Data mapping method and system in meta-universe game environment

By normal distribution processing and action inference model supplementing the action amplitude data in the metaverse game, the problem of excessive data volume and rigid interaction is solved, the data volume is reduced and the smoothness of virtual character movement is improved, and the user is enhanced Experience.

CN118987628AInactive Publication Date: 2024-11-22上海和数软件有限公司 +1
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411092328.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-09
Publication Date
2024-11-22
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing meta-universe game technology generates a large amount of data during user motion capture, resulting in high upload bandwidth usage, delayed user experience, and stiff interaction between virtual images and objects, affecting the smoothness of the experience.

Method used

By normal distribution of the action amplitude data, the active data is filtered out, the amount of uploaded data and bandwidth occupancy is reduced, and the active data is supplemented by the action inference model to achieve more reasonable and smooth movement of virtual characters.

Benefits of technology

It effectively reduces the amount of data upload and bandwidth usage, alleviates server pressure, improves the rationality and fluency of virtual character movements, and improves the user experience.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118987628A_ABST
    Figure CN118987628A_ABST
Patent Text Reader

Abstract

The invention belongs to the field of meta-universe games, and provides a data mapping method and system in a meta-universe game environment. Comprising the steps of collecting action amplitude data, calculating normal distribution data, calculating a normal distribution mean value and a normal distribution standard deviation, obtaining confidence data, matching main action data, uploading the main action data, predicting an action inference model and updating all-node data. According to the method, normal distribution processing is carried out on the action amplitude data, active action data are screened out, the uploaded data size and bandwidth occupation are reduced, and the server pressure is relieved; and the active action data is supplemented through the action inference model, so that the virtual character movement is more reasonable and smooth.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of metaverse games, and in particular to a data mapping method and system in a metaverse game environment. Background Art

[0002] The metaverse is a virtual digital world consisting of various virtual reality platforms, social networks, games, and virtual assets, in which users can interact, create, and experience various virtual content. Metaverse games are an application scenario in the metaverse, usually including elements such as social interaction, digital economy, and virtual experience. They are games in which people can experience a richer, more diverse, and immersive virtual world. Patent CN 114089829 A provides a virtual reality metaverse system that provides a basic platform for users to use the metaverse for communication and games.

[0003] However, existing technologies often generate a large amount of data when performing motion capture of users, occupying a large amount of parallel upload bandwidth, resulting in a high delay in the actual user experience. In addition, the platform's interaction between the user's virtual image and metaverse objects is still rigid, making the user experience not smooth. Summary of the invention

[0004] In order to overcome the shortcomings of the prior art, the purpose of the present invention is to provide a data mapping method and system in a metaverse game environment, which can screen out the main action data by performing normal distribution processing on the action amplitude data, reduce the amount of uploaded data and bandwidth occupancy, and alleviate the pressure on the server; supplement the main action data through the action inference model to make the movement of the virtual character more reasonable and smooth.

[0005] To achieve the above object, the present invention provides the following solutions:

[0006] A data mapping method in a metaverse game environment, comprising:

[0007] Collect motion amplitude data of key nodes;

[0008] The motion amplitude data is normalized by using the BOX-COX algorithm to obtain normal distribution data;

[0009] Calculating the normal distribution mean and normal distribution standard deviation of the normal distribution data;

[0010] Eliminate the data that is smaller than 2 standard deviations of the normal distribution mean and larger than 2 standard deviations of the normal distribution mean in the normal distribution data to obtain confidence data;

[0011] Matching the data in the motion amplitude data corresponding to the confidence data to obtain main motion data;

[0012] Sending the main action data to a game cloud server;

[0013] Based on the game cloud server, the main action data is predicted using the trained action inference model to obtain full node data;

[0014] The full node data is updated to the virtual character in the metaverse game environment.

