CN116740238A

CN116740238A - Personalized configuration method, device, electronic equipment and storage medium

Info

Publication number: CN116740238A
Application number: CN202310612316.5A
Authority: CN
Inventors: 柴金详; 谭宏冰; 栾欣洋
Original assignee: Shanghai Movu Technology Co Ltd; Mofa Shanghai Information Technology Co Ltd
Current assignee: Shanghai Movu Technology Co Ltd; Mofa Shanghai Information Technology Co Ltd
Priority date: 2023-05-28
Filing date: 2023-05-28
Publication date: 2023-09-12

Abstract

The application provides a personalized configuration method, a personalized configuration device, electronic equipment and a computer readable storage medium, which are used for carrying out personalized configuration on a virtual object, wherein the method comprises the following steps: obtaining preference information of configuration personnel for the virtual object, wherein the preference information is used for indicating at least one of the following: personality, gender, age, long-term looks, style of grooming and application scene; inputting the preference information into a personalized configuration model to obtain an initial personalized policy of the virtual object, wherein the initial personalized policy comprises at least one of the following components: facial figures, makeup, hairstyles, languages, timbres, gestures, apparel, scenes, places, drawings, props and shots; and based on the initial personalized policy, utilizing a preview interface to perform personalized display on the virtual object. The configurator can automatically generate the corresponding initial personalized strategy by only inputting the preference information into the personalized configuration model, so that the complicated process of personalized configuration is simplified, and the configuration efficiency is improved.

Description

Personalized configuration method, device, electronic equipment and storage medium

Technical Field

The application relates to the technical fields of virtual persons, interactive designs and artificial intelligence, in particular to a personalized configuration method, a personalized configuration device, electronic equipment and a computer readable storage medium.

Background

The virtual objects include virtual humans, virtual animals, virtual cartoon figures, and the like. The virtual person is a personified image constructed by CG technology and operated in a code form, and has various interaction modes such as language communication, expression, action display and the like. The technology of the dummy person has been rapidly developed in the field of artificial intelligence and has been applied in many technical fields such as video, media, games, finance, travel, education, medical and so on.

The existing personalized configuration mode of the virtual object is as follows: the configurator needs to configure and view the corresponding effects item by item for each item (face image, makeup, hairstyle, language, tone, gesture, clothes, etc.), the configuration efficiency is low and the configuration workload is large.

Based on this, the application provides a personalized configuration method, a personalized configuration device, an electronic device and a computer readable storage medium, so as to improve the prior art.

Disclosure of Invention

The application aims to provide a personalized configuration method, a personalized configuration device, electronic equipment and a computer readable storage medium, which simplify the complicated process of personalized configuration and improve the configuration efficiency.

The application adopts the following technical scheme:

in a first aspect, the present application provides a personalized configuration method, configured to perform personalized configuration on a virtual object, where the method includes:

obtaining preference information of configuration personnel for the virtual object, wherein the preference information is used for indicating at least one of the following: personality, gender, age, long-term looks, style of grooming and application scene;

inputting the preference information into a personalized configuration model to obtain an initial personalized policy of the virtual object, wherein the initial personalized policy comprises at least one of the following components: facial figures, makeup, hairstyles, languages, timbres, gestures, apparel, scenes, places, drawings, props and shots;

and based on the initial personalized policy, utilizing a preview interface to perform personalized display on the virtual object.

The beneficial effect of this technical scheme lies in: the configurator is not required to configure each item and check the corresponding effect, and only needs to provide approximate preference information, and the preference information is input into the personalized configuration model to automatically generate the corresponding initial personalized strategy, so that the complicated process of personalized configuration is simplified, and the configuration efficiency is improved. The configurator is typically a customer's staff, whose work content includes personalized configuration of the virtual objects. The clients refer to clients of virtual object interactive application, typically B-end clients of enterprises, institutions, schools, hospitals and the like, and a small number of C-end clients; the display content of the virtual object is typically the display content provided by the client to its user, so that the user (of the client) can know the relevant information that the client wants to present to its user, such as enterprise history, enterprise culture, types of services and service offers that the enterprise can provide, office business services guides, educational and tutorial courses, health and life knowledge, etc.

