Method and device for adjusting virtual image in virtual world
By extracting the cold start and interactive data preference characteristics of device users in the virtual world, and reconstructing and adjusting virtual images, the problem of virtual images being not personalized enough is solved, and more adaptable virtual images are achieved, which improves user experience.
Patent Information
- Application Number
- CN202211085399.9
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-09-06
- Publication Date
- 2024-11-05
- Estimated Expiration
- 2042-09-06
AI Technical Summary
In the virtual world, it is difficult for the existing technology to effectively adjust the user's virtual image to adapt to their preferences and interactive behavior, resulting in the virtual image not being personalized and adaptable enough.
By obtaining cold start data of device users who access the virtual world, selecting preference features and performing virtual image reconstruction, and then further adjusting the image based on the virtual interaction data to generate a more personalized and adaptive target virtual image.
It realizes the generation of virtual images that are more adaptable and personalized to the preferences of the device user in the virtual world, and improves the adaptability and user experience of the virtual images.
Smart Images

Figure CN116188731B_ABST
Abstract
Description
Technical Field
[0001] This document relates to the field of virtualization technology, and in particular to a method and device for adjusting a virtual image in a virtual world. Background Art
[0002] The virtual world provides a simulation of the real world, and can even provide scenes that are difficult to achieve in the real world. Therefore, the virtual world is increasingly used in various scenarios. In the virtual world scenario, users log in to the three-dimensional virtual world through a specific identity ID and carry out activities through virtual user roles in the virtual world. Usually, there are different user roles in the virtual world, and these user roles each carry out different activities. Summary of the invention
[0003] One or more embodiments of the present specification provide a method for adjusting a virtual image in a virtual world, comprising: obtaining cold start data of a device user of an access device in the virtual world, and extracting preference features from the cold start data to obtain a cold start preference feature. Reconstructing a virtual image based on the cold start preference feature to obtain an initial virtual image of the device user in the virtual world. Extracting preference features from virtual interaction data of the device user in the virtual world to obtain an interaction preference feature. Adjusting the initial virtual image based on the interaction preference feature to obtain a target virtual image and outputting it to the access device.
[0004] One or more embodiments of the present specification provide a virtual image adjustment device for a virtual world, including: a cold start preference feature extraction module, configured to obtain cold start data of a device user of an access device in the virtual world, and perform preference feature extraction on the cold start data to obtain a cold start preference feature. A virtual image reconstruction module, configured to perform virtual image reconstruction based on the cold start preference feature to obtain an initial virtual image of the device user in the virtual world. An interaction preference feature extraction module, configured to perform preference feature extraction on virtual interaction data of the device user in the virtual world to obtain an interaction preference feature. A virtual image adjustment module, configured to adjust the initial virtual image based on the interaction preference feature to obtain a target virtual image and output it to the access device.
[0005] One or more embodiments of the present specification provide a virtual image adjustment device for a virtual world, including: a processor; and a memory configured to store computer executable instructions, wherein when the computer executable instructions are executed, the processor: obtains cold start data of a device user of a device accessing the virtual world, and extracts preference features for the cold start data to obtain a cold start preference feature. Reconstructs a virtual image based on the cold start preference feature to obtain an initial virtual image of the device user in the virtual world. Extracts preference features for virtual interaction data of the device user in the virtual world to obtain an interaction preference feature. Adjusts the initial virtual image based on the interaction preference feature to obtain a target virtual image and outputs it to the access device.
[0006] One or more embodiments of the present specification provide a storage medium for storing computer executable instructions, which implement the following process when executed by a processor: obtain cold start data of a device user of a device accessing a virtual world, and extract preference features for the cold start data to obtain a cold start preference feature. Reconstruct a virtual image based on the cold start preference feature to obtain an initial virtual image of the device user in the virtual world. Extract preference features for virtual interaction data of the device user in the virtual world to obtain an interaction preference feature. Adjust the initial virtual image based on the interaction preference feature to obtain a target virtual image and output it to the access device. BRIEF DESCRIPTION OF THE DRAWINGS
[0007] In order to more clearly illustrate one or more embodiments of the present specification or the technical solutions in the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present specification. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative labor.
[0008] Figure 1 A processing flow chart of a method for adjusting a virtual image in a virtual world provided by one or more embodiments of this specification;
[0009] Figure 2 A processing flow chart of a method for adjusting a virtual image in a virtual world applied to an actual application scenario provided by one or more embodiments of this specification;
[0010] Figure 3 A schematic diagram of a virtual image adjustment device for a virtual world provided by one or more embodiments of this specification;
[0011] Figure 4A schematic diagram of the structure of a virtual image adjustment device for a virtual world provided in one or more embodiments of the present specification. DETAILED DESCRIPTION
[0012] In order to enable those skilled in the art to better understand the technical solutions in one or more embodiments of this specification, the technical solutions in one or more embodiments of this specification will be clearly and completely described below in conjunction with the drawings in one or more embodiments of this specification. Obviously, the described embodiments are only part of the embodiments of this specification, not all of them. Based on one or more embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this document.