[0015] Preferably, the training process of the motion inference model includes:

[0016] Collect training action dataset;

[0017] Normalizing the training action data set by using the BOX-COX algorithm to obtain a normal training data set;

[0018] Calculating a training mean set and a training standard deviation set of the normal training data set;

[0019] Based on the training mean set and the training standard deviation set, obtaining a main action training set;

[0020] Determine the loss function; the loss function is m is the total number of action behaviors in the main action training set; n is the number of key nodes; are the frequency weights of the x-axis, y-axis, and z-axis directions corresponding to the i-th node; bi ,y bi 、z bi are the x-axis, y-axis, and z-axis coordinates of the ith node of the bth action in the training action dataset; are the predicted values ​​of the i-th node of the b-th action respectively;

[0021] Based on the loss function and the training action data set, the main action training set is trained using a convolutional neural network to obtain the action inference model.

[0022] Preferably, collecting the motion amplitude data of key nodes includes:

[0023] Create a 3D virtual image of the target player through a holographic camera;

[0024] Based on the 3D virtual image, collecting the hand joint motion data of the target player through the data gloves;

[0025] Based on the 3D virtual image, collecting head movement data of the target player through a head tracker;

[0026] Based on the 3D virtual image, collecting the body motion data of the target player through a somatosensory capture kit;

[0027] The hand joint motion data, the head motion data and the body motion data are integrated to obtain the motion amplitude data.

[0028] Preferably, the process of acquiring the frequency weight includes:

[0029] Based on the normal training data set, counting the number of occurrences of each of the key nodes;

[0030] Based on the training action data set, counting the sum of the action amplitudes of each of the key nodes;

[0031] The frequency weight is calculated based on the number of occurrences and the sum of the action amplitudes; the calculation formula of the frequency weight is: as well as c q is the number of occurrences of the qth key node; c i For the i The number of occurrences of the key nodes; Δd xq is the sum of the movement amplitudes of the qth key node in the x-axis direction; is the sum of the movement amplitudes of the i-th key node in the x-axis direction; Δd yq is the sum of the movement amplitudes of the qth key node in the y-axis direction; is the sum of the movement amplitudes of the i-th key node in the y-axis direction; Δd zq is the sum of the motion amplitudes of the qth key node in the z-axis direction; is the sum of the movement amplitudes of the i-th key node in the z-axis direction.

[0032] Preferably, sending the main action data to the game cloud server includes:

[0033] When the main motion data is matched, the data in the motion amplitude data that is not selected as the main motion data is set to 0 to obtain temporary motion data;

[0034] Matching the temporary action data according to the sequence of the key nodes to obtain sequential action data;

[0035] The sequential action data is packaged to obtain an action data packet, and the action data packet is sent to the game cloud server.

[0036] Preferably, the key nodes correspond to 78 joints of the human body.

[0037] Preferably, the acquisition frequency of the motion amplitude data is 24 times per second.

[0038] Preferably, the joints of the virtual character correspond to the key nodes one by one.

[0039] Preferably, a data mapping system in a metaverse game environment comprises:

[0040] A data acquisition module, used for acquiring the motion amplitude data of the key nodes;

[0041] A normal conversion module, used for normalizing the motion amplitude data using a BOX-COX algorithm to obtain the normal distribution data;

[0042] A parameter calculation module, used for calculating the normal distribution mean and the normal distribution standard deviation of the normal distribution data;

[0043] A data elimination module is used to eliminate data in the normal distribution data that is less than 2 normal distribution standard deviations of the normal distribution mean and data that is greater than 2 normal distribution standard deviations of the normal distribution mean, to obtain the confidence data;

[0044] A data matching module, used for matching the data in the motion amplitude data corresponding to the confidence data to obtain the main motion data;

[0045] A data generation module, used for sending the main action data to the game cloud server;

[0046] A data prediction module, used to predict the main action data based on the game cloud server using the action inference model to obtain the full node data;

[0047] A data update module is used to update the full node data to the virtual character in the metaverse game environment.

[0048] The present invention discloses the following technical effects:

[0049] The present invention provides a data mapping method and system in a metaverse game environment. By performing normal distribution processing on motion amplitude data, the problem of excessive data collection is solved, and the function of reducing the amount of uploaded data and bandwidth occupancy is achieved. Through the motion inference model, the problem of incomplete motion data is solved, and the supplement of main motion data and the smooth movement of virtual characters are achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative labor.