Specifically, the virtual object can be enabled to better meet the requirements of users through personalized configuration, and the satisfaction degree and the using viscosity of the users to the virtual object are improved. The traditional personalized configuration of the virtual object needs to consume a great deal of time, and the initial personalized strategy can be quickly generated by utilizing the personalized configuration model to analyze and process the preference information, and the personalized strategy of the virtual object is generated according to the preference information of the configuration personnel, so that the complicated process of manual configuration is avoided, and the configuration efficiency is improved. The personalized effect of the virtual object can be displayed in real time through the preview interface, configuration personnel are helped to adjust personalized strategies in time, and configuration accuracy is improved. The method can be applied to various virtual objects, such as game roles, virtual assistants, virtual teaching and the like, and user experience can be improved through personalized configuration.

In some alternative embodiments, the method further comprises:

responding to the adjustment operation of the configurator for the initial personalized policy, acquiring the final personalized policy of the virtual object, updating the virtual object, and displaying the updated virtual object in real time by utilizing the preview interface;

Wherein the adjusting operation is used for adjusting part or all of the initial personalized policy.

The beneficial effect of this technical scheme lies in: the configurator can timely adjust the personalized strategy according to the feedback of the real-time preview interface, timely discover and correct errors in personalized configuration, and improve the accuracy and efficiency of configuration. Through real-time adjustment and updating of the personalized strategies, configuration personnel can more deeply understand the design and characteristics of the virtual object, and strengthen the sense of participation and the use wish of the virtual object. The personalized strategy of the virtual object can be iterated rapidly, the adjustment cost is reduced, the iteration speed is improved, and the user requirements are met and the user experience is optimized rapidly. The configuration personnel can adjust and customize the personalized strategy of the virtual object according to the self requirements, and the customization and flexibility of the virtual object are improved.

In some alternative embodiments, the method further comprises:

and updating model parameters of the personalized configuration model by using the preference information and the final personalized policy.

The beneficial effect of this technical scheme lies in: the model parameters of the personalized configuration model are one of important factors influencing the personalized effect, and the accuracy and precision of the personalized configuration can be improved by updating the model parameters, the working principle of the personalized configuration model can be better understood, and the interpretability of the personalized configuration is improved. By updating the model parameters based on the preference information and the final personalization policy, the risk of overfitting can be better reduced. Model parameters of the personalized configuration model are optimized for different configuration personnel and different scenes, and personalized optimization can be better performed according to the requirements of specific configuration personnel when the model parameters are updated, so that a better personalized effect is achieved.

In some alternative embodiments, the training process of the personalized configuration model includes:

acquiring a training set, wherein the training set comprises a plurality of training data, and each training data comprises sample preference information and labeling data of a final personalized strategy of the sample preference information;

for each training data in the training set, performing the following processing:

inputting sample preference information in the training data into a preset deep learning model to obtain prediction data of a final personalized strategy of the sample preference information;

updating model parameters of the deep learning model based on the prediction data and the labeling data of the final personalized strategy;

detecting whether a preset training ending condition is met; if yes, taking the trained deep learning model as the personalized configuration model; if not, continuing to train the deep learning model by using the next training data.

The beneficial effect of this technical scheme lies in: model parameters of the personalized configuration model are continuously optimized by using the labeling data, and the method has strong self-adaptability, so that the user requirements are better met. And inputting sample preference information in the training set into a preset deep learning model to obtain prediction data of a final personalized strategy, and combining the prediction data with labeling data to update model parameters of the deep learning model. The method can effectively improve the accuracy of personalized configuration. And performing repeated iterative optimization on the deep learning model through a plurality of training data in the training set to finally obtain an accurate and efficient personalized configuration model. And training by using the deep learning model, and deeply analyzing the relation between the sample preference information and the final personalized policy, so that the interpretability of the personalized configuration model is enhanced.