[0013] This specification provides an embodiment of a method for adjusting a virtual image in a virtual world:
[0014] The virtual image adjustment method of the virtual world provided by the present application extracts preference features from cold start data of a device user of a device accessing the virtual world, and reconstructs the virtual image based on the obtained cold start preference features to obtain a corresponding initial virtual image, generates a virtual image that is more compatible with the preferences of the device user in the cold start phase, then extracts preference features from the virtual interaction data of the device user in the virtual world, and further adjusts the initial virtual image based on the obtained interaction preference features, thereby obtaining a virtual image that is more compatible with the preferences of the device user and more personalized.
[0015] Step S102, obtaining cold start data of a device user of an access device in the virtual world, and extracting preference features from the cold start data to obtain cold start preference features.
[0016] The virtual world described in this embodiment refers to a virtual simulation world based on decentralized collaboration and with an open economic system, such as a game virtual world in which users access devices to interact with each other in games, a meeting virtual world in which users access devices to hold online meetings, or a transaction virtual world in which users access devices to conduct online transactions; optionally, decentralized transactions are performed in the virtual world by generating non-homogeneous identifiers, and ownership of virtual assets is obtained through transactions.
[0017] Furthermore, it is also possible to establish an identity mapping between the virtual image and the device user in the virtual world, and to trade virtual assets in the virtual world by generating non-homogeneous identifiers based on the established identity mapping, and to possess ownership of virtual assets through transactions. Specifically, users in the real world (physical world) can access the virtual world through access devices to conduct decentralized transactions and other behaviors in the virtual world. The access device to the virtual world can be a VR (Virtual Reality) device connected to the virtual world, an AR (Augmented Reality) device, etc., such as a head-mounted VR device connected to the virtual world.
[0018] The cold start data includes the access data or behavior data of the user during the process of registering and cold starting the virtual world. In the specific implementation process, when the device user accesses the virtual world through the access device of the virtual world, the access data before the device user interacts with the virtual image in the virtual world through the access device is collected as the cold start data. Optionally, the cold start data can be multimodal data, that is, the cold start data includes modal data of at least one modality, such as cold start data composed of image data of image modality and voice data of voice modality.
[0019] In a specific implementation, in the process of extracting preference features for the cold start data, in order to improve the extraction efficiency and extraction accuracy of preference feature extraction, in an optional implementation provided by this embodiment, preference feature extraction is performed on the cold start data to obtain cold start preference features, including:
[0020] Inputting modality data of at least one modality included in the cold start data into a first feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality;
[0021] The preference feature of the at least one modality is input into a first classifier for preference classification processing, and the output feature of a specific layer of the first classifier is extracted as the cold start preference feature.
[0022] Among them, the modality includes image modality, voice modality and / or text modality, and correspondingly, the modality data includes image data of image modality, voice data of voice modality and / or text data of text modality; the preference feature extraction refers to extracting features related to the behavioral preferences of the device user from the corresponding data.
[0023] For example, when user A accesses the virtual world through an access device of the virtual world, access data is collected before user A interacts with the virtual image in the virtual world through the access device, specifically, cold start data of user A during the access registration process of the virtual world and the cold start process of the virtual world is collected; after collecting the cold start data, the image data of the image modality contained in the cold start data is input into a deep residual network (such as ResNet34) for preference feature extraction to obtain image preference features of the image modality; and the voice data of the voice modality contained in the cold start data is input into an RNN (ecurrent Neural Network) for preference feature extraction to obtain voice preference features of the voice modality.
[0024] In practical applications, the training of the first feature extraction network and the first classifier can be completed in advance, for example, the first feature extraction network and the first classifier are trained separately on a cloud server. In order to improve the training efficiency and also to make the preference feature extraction performed by the first feature extraction network and the preference classification performed by the first classifier more accurate, in an optional implementation manner provided in this embodiment, the first feature extraction network and the first classifier are trained in the following manner:
[0025] Inputting modal data of at least one modality contained in the cold start sample data into a to-be-trained network of the corresponding modality to extract preference features, thereby obtaining preference features of at least one modality;
[0026] Inputting the preference feature of at least one modality into the classifier to be trained for preference classification processing, and extracting the output feature of a specific layer of the classifier to be trained;
[0027] The loss function is used to calculate the training loss of the sample labels corresponding to the output features and the cold start sample data, and the parameters of the network to be trained and the classifier to be trained are adjusted based on the training loss.