[0051] Figure 1 A data mapping flow chart provided for an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0053] The purpose of the present invention is to provide a data mapping method and system in a metaverse game environment. By performing normal distribution processing on the motion amplitude data, the main motion data is screened out, the amount of uploaded data and bandwidth occupancy are reduced, and the pressure on the server is relieved; the main motion data is supplemented by a motion inference model to make the movement of virtual characters more reasonable and smooth.

[0054] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0055] Figure 1 The data mapping flow chart provided by the embodiment of the present invention is as follows: Figure 1 As shown, the present invention provides a data mapping method in a metaverse game environment, comprising:

[0056] Step 100, collecting motion amplitude data of key nodes;

[0057] Step 200, using the BOX-COX algorithm to normalize the motion amplitude data to obtain normally distributed data;

[0058] Step 300, calculating the normal distribution mean and normal distribution standard deviation of the normal distribution data; the calculation formula of the normal distribution mean is: The formula for calculating the standard deviation of a normal distribution is: is the mean of the normal distribution; a can be a data set of motion amplitude data in any direction of xyz; a i is the data corresponding to the i-th joint in data set a; s is the standard deviation of the normal distribution;

[0059] Step 400, eliminating data with a value less than 2 standard deviations of the normal distribution and data with a value greater than 2 standard deviations of the normal distribution mean from the normal distribution data, to obtain confidence data;

[0060] Step 500, matching the data in the motion amplitude data corresponding to the confidence data to obtain main motion data;

[0061] Step 600, sending the main action data to the game cloud server;

[0062] Step 700, based on the game cloud server, using the trained action inference model to predict the main action data to obtain full node data;

[0063] Step 800, updating the full node data to the virtual character in the metaverse game environment.

[0064] Specifically, the training process of the action inference model includes:

[0065] Collect training action dataset;

[0066] Normalizing the training action data set by using the BOX-COX algorithm to obtain a normal training data set;

[0067] Calculating a training mean set and a training standard deviation set of the normal training data set;

[0068] Based on the training mean set and the training standard deviation set, obtaining a main action training set;

[0069] Determine the loss function; the loss function is m is the total number of action behaviors in the main action training set; n is the number of key nodes; are the frequency weights of the x-axis, y-axis, and z-axis directions corresponding to the i-th node; bi ,y bi 、z bi are the x-axis, y-axis, and z-axis coordinates of the ith node of the bth action in the training action dataset; are the predicted values ​​of the i-th node of the b-th action respectively;

[0070] Based on the loss function and the training action data set, the main action training set is trained using a convolutional neural network to obtain the action inference model.

[0071] Optionally, the motion amplitude data of key nodes are collected, including:

[0072] Create a 3D virtual image of the target player through a holographic camera;

[0073] Based on the 3D virtual image, collecting the hand joint motion data of the target player through the data gloves;

[0074] Based on the 3D virtual image, collecting head movement data of the target player through a head tracker;

[0075] Based on the 3D virtual image, collecting the body motion data of the target player through a somatosensory capture kit;

[0076] The hand joint motion data, the head motion data and the body motion data are integrated to obtain the motion amplitude data.

[0077] Furthermore, the process of obtaining the frequency weight includes:

[0078] Based on the normal training data set, counting the number of occurrences of each of the key nodes;

[0079] Based on the training action data set, counting the sum of the action amplitudes of each of the key nodes;

[0080] The frequency weight is calculated based on the number of occurrences and the sum of the action amplitudes; the calculation formula of the frequency weight is: as well as c q is the number of occurrences of the qth key node; c i For the i The number of occurrences of the key nodes; Δd xq is the sum of the movement amplitudes of the qth key node in the x-axis direction; is the sum of the movement amplitudes of the i-th key node in the x-axis direction; Δd yq is the sum of the movement amplitudes of the qth key node in the y-axis direction; is the sum of the movement amplitudes of the i-th key node in the y-axis direction; Δd zq is the sum of the motion amplitudes of the qth key node in the z-axis direction; is the sum of the movement amplitudes of the i-th key node in the z-axis direction.

[0081] Specifically, sending the main action data to the game cloud server includes:

[0082] When the main motion data is matched, the data in the motion amplitude data that is not selected as the main motion data is set to 0 to obtain temporary motion data;

[0083] Matching the temporary action data according to the sequence of the key nodes to obtain sequential action data;

[0084] The sequential action data is packaged to obtain an action data packet, and the action data packet is sent to the game cloud server.