In some optional embodiments, the virtual object is used for explaining and introducing a preset product, and the method further includes:

obtaining product information of the preset product;

the inputting the preference information into a personalized configuration model to obtain an initial personalized policy of the virtual object comprises the following steps:

and inputting the preference information and the product information into the personalized configuration model to obtain an initial personalized policy of the virtual object.

The beneficial effect of this technical scheme lies in: the personalized configuration of the virtual object can be combined with the product, and the product selling point is applied to the personalized display of the virtual object, for example, when the preset product is a makeup product, the makeup of the virtual object can be configured into a corresponding makeup; when the preset product is a mobile phone, the clothing of the virtual object can be configured as a shirt and/or a skirt printed with images corresponding to the mobile phone product. By combining the preference information for the virtual object with the product information, the initial personalized policy more meeting the requirements of the configurator can be provided in a customized manner, so that the user experience is improved, and the products most suitable for the configurator can be recommended to the user more accurately, so that the sales opportunity is increased. The personalized strategy of the virtual object is adjusted and optimized by utilizing the personalized configuration model, so that the user requirements and preferences can be better met, and the sales conversion rate is improved.

In some optional embodiments, the personalized configuration model is a semantic extraction model, and the obtaining preference information of the configurator for the virtual object includes:

and responding to the text input operation, the voice input operation or the image input operation of the configurator, and acquiring the preference information of the configurator for the virtual object.

The beneficial effect of this technical scheme lies in: the semantic extraction model can carry out deep analysis according to preference information input by the configurator, and accurately understand the intention of the configurator, so that a personalized strategy which meets the requirements of the configurator better is generated. By adopting an automatic semantic extraction model, the intervention of manpower in the personalized configuration process can be reduced, and the configuration cost is reduced. By responding to the text input operation or the voice input operation, preference information of configuration personnel for the virtual object can be obtained more quickly, and configuration efficiency is improved.

In some alternative embodiments, the method further comprises:

acquiring historical configuration data of the configuration personnel, wherein the historical configuration data comprises personalized strategies used by the configuration personnel;

Inputting the preference information into the personalized configuration model to obtain a plurality of alternative personalized strategies of the virtual object;

and selecting one of a plurality of alternative personalized policies from the plurality of alternative personalized policies as the initial personalized policy based on the historical configuration data.

The beneficial effect of this technical scheme lies in: by considering the historical configuration data, the use habit of the configurator can be better understood, an initial personalized strategy which meets the requirements of the configurator is provided for the configurator, and the user satisfaction is improved. Specifically, a plurality of candidate personalized strategies can be generated according to preference information provided by configuration personnel, the personalized strategy with the highest matching degree with the historical configuration data is selected from the plurality of candidate personalized strategies to serve as an initial personalized strategy, the obtained initial personalized strategy accords with the use habit of the configuration personnel, the requirements of the configuration personnel are met more easily, the configuration personnel are prevented from adjusting and modifying for many times, and the configuration efficiency is provided.

In a second aspect, the present application provides a personalized configuration apparatus for personalized configuration of a virtual object, the apparatus comprising:

a preference obtaining module, configured to obtain preference information of a configurator for the virtual object, where the preference information is used to indicate at least one of the following: personality, gender, age, long-term looks, style of grooming and application scene;

The initial policy acquisition module is used for inputting the preference information into the personalized configuration model to obtain an initial personalized policy of the virtual object, and the initial personalized policy comprises at least one of the following: facial figures, makeup, hairstyles, languages, timbres, gestures, apparel, scenes, places, drawings, props and shots;

and the personalized display module is used for utilizing a preview interface to perform personalized display on the virtual object based on the initial personalized policy.