[0028] The network to be trained and the classifier to be trained are trained using the above training method, and after the training is completed, the first feature extraction network and the first preference classification network are obtained.
[0029] It should be noted that in addition to the above-mentioned implementation method of extracting preference features based on the first feature extraction network and performing preference classification processing based on the first classifier, in the process of extracting preference features for the cold start data, the following optional implementation method can also be used to extract preference features: extracting preference features of the corresponding modality for the modal data of at least one modality contained in the cold start data to obtain the preference features of the at least one modality; and determining the cold start preference features based on the preference features of the at least one modality.
[0030] Step S104: reconstructing the virtual image based on the cold start preference feature to obtain an initial virtual image of the device user in the virtual world.
[0031] In this embodiment, the virtual image refers to a three-dimensional digital image (three-dimensional digital image) obtained by reconstructing the user image using image reconstruction technology from the user image for displaying the user image in the virtual world, or a digital twin image in the virtual world that has a digital twin relationship or identity mapping with the user obtained by reconstructing the user image using image reconstruction technology from the user image. For example, based on the user images of multiple perspectives of the user in the physical world, the user images of multiple perspectives are input into the Neural Radiance Fields (NeRF) model for image reconstruction to obtain a three-dimensional digital image, digital twin image, or identity mapping image in the virtual world.
[0032] The initial virtual image refers to the virtual image recommended and displayed to the device user when the device user first accesses the virtual image during the process of accessing the virtual world through the access device. Specifically, the initial virtual image can be reconstructed and generated based on the cold start preference characteristics of the device user; in addition, the initial virtual image can also be reconstructed and generated based on the cold start preference characteristics and a multi-view image stored in advance as a template.
[0033] In specific implementation, in order to improve the efficiency of reconstructing the initial virtual image by reconstructing the virtual image based on the cold start preference features, an optional implementation provided in this embodiment is to reconstruct the virtual image based on the cold start preference features, including: inputting the cold start preference features into an image reconstruction model to reconstruct the virtual image, and outputting the initial virtual image.
[0034] In practical applications, the model training of the image reconstruction model can be completed in advance, such as training the image reconstruction model on a cloud server. In order to improve the efficiency of model training and also to make the virtual image reconstruction performed by the image reconstruction model more accurate, this embodiment provides an optional implementation method for training the image reconstruction model as follows: input the preferred feature samples into the image reconstruction model to be trained to reconstruct the virtual image and output the virtual image; calculate the loss between the output virtual image and the pre-constructed virtual image, and adjust the parameters of the image reconstruction model to be trained based on the loss.
[0035] For example, the StyleGAN architecture and the 3D-CNN (3D convolutional neural network) architecture are used to construct an image reconstruction model to be trained, and the image reconstruction model to be trained is supervised by predetermined preferred feature samples and pre-generated virtual images until the image reconstruction model to be trained converges to obtain an image reconstruction model.
[0036] Step S106, extracting preference features from the virtual interaction data of the device user in the virtual world to obtain interaction preference features.
[0037] The virtual interaction data described in this embodiment includes user interaction data of the device user interacting with other users in the virtual world based on the initial virtual image, access data of the device user performing access processing in the virtual world, and / or image adjustment data of the device user performing adjustment processing on the initial virtual image in the virtual world.
[0038] In a specific implementation, in the process of extracting preference features from the virtual interaction data, in order to improve the extraction efficiency and extraction accuracy of preference features, in an optional implementation provided by this embodiment, preference features are extracted from the virtual interaction data of the device user in the virtual world to obtain interaction preference features, including:
[0039] Inputting the interaction data of at least one modality included in the virtual interaction data into a second feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality;
[0040] The preference features of the at least one modality are input into a second classifier for classification, and the output features of a specific layer of the second classifier are extracted as the interaction preference features.
[0041] Among them, the modality includes image modality, voice modality and / or text modality, and correspondingly, the modality data includes image data of image modality, voice data of voice modality and / or text data of text modality; the preference feature extraction refers to extracting features related to the behavioral preferences of the device user from the corresponding data.
[0042] For example, after user A logs in to the virtual world through the access device of the virtual world, in the process of interacting in the virtual world based on the initial virtual image, user interaction data of user A interacting with other users based on the initial virtual image, as well as access data of user A performing access processing in the virtual world are collected, and the collected user interaction data and access data are used as virtual interaction data; after collecting the virtual interaction data, the image data of the image modality contained in the virtual interaction data is input into a lightweight convolutional neural network (such as MobileNetV2) for preference feature extraction to obtain image preference features of the image modality; and, the voice data of the voice modality contained in the virtual interaction data is input into an RNN (ecurrent Neural Network) for preference feature extraction to obtain voice preference features of the voice modality.
