number of steps, and the sample set that meet their own needs through this first set input interface.
[0067] In one example, the first set input interface may specifically include a first interface, a second interface, and a third interface. Among them, the first interface is used for the user to input the predicted number of steps that meet their own needs, the second interface is used for the user to input the sample number of steps that meet their own needs, and the third interface is used for the user to input the sample set that meet their own needs.
[0068] Step S2200: Set the number of neurons in the output layer to the predicted number of steps, and set the number of neurons in the input layer to the sample number of steps, to obtain a pose prediction model to be trained.
[0069] In this embodiment, the input layer is the first layer of the model and is used to input samples in the training samples during model training. The number of neurons in the input layer corresponds to the number of features included in the samples in a training sample. Based on this, in this embodiment, the number of neurons in the input layer is set to the sample number of steps. For example, when the sample number of steps is window_size - i + 1, the number of neurons in the input layer is set to window_size - i + 1.
[0070] The output layer is the last layer of the model and is used to output the pose predicted by the model during model training. The number of neurons in the output layer corresponds to the number of features included in the labels in a training sample. Based on this, in this embodiment, the number of neurons in the output layer is set to the predicted number of steps. For example, when the predicted number of steps is k + 1, the number of neurons in the output layer is k + 1.
[0071] Through the above step S2200, the personalized customization of the pose prediction model to be trained can be completed.
[0072] It can be understood that as Figure 3 shown, on the basis of including an input layer and an output layer, the model further includes a hidden layer. In this regard, in an embodiment of the present application, the above step S2200 is specifically implemented through the following step S2210.
[0073] Step S2210: Set the number of neurons in the output layer to the predicted number of steps, the number of neurons in the input layer to the sample number of steps, the number of layers of the hidden layer to a preset number of layers, and the number of neurons in any hidden layer to a corresponding preset quantity, to obtain a pose prediction model to be trained.
[0074] In one example, assuming the preset number of layers is 2, and the hidden layers are hidden layer 1 and hidden layer 2 respectively, the number of neurons in hidden layer 1 is the preset quantity a, and the number of neurons in hidden layer 2 is the preset quantity b, where a and b may or may not be the same.
[0075] Through the above step S2210, the complete customization of the pose prediction model to be trained can be completed.
[0076] It should be noted that Figure 3 In [the reference], an example is shown where the input layer, hidden layers, and output layer all include 5 neurons, and the hidden layers are hidden layer 1 and hidden layer 2 respectively.
[0077] In this embodiment, the preset number of layers and the preset quantity in the above step S2210 can be set in advance by the user according to their needs. That is to say, the training method of the pose prediction model provided in this application does not limit the specific values of the preset number of layers and the preset quantity.
[0078] In an embodiment of this application, the above step S2210 is specifically implemented through the following step S2210-1 and step S2210-2.
[0079] Step S2210-1, display the second set input interface.
[0080] Step S2210-2, receive the preset number of layers and the preset quantity input through the second set input interface.
[0081] In this embodiment, in order to allow the user to input the preset number of layers and the corresponding preset quantity that meet their own needs, the electronic device provides a second set input interface. The user can input the preset number of layers and the corresponding preset quantity that meet their own needs through this second set input interface.
[0082] In one example, the second set input interface includes a fourth interface, which is used for the user to input the preset number of layers that meet their own needs. The second set input interface also includes a fifth interface with the same number as the preset number of layers, and the fifth interface is used for the user to sequentially input the corresponding preset quantity.
[0083] Through the above step S2210-1 and step S2210-2, the complete personalized customization of the pose prediction model to be trained can be completed.
[0084] Step S2300, split the sample set into multiple training samples according to the sample step number and the prediction step number.
[0085] Among them, one training sample includes the consecutive sample step number of poses as the sample and the consecutive prediction step number of poses as the label.
[0086] In one example, taking the predicted number of steps as k + 1, the sample number of steps as window_size - i + 1, and the sample set represented as [x1, x2, …, x n as an example, based on the above step S2300, the training samples [datax i , datay i can be split and obtained.