[0085] Specifically, the key nodes correspond to 78 joints of the human body.

[0086] Optionally, the motion amplitude data is collected at a frequency of 24 times per second.

[0087] Specifically, the joints of the virtual character correspond to the key nodes one by one.

[0088] Furthermore, a data mapping system in a metaverse game environment includes:

[0089] A data acquisition module, used for acquiring the motion amplitude data of the key nodes;

[0090] A normal conversion module, used for normalizing the motion amplitude data using a BOX-COX algorithm to obtain the normal distribution data;

[0091] A parameter calculation module, used for calculating the normal distribution mean and the normal distribution standard deviation of the normal distribution data;

[0092] A data elimination module is used to eliminate data in the normal distribution data that is less than 2 normal distribution standard deviations of the normal distribution mean and data that is greater than 2 normal distribution standard deviations of the normal distribution mean, to obtain the confidence data;

[0093] A data matching module, used for matching the data in the motion amplitude data corresponding to the confidence data to obtain the main motion data;

[0094] A data generation module, used for sending the main action data to the game cloud server;

[0095] A data prediction module, used to predict the main action data based on the game cloud server using the action inference model to obtain the full node data;

[0096] A data update module is used to update the full node data to the virtual character in the metaverse game environment.

[0097] Specifically, the motion amplitude data of this embodiment is collected from 78 main joints of the human body. During the collection process, the coordinate changes of the same joint before and after collection represent the motion process; wherein, all joints use the same set of coordinate systems. When performing normalized data screening, the motion data of the joints are analyzed according to the x-axis, y-axis, and z-axis, and the three sets of data do not affect each other.

[0098] Alternatively, the pseudo code of the BOX-COX algorithm is as follows:

[0099] from scipy import stats

[0100] #Suppose there is a set of data a; a can be any of the three sets of data xyz

[0101] a=[a1,a2,...,an]

[0102] #Perform Box-Cox transformation convert_res is the output result

[0103] convert_res,_=stats.boxcox(a)

[0104] print(convert_res)

[0105] Specifically, since this embodiment is in a metaverse game environment and the key interaction is the hands, the hand movement data is analyzed separately, and its data screening process can also be processed using normalization.

[0106] Optionally, the motion range data of this embodiment is recorded using Microsoft Excel, the file format is csv, and the storage format is: joint number + acquisition time + motion range (x-axis range + y-axis range + z-axis range). The data when the motion inference model is trained also uses the above format.

[0107] Furthermore, the convolutional neural network used in the training of this embodiment includes two convolutional layers and one fully connected layer. During training, the remaining joint motion data are supplemented by the main action training set, and feedback and iteration are performed by calculating the difference between the supplemented joint data and the actual joint data until the similarity between the predicted data and the actual data reaches 90%, then the iteration is stopped and the motion inference model is output.

[0108] The beneficial effects of the present invention are as follows:

[0109] The present invention processes the motion amplitude data into normal distribution, screens out the main motion data, reduces the uploaded data volume and bandwidth occupancy, and relieves the server pressure; and supplements the main motion data through the motion inference model, making the virtual character movement more reasonable and smooth.

[0110] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the system disclosed in the embodiment, since it corresponds to the method disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the method part.

[0111] The principles and implementation methods of the present invention are described in this article using specific examples. The description of the above embodiments is only used to help understand the method and core idea of ​​the present invention. At the same time, for those skilled in the art, according to the idea of ​​the present invention, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present invention.

Claims

1. A data mapping method in a metaverse game environment, characterized in that: include: Collect motion amplitude data of key nodes; The motion amplitude data is normalized by using the BOX-COX algorithm to obtain normal distribution data; Calculating the normal distribution mean and normal distribution standard deviation of the normal distribution data; Eliminate the data that is smaller than 2 standard deviations of the normal distribution mean and larger than 2 standard deviations of the normal distribution mean in the normal distribution data to obtain confidence data; Matching the data in the motion amplitude data corresponding to the confidence data to obtain main motion data; Sending the main action data to a game cloud server; Based on the game cloud server, the main action data is predicted using the trained action inference model to obtain full node data; The full node data is updated to the virtual character in the metaverse game environment.