[0043] In practical applications, the training of the second feature extraction network and the second classifier can be completed in advance, for example, the second feature extraction network and the second classifier are trained separately on a cloud server. In order to improve the training efficiency and make the preference feature extraction performed by the second feature extraction network and the preference classification processing performed by the second classifier more accurate, in an optional implementation manner provided by this embodiment, the second feature extraction network and the second classifier are trained in the following manner:
[0044] Inputting modal data of at least one modality contained in the interactive sample data into a to-be-trained network of the corresponding modality to extract preference features, thereby obtaining preference features of at least one modality;
[0045] Inputting the preference feature of at least one modality into the classifier to be trained for classification processing, and extracting the output of a specific layer of the classifier to be trained as a training output feature;
[0046] The loss function is used to calculate the training loss of the sample labels corresponding to the training output features and the interactive sample data, and the parameters of the network to be trained and the classifier to be trained are adjusted based on the training loss.
[0047] It should be noted that, in addition to the above-mentioned implementation method of performing preference feature extraction based on the second feature extraction network and performing preference classification processing based on the second classifier, performing preference feature extraction on the virtual interaction data of the device user in the virtual world to obtain interaction preference features can also be implemented by the following optional implementation method: performing preference feature extraction on the modal data of at least one modality contained in the virtual interaction data to obtain the preference features of the at least one modality; and determining the virtual interaction data based on the preference features of the at least one modality.
[0048] Step S108: adjusting the initial virtual image based on the interaction preference feature, obtaining a target virtual image and outputting it to the access device.
[0049] In specific implementation, in order to improve the accuracy and effectiveness of adjusting the initial virtual image based on the interaction preference characteristics, this embodiment provides an optional implementation method for adjusting the initial virtual image based on the interaction preference characteristics, including: inputting the initial virtual image and the interaction preference characteristics into an image adjustment model to adjust the initial virtual image, and outputting the target virtual image.
[0050] It should be noted that the image adjustment model has the processing capability to adjust the input initial virtual image according to the input interaction preference characteristics, and the processing capability can be obtained through model training. In practical applications, the model training of the image adjustment model can be completed in advance, such as training the image adjustment model on a cloud server. In order to improve the model training efficiency and also to make the adjustment processing of the virtual image performed by the image adjustment model more accurate, in an optional implementation manner provided in this embodiment, the image adjustment model is trained in the following manner:
[0051] Inputting the training sample into the image adjustment model to be trained to adjust the virtual image, and inputting the output virtual image into the image encoder for encoding processing to obtain the virtual image encoding; the training sample is composed of the virtual image and the preference feature;
[0052] The training loss of the virtual image encoding and the training samples is calculated, and the parameters of the image adjustment model to be trained are adjusted according to the training loss.
[0053] For example, a lightweight 3D-CNN (3D convolutional neural network) and / or encoder architecture is used to construct an image adjustment model to be trained, and supervised training is performed on the image adjustment model to be trained according to predetermined training samples and predetermined virtual images (as sample labels of training samples) until the image adjustment model to be trained converges to obtain an image adjustment model.
[0054] Optionally, the target virtual image has an identity mapping with the device user, and the device user generates a non-homogeneous identifier in the virtual world based on the identity mapping to trade virtual assets and obtain ownership of the virtual assets through the transaction.
[0055] The following takes the application of a virtual image adjustment method of a virtual world provided by this embodiment in an actual application scenario as an example. Figure 2 , the virtual image adjustment method of the virtual world provided by this embodiment is further described, see Figure 2, a virtual image adjustment method for a virtual world applied in actual application scenarios specifically includes the following steps.
[0056] Step S202, obtaining cold start data of a device user of a device accessing the virtual world.
[0057] Step S204: input the modal data of at least one modality included in the cold start data into a first feature extraction network of the corresponding modality to extract preference features, so as to obtain preference features of at least one modality.
[0058] Step S206: input the preference features of at least one modality into the first classifier for preference classification processing, and extract the output features of a specific layer of the first classifier as cold start preference features.
[0059] Step S208: Input the cold start preference features into the image reconstruction model to reconstruct the virtual image, obtain the initial virtual image of the device user in the virtual world, and output it to the access device.
[0060] Step S210: Input the interaction data of at least one modality contained in the virtual interaction data of the device user in the virtual world into the second feature extraction network of the corresponding modality to extract the preference feature, so as to obtain the preference feature of at least one modality.
[0061] Step S212: input the preference features of at least one modality into a second classifier for classification processing, and extract the output features of a specific layer of the second classifier as interaction preference features.