[0087] Among them, datax i is the sample in the training sample, and datay i is the label corresponding to the sample datax i . datax i = [x i , x i+1 , … x window_size , datay i = [x window_size+j , x window_size+j+1 , … x window_size+j+k . i can start taking values from 1, and windows_size + j represents the starting step of prediction, where j is greater than or equal to 1.
[0088] Step S2400, train the pose prediction model to be trained according to multiple training samples to obtain a preset pose prediction model.
[0089] In this embodiment, input multiple training samples into the pose prediction model to be trained, and control the pose prediction model to be trained to run, then the preset pose prediction model can be obtained.
[0090] Through the above steps S2100 to S2400, the training of the pose prediction model to be trained for the predicted number of steps and the sample number of steps can be realized. This training process is independent of the specific values of the predicted number of steps and the sample number of steps. The predicted number of steps and the sample number of steps can be set by the user according to needs. Therefore, the above steps S2100 to S2400 provide a general training method for the pose prediction model for different predicted number of steps and sample number of steps requirements.
[0091] This application provides a training method for a pose prediction model, including: obtaining the predicted number of steps, the sample number of steps, and a sample set, where the sample set includes multiple poses arranged in time sequence; setting the number of neurons in the output layer as the predicted number of steps and the number of neurons in the input layer as the sample number of steps to obtain the pose prediction model to be trained; according to the sample number of steps and the predicted number of steps, split the sample set into multiple training samples, and a training sample includes consecutive sample number of steps poses as samples and consecutive predicted number of steps poses as labels; train the pose prediction model to be trained according to multiple training samples to obtain a preset pose prediction model. This method provides a general training method for the pose prediction model for different predicted number of steps and sample number of steps requirements.
[0092] In an embodiment of the present application, the above step S2300 is specifically implemented through the following steps S2310 and S2311.
[0093] Step S2310, obtain the sample splitting rule.
[0094] Among them, the sample splitting rule includes: the starting step number of the samples in the training samples is less than or equal to the difference between the total number of poses in the sample set and the cut-off step number of the labels in the training samples.
[0095] In an example, the training samples are [datax i , datay i , datax i = [x i , x i+1 , … x window_size , datay i = [x window_size+j , x window_size+j+1 , … x window_size+j+k , and the sample set is [x1, x2, …, x n as an example. Then the sample splitting rule is: 0 < i ≤ n - (window_size + j + k).
[0096] Through the sample splitting rule in the above step S2310, it can be ensured that the poses with continuous prediction step numbers as labels and the poses with continuous sample step numbers as samples are split from the sample set.
[0097] Step S2311, split the sample set into multiple training samples according to the sample step number, prediction step number and sample splitting rule.
[0098] In this embodiment, the specific implementation of the above step S2311 is: for a training sample, extract the poses with continuous sample step numbers from the sample as the samples in the training sample, and extract the poses with continuous prediction step numbers from the sample as the labels in the training sample, where the starting step number of the sample is less than or equal to the difference between the total number of poses in the sample set and the cut-off step number of the labels in the sample set.
[0099] In an embodiment of the present application, the training method of the pose prediction model provided by the present application further includes the following steps S2500 and S2600.
[0100] Step S2500, obtain the poses for pose prediction with a continuous number of sample steps.
[0101] Step S2600, obtain the predicted poses with a prediction step number according to the poses for pose prediction with a sample step number and a preset pose prediction model.
[0102] In this embodiment, based on the above step S2400, a trained preset pose prediction model can be obtained, and this preset pose prediction model can achieve accurate pose prediction. On this basis, if a continuous number of sample steps of poses for pose prediction are input into the preset pose prediction model, then the preset pose prediction model can output a predicted number of prediction poses corresponding to the number of prediction steps.
[0103] In addition, the following description is made for the above output layer.
[0104] The output layer is a fully connected layer, also known as a dense layer or a linear layer. Each neuron in the output layer is connected to all neurons in the previous hidden layer, and is used to synthesize the features extracted previously.
[0105] Suppose the input vector of a model is x and the output vector is y. The calculation process of the fully connected layer can be expressed as: y = Wx + b, where: W is the weight matrix, which contains the weights of all connections. b is the bias vector, which is usually used to adjust the linear transformation of the output. The input dimension of the output layer is equal to the number of neurons in the hidden layer, denoted as hidden_size.