2. A data mapping method in a metaverse game environment according to claim 1, characterized in that: The training process of the action inference model includes: Collect training action dataset; Normalizing the training action data set by using the BOX-COX algorithm to obtain a normal training data set; Calculating a training mean set and a training standard deviation set of the normal training data set; Based on the training mean set and the training standard deviation set, obtaining a main action training set; Determine the loss function; the loss function is m is the total number of action behaviors in the main action training set; n is the number of key nodes; p xi 、p yi 、p zi are the frequency weights of the x-axis, y-axis, and z-axis directions corresponding to the i-th node; bi ,y bi 、z bi are the x-axis, y-axis, and z-axis coordinates of the ith node of the bth action in the training action dataset; are the predicted values ​​of the i-th node of the b-th action respectively; Based on the loss function and the training action data set, the main action training set is trained using a convolutional neural network to obtain the action inference model.

3. The data mapping method in a metaverse game environment according to claim 1, characterized in that: Collect the motion amplitude data of key nodes, including: Create a 3D virtual image of the target player through a holographic camera; Based on the 3D virtual image, collecting the hand joint motion data of the target player through the data gloves; Based on the 3D virtual image, collecting head movement data of the target player through a head tracker; Based on the 3D virtual image, collecting the body motion data of the target player through a somatosensory capture kit; The hand joint motion data, the head motion data and the body motion data are integrated to obtain the motion amplitude data.

4. The data mapping method in a metaverse game environment according to claim 2, characterized in that: The process of obtaining the frequency weight includes: Based on the normal training data set, counting the number of occurrences of each of the key nodes; Based on the training action data set, counting the sum of the action amplitudes of each of the key nodes; The frequency weight is calculated based on the number of occurrences and the sum of the action amplitudes; the calculation formula of the frequency weight is: as well as c q is the number of occurrences of the qth key node; c i For the i The number of occurrences of the key nodes; Δd xq is the sum of the movement amplitudes of the qth key node in the x-axis direction; is the sum of the movement amplitudes of the i-th key node in the x-axis direction; Δd yq is the sum of the movement amplitudes of the qth key node in the y-axis direction; is the sum of the movement amplitudes of the i-th key node in the y-axis direction; Δd zq is the sum of the movement amplitudes of the qth key node in the z-axis direction; Δd zi is the sum of the movement amplitudes of the i-th key node in the z-axis direction.

5. The data mapping method in a metaverse game environment according to claim 1, characterized in that: Sending the main action data to the game cloud server includes: When the main motion data is matched, the data in the motion amplitude data that is not selected as the main motion data is set to 0 to obtain temporary motion data; Matching the temporary action data according to the sequence of the key nodes to obtain sequential action data; The sequential action data is packaged to obtain an action data packet, and the action data packet is sent to the game cloud server.

6. The data mapping method in a metaverse game environment according to claim 1, characterized in that: The key nodes correspond to 78 joints of the human body.

7. The data mapping method in a metaverse game environment according to claim 1, characterized in that: The acquisition frequency of the motion amplitude data is 24 times per second.

8. The data mapping method in a metaverse game environment according to claim 1, characterized in that: The joints of the virtual character correspond to the key nodes one by one.

9. A data mapping system in a metaverse game environment, characterized in that: include: A data acquisition module, used for acquiring the motion amplitude data of the key nodes; A normal conversion module, used for normalizing the motion amplitude data using a BOX-COX algorithm to obtain the normal distribution data; A parameter calculation module, used for calculating the normal distribution mean and the normal distribution standard deviation of the normal distribution data; A data elimination module is used to eliminate data in the normal distribution data that is less than 2 normal distribution standard deviations of the normal distribution mean and data that is greater than 2 normal distribution standard deviations of the normal distribution mean, to obtain the confidence data; A data matching module, used for matching the data in the motion amplitude data corresponding to the confidence data to obtain the main motion data; A data generation module, used for sending the main action data to the game cloud server; A data prediction module, used to predict the main action data based on the game cloud server using the action inference model to obtain the full node data; A data update module is used to update the full node data to the virtual character in the metaverse game environment.