[0062] Step S214, inputting the initial virtual image and the interaction preference features into the image adjustment model to adjust the initial virtual image, outputting the target virtual image and outputting it to the access device.
[0063] An embodiment of a virtual image adjustment device for a virtual world provided in this specification is as follows:
[0064] In the above-mentioned embodiment, a method for adjusting a virtual image in a virtual world is provided, and correspondingly, a device for adjusting a virtual image in a virtual world is also provided, which will be described below with reference to the accompanying drawings.
[0065] Reference Figure 3 , which shows a schematic diagram of a virtual image adjustment device in a virtual world provided by this embodiment.
[0066] Since the device embodiment corresponds to the method embodiment, the description is relatively simple, and the relevant parts can refer to the corresponding description of the method embodiment provided above. The device embodiment described below is only illustrative.
[0067] This embodiment provides a virtual image adjustment device for a virtual world, comprising:
[0068] The cold start preference feature extraction module 302 is configured to obtain cold start data of a device user of an access device in the virtual world, and extract preference features from the cold start data to obtain cold start preference features;
[0069] A virtual image reconstruction module 304 is configured to perform virtual image reconstruction based on the cold start preference feature to obtain an initial virtual image of the device user in the virtual world;
[0070] An interaction preference feature extraction module 306 is configured to extract preference features from virtual interaction data of the device user in the virtual world to obtain interaction preference features;
[0071] The virtual image adjustment module 308 is configured to adjust the initial virtual image based on the interaction preference characteristics, obtain a target virtual image, and output it to the access device.
[0072] An embodiment of a virtual image adjustment device for a virtual world provided in this specification is as follows:
[0073] Corresponding to the virtual image adjustment method of a virtual world described above, based on the same technical concept, one or more embodiments of this specification further provide a virtual image adjustment device for a virtual world, wherein the virtual image adjustment device for a virtual world is used to execute the virtual image adjustment method of a virtual world provided above. Figure 4 A schematic diagram of the structure of a virtual image adjustment device for a virtual world provided in one or more embodiments of the present specification.
[0074] This embodiment provides a virtual image adjustment device for a virtual world, including:
[0075] like Figure 4As shown, the virtual image adjustment device of the virtual world may have relatively large differences due to different configurations or performances, and may include one or more processors 401 and memory 402, and the memory 402 may store one or more storage applications or data. Among them, the memory 402 may be a temporary storage or a permanent storage. The application stored in the memory 402 may include one or more modules (not shown in the figure), and each module may include a series of computer executable instructions in the virtual image adjustment device of the virtual world. Furthermore, the processor 401 may be configured to communicate with the memory 402, and execute a series of computer executable instructions in the memory 402 on the virtual image adjustment device of the virtual world. The virtual image adjustment device of the virtual world may also include one or more power supplies 403, one or more wired or wireless network interfaces 404, one or more input / output interfaces 405, one or more keyboards 406, etc.
[0076] In a specific embodiment, a virtual image adjustment device for a virtual world includes a memory and one or more programs, wherein the one or more programs are stored in the memory, and the one or more programs may include one or more modules, and each module may include a series of computer executable instructions for adjusting the virtual image in the virtual world, and the one or more programs are configured to be executed by one or more processors, including computer executable instructions for performing the following:
[0077] Acquire cold start data of a device user of an access device in the virtual world, and extract preference features from the cold start data to obtain cold start preference features;
[0078] Reconstructing the virtual image based on the cold start preference feature to obtain the initial virtual image of the device user in the virtual world;
[0079] Extracting preference features from virtual interaction data of the device user in the virtual world to obtain interaction preference features;
[0080] The initial virtual image is adjusted based on the interaction preference feature, a target virtual image is obtained and output to the access device.
[0081] An embodiment of a storage medium provided in this specification is as follows:
[0082] Corresponding to the above-described method for adjusting a virtual image in a virtual world, based on the same technical concept, one or more embodiments of this specification further provide a storage medium.
[0083] The storage medium provided in this embodiment is used to store computer executable instructions, and the computer executable instructions implement the following process when executed by a processor:
[0084] Acquire cold start data of a device user of an access device in the virtual world, and extract preference features from the cold start data to obtain cold start preference features;
[0085] Reconstructing the virtual image based on the cold start preference feature to obtain the initial virtual image of the device user in the virtual world;
[0086] Extracting preference features from virtual interaction data of the device user in the virtual world to obtain interaction preference features;
[0087] The initial virtual image is adjusted based on the interaction preference feature, a target virtual image is obtained and output to the access device.
[0088] It should be noted that the embodiment of the storage medium in this specification and the embodiment of the virtual image adjustment method in the virtual world in this specification are based on the same inventive concept, so the specific implementation of this embodiment can refer to the implementation of the aforementioned corresponding method, and the repeated parts will not be repeated.