[0106] For the single-step prediction scenario, the number of prediction steps is 1. At this time, the shape of the input X of the output layer is set to hidden_size×1, as shown in the following expression (1):
[0107]
[0108] Based on the above expression (1), the shape of the weight matrix W is output_size×hidden_size, that is, the following expression (2):
[0109]
[0110] Among them, output_size represents the number of neurons in the output layer.
[0111] The shape of the bias vector b is output_size×1, that is, the following expression (3):
[0112]
[0113] Based on the above expressions (1), (2) and (3), the output of the output layer is expressed as the following expression (4)
[0114]
[0115] Among them, for each output vector y in the output layer iExpressed as the following Expression Five:
[0116]
[0117] Where l represents the row number of the weight w, and m represents the column number of the weight w.
[0118] In an embodiment of the present application, the present application further provides a training device 400 for a pose prediction model, as Figure 4 shown, including:
[0119] An acquisition module 410, configured to acquire the prediction steps, the sample steps, and a sample set, where the sample set includes a plurality of poses arranged in time sequence;
[0120] A setting module 420, configured to set the number of neurons in the output layer to the prediction steps and the number of neurons in the input layer to the sample steps, to obtain a pose prediction model to be trained;
[0121] A splitting module 430, configured to split the sample set into a plurality of training samples according to the sample steps and the prediction steps, where one training sample includes consecutive sample steps of poses as samples and consecutive prediction steps of poses as labels;
[0122] A training module 440, configured to train the pose prediction model to be trained according to the plurality of training samples to obtain a preset pose prediction model.
[0123] In an embodiment of the present application, the splitting module 430 is specifically configured to:
[0124] Obtain a sample splitting rule;
[0125] Split the sample set into a plurality of training samples according to the sample steps, the prediction steps, and the sample splitting rule;
[0126] The sample splitting rule includes: the starting step of the sample in the training sample is less than or equal to the difference between the total number of poses in the sample set and the cut-off step of the label in the training sample.
[0127] In an embodiment of the present application, the acquisition module 410 is specifically configured to:
[0128] Display a first set input interface;
[0129] Receive the prediction steps, the sample steps, and the sample set input through the first set input interface.
[0130] In one embodiment of the present application, the setting module 420 is specifically configured to set the number of neurons in the output layer to the prediction steps, the number of neurons in the input layer to the sample steps, the number of hidden layers to a preset number of layers, and the number of neurons in any one of the hidden layers to a corresponding preset quantity, so as to obtain a pose prediction model to be trained.
[0131] In one embodiment of the present application, the setting module 420 is further configured to:
[0132] Display a second set input interface;
[0133] Receive the preset number of layers and the corresponding preset quantity input through the second set input interface.
[0134] In one embodiment of the present application, the training device 400 for the pose prediction model provided by the present application further includes:
[0135] A prediction module, configured to obtain the consecutive sample step numbers of poses for pose prediction;
[0136] According to the sample step numbers of poses for pose prediction and the preset pose prediction model, obtain the prediction step numbers of predicted poses.
[0137] The present application also provides an electronic device 500, and the electronic device 500 includes any one of the training devices 400 for the pose prediction model provided by the above device embodiments;
[0138] Or, as Figure 5 shown, the electronic device 500 includes a memory 510 and a processor 520. The memory 510 is used to store computer instructions, and the processor 520 is used to call the computer instructions from the memory 510 to execute any one of the training methods for the pose prediction model provided by the above method embodiments.
[0139] The present application also provides a computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, it implements any one of the training methods for the pose prediction model provided by the above method embodiments.
[0140] The present application may be a system, a method, and / or a computer program product. The computer program product may include a computer-readable storage medium having computer-readable program instructions thereon for causing a processor to implement various aspects of the present application.
[0141] A computer-readable storage medium can be a tangible device that can retain and store instructions for use by an instruction execution device. A computer-readable storage medium may be, for example—but not limited to—an electrical storage device, a magnetic storage device, an optical storage device, an electromagnetic storage device, a semiconductor storage device, or any suitable combination of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium include: a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), a static random access memory (SRAM), a portable compact disc read-only memory (CD-ROM), a digital versatile disc (DVD), a memory stick, a floppy disk, a mechanically encoded device such as a punched card or raised structures in grooves having instructions stored thereon, and any suitable combination of the foregoing. The computer-readable storage medium as used herein is not construed as being a transient signal per se, such as a radio wave or other freely propagating electromagnetic wave, an electromagnetic wave propagated through a waveguide or other transmission medium (e.g., an optical pulse through an optical fiber cable), or an electrical signal transmitted through a wire.
[0142] The computer-readable program instructions described herein can be downloaded from a computer-readable storage medium to respective computing / processing devices, or can be downloaded to an external computer or an external storage device through a network, such as the Internet, a local area network, a wide area network, and / or a wireless network. The network may include a copper transmission cable, an optical fiber transmission, a wireless transmission, a router, a firewall, a switch, a gateway computer, and / or an edge server. A network adapter card or network interface in each computing / processing device receives the computer-readable program instructions from the network and forwards the computer-readable program instructions for storage in a computer-readable storage medium in each computing / processing device.
[0143] The computer program instructions for performing the operations of this application may be assembly instructions, instruction set architecture (ISA) instructions, machine instructions, machine - related instructions, microcode, firmware instructions, state - setting data, or source code or object code written in any combination of one or more programming languages, including object - oriented programming languages such as Smalltalk, C++, etc., and conventional procedural programming languages such as the "C" language or similar programming languages. The computer - readable program instructions may be executed entirely on the user's computer, partially on the user's computer, executed as a stand - alone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or it may be connected to an external computer (e.g., through the Internet using an Internet service provider). In some embodiments, by using the state information of the computer - readable program instructions to customize an electronic circuit, such as a programmable logic circuit, a field - programmable gate array (FPGA), or a programmable logic array (PLA), the electronic circuit can execute the computer - readable program instructions to implement various aspects of this application.
[0144] Aspects of the present application are described herein with reference to the flowcharts and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the present application. It should be understood that each block of the flowcharts and / or block diagrams, and combinations of blocks in the flowcharts and / or block diagrams, can be implemented by computer - readable program instructions.
[0145] These computer - readable program instructions can be provided to a processor of a general - purpose computer, a special - purpose computer, or other programmable data - processing apparatus to produce a machine such that when these instructions are executed by the processor of the computer or other programmable data - processing apparatus, a device is produced that implements the functions / actions specified in one or more blocks of the flowcharts and / or block diagrams. These computer - readable program instructions can also be stored in a computer - readable storage medium, and these instructions cause a computer, a programmable data - processing apparatus, and / or other devices to work in a specific manner. Thus, the computer - readable medium storing the instructions includes a manufactured article that includes instructions for implementing various aspects of the functions / actions specified in one or more blocks of the flowcharts and / or block diagrams.
[0146] Computer-readable program instructions may also be loaded onto a computer, other programmable data processing apparatus, or other device, causing a series of operational steps to be performed on the computer, other programmable data processing apparatus, or other device to produce a computer-implemented process such that the instructions executed on the computer, other programmable data processing apparatus, or other device implement the functions / acts specified in one or more boxes of the flowchart and / or block diagram.
[0147] The flowcharts and block diagrams in the figures illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of the present application. In this regard, each box in the flowchart or block diagram may represent a module, a segment of a program, or a portion of an instruction, which contains one or more executable instructions for implementing the specified logical function. In some alternative implementations, the functions noted in the boxes may occur in a different order than noted in the figures. For example, two consecutive boxes may in fact be executed substantially in parallel, or they may sometimes be executed in the reverse order, depending on the functions involved. It should also be noted that each box in the block diagrams and / or flowcharts, and combinations of boxes in the block diagrams and / or flowcharts, may be implemented by a dedicated hardware-based system that performs the specified functions or acts, or by a combination of dedicated hardware and computer instructions. As is well known to those skilled in the art, implementation by hardware, implementation by software, and implementation by a combination of software and hardware are equivalent.