[0089] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.
[0090] In the 1930s, improvements to a technology could be clearly distinguished as hardware improvements (for example, improvements to the circuit structure of diodes, transistors, switches, etc.) or software improvements (improvements to the method flow). However, with the development of technology, many improvements to the method flow today can be regarded as direct improvements to the hardware circuit structure. Designers almost always obtain the corresponding hardware circuit structure by programming the improved method flow into the hardware circuit. Therefore, it cannot be said that an improvement in a method flow cannot be implemented using a hardware entity module. For example, a programmable logic device (PLD) (such as a field programmable gate array (FPGA)) is such an integrated circuit whose logical function is determined by the user's programming of the device. Designers can "integrate" a digital system on a PLD by programming it themselves, without having to ask a chip manufacturer to design and produce a dedicated integrated circuit chip. Moreover, nowadays, instead of manually making integrated circuit chips, this kind of programming is mostly implemented by "logic compiler" software, which is similar to the software compiler used when developing and writing programs, and the original code before compilation must also be written in a specific programming language, which is called hardware description language (HDL). There is not only one HDL, but many kinds, such as ABEL (Advanced Boolean Expression Language), AHDL (Altera Hardware Description Language), Confluence, CUPL (Cornell University Programming Language), HDCal, JHDL (Java Hardware Description Language), Lava, Lola, MyHDL, PALASM, RHDL (Ruby Hardware Description Language), etc. The most commonly used ones are VHDL (Very-High-Speed Integrated Circuit Hardware Description Language) and Verilog. Those skilled in the art should also know that it is only necessary to program the method flow slightly in the above-mentioned hardware description languages and program it into the integrated circuit, and then it is easy to obtain the hardware circuit that implements the logic method flow.
[0091] The controller can be implemented in any appropriate manner, for example, the controller can take the form of a microprocessor or processor and a computer-readable medium storing a computer-readable program code (such as software or firmware) that can be executed by the (micro)processor, a logic gate, a switch, an application-specific integrated circuit (ASIC), a programmable logic controller, and an embedded microcontroller. Examples of controllers include, but are not limited to, the following microcontrollers: ARC625D, Atmel AT91SAM, Microchip PIC18F26K20, and Silicone Labs C8051F320. The memory controller can also be implemented as part of the control logic of the memory. Those skilled in the art also know that in addition to implementing the controller in a purely computer-readable program code manner, the controller can be implemented in the form of a logic gate, a switch, an application-specific integrated circuit, a programmable logic controller, and an embedded microcontroller by logically programming the method steps. Therefore, this controller can be considered as a hardware component, and the devices included therein for implementing various functions can also be regarded as structures within the hardware component. Or even, the devices for implementing various functions can be regarded as both software modules for implementing the method and structures within the hardware component.
[0092] The systems, devices, modules or units described in the above embodiments may be implemented by computer chips or entities, or by products with certain functions. A typical implementation device is a computer. Specifically, the computer may be, for example, a personal computer, a laptop computer, a cellular phone, a camera phone, a smart phone, a personal digital assistant, a media player, a navigation device, an email device, a game console, a tablet computer, a wearable device, or a combination of any of these devices.
[0093] For the convenience of description, the above devices are described in terms of functions and are divided into various units. Of course, when implementing the embodiments of this specification, the functions of each unit can be implemented in the same or multiple software and / or hardware.
[0094] It should be understood by those skilled in the art that one or more embodiments of this specification may be provided as a method, system or computer program product. Therefore, one or more embodiments of this specification may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this specification may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.
[0095] This specification is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of this specification. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0096] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0097] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.
[0098] In a typical configuration, a computing device includes one or more processors (CPU), input / output interfaces, network interfaces, and memory.
[0099] The memory may include non-permanent storage in a computer-readable medium, random access memory (RAM) and / or non-volatile memory in the form of read-only memory (ROM) or flash RAM. The memory is an example of a computer-readable medium.
[0100] Computer readable media include permanent and non-permanent, removable and non-removable media that can be implemented by any method or technology to store information. Information can be computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, compact disk read-only memory (CD-ROM), digital versatile disk (DVD) or other optical storage, magnetic cassettes, magnetic disk storage or other magnetic storage devices or any other non-transmission media that can be used to store information that can be accessed by a computing device. As defined herein, computer readable media does not include temporary computer readable media (transitory media), such as modulated data signals and carrier waves.
[0101] It should also be noted that the terms "include", "comprises" or any other variations thereof are intended to cover non-exclusive inclusion, so that a process, method, commodity or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, commodity or device. In the absence of more restrictions, the elements defined by the sentence "comprises a ..." do not exclude the existence of other identical elements in the process, method, commodity or device including the elements.