[0148] The embodiments of the present application have been described above. The above description is exemplary, not exhaustive, and is not limited to the disclosed embodiments. Many modifications and variations will be apparent to those of ordinary skill in the art without departing from the scope and spirit of the described embodiments. The selection of the terms used herein is intended to best explain the principles of the embodiments, the practical application, or the technical improvement of the technology in the market, or to enable other ordinary skilled persons in the art to understand the embodiments disclosed herein. The scope of the present application is defined by the appended claims.
Claims
1. A training method for a posture prediction model, characterized in that: include: Obtaining a predicted number of steps, a sample number of steps, and a sample set, wherein the sample set includes a plurality of postures arranged in time sequence; The number of neurons in the output layer is set to the number of prediction steps, and the number of neurons in the input layer is set to the number of sample steps, to obtain a posture prediction model to be trained; According to the sample step number and the predicted step number, the sample set is divided into a plurality of training samples, wherein one training sample includes continuous poses of the sample step number as samples and continuous poses of the predicted step number as labels; The posture prediction model to be trained is trained according to the multiple training samples to obtain a preset posture prediction model.
2. The method according to claim 1, characterized in that: The step of splitting the training sample set into a plurality of training samples according to the sample step number and the prediction step number comprises: Get sample splitting rules; Splitting the sample set into a plurality of training samples according to the sample step number, the prediction step number and the sample splitting rule; The sample splitting rule includes: the starting step number of the sample in the training sample is less than or equal to the difference between the total number of poses in the sample set and the cutoff step number of the label in the training sample.
3. The method according to claim 1, characterized in that The obtaining of the prediction steps, sample steps and sample set includes: Display the first setting input interface; Receive the prediction step number, sample step number and sample set inputted from the first setting input interface.
4. The method according to claim 1, characterized in that The step of setting the number of neurons in the output layer to the number of prediction steps and the number of neurons in the input layer to the number of sample steps, and obtaining a posture prediction model to be trained, comprises: The number of neurons in the output layer is set to the prediction number of steps, the number of neurons in the input layer is set to the sample number of steps, the number of layers in the hidden layer is set to the preset number of layers, and the number of neurons in any hidden layer is set to the corresponding preset number to obtain the posture prediction model to be trained.
5. The method according to claim 4, characterized in that Before obtaining the posture prediction model to be trained, the method further includes: Display the second setting input interface; The preset number of layers and the corresponding preset quantity inputted from the second setting input interface are received.
6. The method according to claim 1, characterized in that After training the posture prediction model to be trained according to the multiple training samples to obtain a preset posture prediction model, the method further includes: Obtaining the continuous number of sample steps for posture prediction; According to the sample step number of postures for posture prediction and the preset posture prediction model, the predicted step number of predicted postures are obtained.
7. A training device for a posture prediction model, characterized in that: include: An acquisition module, used to acquire the predicted number of steps, the number of sample steps and a sample set, wherein the sample set includes a plurality of postures arranged in time sequence; A setting module, used to set the number of neurons in the output layer to the prediction step number and the number of neurons in the input layer to the sample step number, to obtain a posture prediction model to be trained; A splitting module, used for splitting the sample set into multiple training samples according to the sample step number and the predicted step number, wherein one training sample includes continuous poses of the sample step number as samples and continuous poses of the predicted step number as labels; The training module is used to train the posture prediction model to be trained according to the multiple training samples to obtain a preset posture prediction model.
8. The device according to claim 7, characterized in that The splitting module is specifically used for: Get sample splitting rules; Splitting the sample set into a plurality of training samples according to the sample step number, the prediction step number and the sample splitting rule; The sample splitting rule includes: the starting step number of the sample in the training sample is less than or equal to the difference between the total number of poses in the sample set and the cutoff step number of the label in the training sample.
9. An electronic device, characterized in that: The electronic device comprises the device according to claim 7 or 8; Alternatively, the electronic device includes a memory and a processor, the memory is used to store computer instructions, and the processor is used to call the computer instructions from the memory to execute the method as claimed in any one of claims 1 to 6.
10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the method according to any one of claims 1 to 6 is implemented.