[0102] One or more embodiments of the present specification may be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. One or more embodiments of the present specification may also be practiced in distributed computing environments where tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules may be located in local and remote computer storage media, including storage devices.
[0103] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.
[0104] The above description is only an embodiment of this document and is not intended to limit this document. For those skilled in the art, this document may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this document should be included in the scope of the claims of this document.
Claims
1. A method for adjusting a virtual image in a virtual world, comprising: Acquire cold start data of a device user of an access device in the virtual world, and extract preference features for the cold start data to obtain cold start preference features, wherein the cold start data consists of image data in an image modality and voice data in a voice modality; Reconstructing the virtual image based on the cold start preference feature to obtain the initial virtual image of the device user in the virtual world; Extracting preference features from virtual interaction data of the device user in the virtual world to obtain interaction preference features, wherein the virtual interaction data includes user interaction data of the device user interacting with other users in the virtual world based on the initial virtual image; Adjusting the initial virtual image based on the interaction preference feature, obtaining a target virtual image and outputting it to the access device; The extracting preference features from the cold start data to obtain cold start preference features includes: Inputting modality data of at least one modality included in the cold start data into a first feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality; Inputting the preference feature of the at least one modality into a first classifier for preference classification processing, and extracting output features of a specific layer of the first classifier as the cold start preference feature; The extracting preference features of the virtual interaction data of the device user in the virtual world to obtain the interaction preference features includes: Inputting the interaction data of at least one modality included in the virtual interaction data into a second feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality; The preference features of the at least one modality are input into a second classifier for classification, and the output features of a specific layer of the second classifier are extracted as the interaction preference features.
2. The method for adjusting a virtual image in a virtual world according to claim 1, wherein the step of extracting preference features from the cold start data to obtain cold start preference features comprises: Extracting preference features of the corresponding modality from modality data of at least one modality included in the cold start data to obtain preference features of the at least one modality; The cold start preference profile is determined based on a preference profile of the at least one modality.
3. The method for adjusting a virtual image in a virtual world according to claim 1, wherein the first feature extraction network and the first classifier are trained in the following manner: Inputting modal data of at least one modality contained in the cold start sample data into a to-be-trained network of the corresponding modality to extract preference features, thereby obtaining preference features of at least one modality; Inputting the preference feature of at least one modality into the classifier to be trained for preference classification processing, and extracting the output feature of a specific layer of the classifier to be trained; The loss function is used to calculate the training loss of the sample labels corresponding to the output features and the cold start sample data, and the parameters of the network to be trained and the classifier to be trained are adjusted based on the training loss.
4. The method for adjusting a virtual image in a virtual world according to claim 1, wherein the step of reconstructing the virtual image based on the cold start preference feature comprises: The cold start preference feature is input into an image reconstruction model to reconstruct a virtual image, and the initial virtual image is output.
5. According to the virtual image adjustment method of the virtual world of claim 4, the image reconstruction model is trained in the following manner: Inputting the preferred feature samples into the image reconstruction model to be trained to reconstruct the virtual image, and outputting the virtual image; The loss between the output virtual image and the pre-built virtual image is calculated, and the parameters of the image reconstruction model to be trained are adjusted based on the loss.
6. The method for adjusting a virtual image in a virtual world according to claim 1, wherein the second feature extraction network and the second classifier are trained in the following manner: Inputting modal data of at least one modality contained in the interactive sample data into a to-be-trained network of the corresponding modality to extract preference features, thereby obtaining preference features of at least one modality; Inputting the preference feature of at least one modality into the classifier to be trained for classification processing, and extracting the output of a specific layer of the classifier to be trained as a training output feature; The loss function is used to calculate the training loss of the sample labels corresponding to the training output features and the interactive sample data, and the parameters of the network to be trained and the classifier to be trained are adjusted based on the training loss.
7. The method for adjusting a virtual image in a virtual world according to claim 1, wherein the step of extracting preference features from virtual interaction data of the device user in the virtual world to obtain interaction preference features comprises: Extracting preference features of the corresponding modality from modality data of at least one modality included in the virtual interaction data to obtain preference features of the at least one modality; The virtual interaction data is determined based on a preference characteristic of the at least one modality.
8. The method for adjusting a virtual image in a virtual world according to claim 1, wherein adjusting the initial virtual image based on the interaction preference feature comprises: The initial virtual image and the interaction preference feature are input into an image adjustment model to perform adjustment processing on the initial virtual image, and the target virtual image is output.
9. The virtual image adjustment method of the virtual world according to claim 8, wherein the image adjustment model is trained in the following manner: Inputting the training sample into the image adjustment model to be trained to adjust the virtual image, and inputting the output virtual image into the image encoder for encoding processing to obtain the virtual image encoding; the training sample is composed of the virtual image and the preference feature; The training loss of the virtual image encoding and the training samples is calculated, and the parameters of the image adjustment model to be trained are adjusted according to the training loss.
10. According to the virtual image adjustment method of the virtual world according to any one of claims 1 to 9, the target virtual image has an identity mapping with the device user, and the device user trades virtual assets in the virtual world by generating a non-homogeneous identifier based on the identity mapping, and possesses ownership of the virtual assets through the transaction.
11. A virtual image adjustment device for a virtual world, comprising: a cold start preference feature extraction module, configured to obtain cold start data of a device user of an access device in the virtual world, and extract preference features for the cold start data to obtain cold start preference features, wherein the cold start data is composed of image data in an image modality and voice data in a voice modality; A virtual image reconstruction module is configured to perform virtual image reconstruction based on the cold start preference feature to obtain an initial virtual image of the device user in the virtual world; an interaction preference feature extraction module, configured to extract preference features from virtual interaction data of the device user in the virtual world to obtain interaction preference features, wherein the virtual interaction data includes user interaction data of the device user interacting with other users in the virtual world based on the initial virtual image; A virtual image adjustment module is configured to adjust the initial virtual image based on the interaction preference feature, obtain a target virtual image and output it to the access device; The extracting preference features from the cold start data to obtain cold start preference features includes: Inputting modality data of at least one modality included in the cold start data into a first feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality; Inputting the preference feature of the at least one modality into a first classifier for preference classification processing, and extracting output features of a specific layer of the first classifier as the cold start preference feature; The extracting preference features of the virtual interaction data of the device user in the virtual world to obtain the interaction preference features includes: Inputting the interaction data of at least one modality included in the virtual interaction data into a second feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality; The preference features of the at least one modality are input into a second classifier for classification, and the output features of a specific layer of the second classifier are extracted as the interaction preference features.
12. A virtual image adjustment device for a virtual world, comprising: processor; and a memory configured to store computer executable instructions that, when executed, cause the processor to: Acquire cold start data of a device user of an access device in the virtual world, and extract preference features for the cold start data to obtain cold start preference features, wherein the cold start data consists of image data in an image modality and voice data in a voice modality; Reconstructing the virtual image based on the cold start preference feature to obtain the initial virtual image of the device user in the virtual world; Extracting preference features from virtual interaction data of the device user in the virtual world to obtain interaction preference features, wherein the virtual interaction data includes user interaction data of the device user interacting with other users in the virtual world based on the initial virtual image; Adjusting the initial virtual image based on the interaction preference feature, obtaining a target virtual image and outputting it to the access device; The extracting preference features from the cold start data to obtain cold start preference features includes: Inputting modality data of at least one modality included in the cold start data into a first feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality; Inputting the preference feature of the at least one modality into a first classifier for preference classification processing, and extracting output features of a specific layer of the first classifier as the cold start preference feature; The extracting preference features of the virtual interaction data of the device user in the virtual world to obtain the interaction preference features includes: Inputting the interaction data of at least one modality included in the virtual interaction data into a second feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality; The preference features of the at least one modality are input into a second classifier for classification, and the output features of a specific layer of the second classifier are extracted as the interaction preference features.
13. A storage medium for storing computer executable instructions, wherein the computer executable instructions, when executed by a processor, implement the following process: Acquire cold start data of a device user of an access device in the virtual world, and extract preference features for the cold start data to obtain cold start preference features, wherein the cold start data consists of image data in an image modality and voice data in a voice modality; Reconstructing the virtual image based on the cold start preference feature to obtain the initial virtual image of the device user in the virtual world; Extracting preference features from virtual interaction data of the device user in the virtual world to obtain interaction preference features, wherein the virtual interaction data includes user interaction data of the device user interacting with other users in the virtual world based on the initial virtual image; Adjusting the initial virtual image based on the interaction preference feature, obtaining a target virtual image and outputting it to the access device; The extracting preference features from the cold start data to obtain cold start preference features includes: Inputting modality data of at least one modality included in the cold start data into a first feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality; Inputting the preference feature of the at least one modality into a first classifier for preference classification processing, and extracting output features of a specific layer of the first classifier as the cold start preference feature; The extracting preference features of the virtual interaction data of the device user in the virtual world to obtain the interaction preference features includes: Inputting the interaction data of at least one modality included in the virtual interaction data into a second feature extraction network of the corresponding modality to extract preference features, thereby obtaining the preference features of the at least one modality; The preference features of the at least one modality are input into a second classifier for classification, and the output features of a specific layer of the second classifier are extracted as the interaction preference features.