In some alternative embodiments, the apparatus further comprises:

the personalized adjustment module is used for responding to the adjustment operation of the configurator for the initial personalized policy, acquiring the final personalized policy of the virtual object, updating the virtual object and displaying the updated virtual object in real time by utilizing the preview interface;

In some alternative embodiments, the apparatus further comprises:

and the model updating module is used for updating the model parameters of the personalized configuration model by utilizing the preference information and the final personalized policy.

In some optional embodiments, the virtual object is used for explaining and introducing a preset product, and the device further includes:

the product information module is used for acquiring product information of the preset product;

the initial policy acquisition module is configured to:

In some optional embodiments, the personalized configuration model is a semantic extraction model, and the preference acquisition module is configured to:

In some alternative embodiments, the apparatus further comprises:

the history configuration module is used for acquiring history configuration data of the configuration personnel, wherein the history configuration data comprises personalized strategies used by the configuration personnel;

the initial policy acquisition module includes:

the alternative strategy module is used for inputting the preference information into the personalized configuration model to obtain a plurality of alternative personalized strategies of the virtual object;

and the policy determining module is used for selecting one of the alternative personalized policies from the plurality of personalized policies based on the historical configuration data to serve as the initial personalized policy.

In a third aspect, the application provides an electronic device comprising a memory storing a computer program and at least one processor configured to implement the following steps when executing the computer program:

In some alternative embodiments, the at least one processor is further configured to implement the following steps when executing the computer program:

In some alternative embodiments, the virtual object is used to introduce a preset product, and the at least one processor is further configured to implement the following steps when executing the computer program:

Obtaining product information of the preset product;

the at least one processor is configured to input the preference information into a personalization configuration model when executing the computer program to obtain an initial personalization policy for the virtual object in the following manner:

In some alternative embodiments, the personalized configuration model is a semantic extraction model, and the at least one processor is configured to obtain configuration personnel preference information for the virtual object when executing the computer program by:

In a fourth aspect, the present application provides a computer-readable storage medium storing a computer program which, when executed by at least one processor, performs the steps of any of the methods described above or performs the functions of the electronic device described above.

Drawings

The application will be further described with reference to the drawings and embodiments.

Fig. 1 shows a flow diagram of a personalized configuration method according to an embodiment of the present application.

Fig. 2 is a schematic flow chart of another personalized configuration method according to an embodiment of the present application.

Fig. 3 is a schematic flow chart of another personalized configuration method according to an embodiment of the present application.

Fig. 4 shows a flowchart of acquiring an initial personalized policy according to an embodiment of the present application.

Fig. 5 is a schematic structural diagram of a personalized configuration device according to an embodiment of the present application.

Fig. 6 is a block diagram of an electronic device according to an embodiment of the present application.

Fig. 7 is a schematic structural diagram of a program product according to an embodiment of the present application.

Detailed Description

The technical scheme of the present application will be described below with reference to the drawings and the specific embodiments of the present application, and it should be noted that, on the premise of no conflict, new embodiments may be formed by any combination of the embodiments or technical features described below.

In embodiments of the present application, "at least one" means one or more, and "a plurality" means two or more. "and/or", describes an association relationship of an association object, and indicates that there may be three relationships, for example, a and/or B, and may indicate: a alone, a and B together, and B alone, wherein a, B may be singular or plural. The character "/" generally indicates that the context-dependent object is an "or" relationship. "at least one of" or the like means any combination of these items, including any combination of single item(s) or plural items(s). For example, at least one (one) of a, b or c may represent: a, b, c, a and b, a and c, b and c, a and b and c, wherein a, b and c can be single or multiple. It is noted that "at least one" may also be interpreted as "one (a) or more (a)".

It is also noted that, in embodiments of the present application, words such as "exemplary" or "such as" are used to mean serving as an example, instance, or illustration. Any implementation or design described as "exemplary" or "e.g." in the examples of this application should not be construed as preferred or advantageous over other implementations or designs. Rather, the use of words such as "exemplary" or "such as" is intended to present related concepts in a concrete fashion.

The technical field and related terms of the embodiments of the present application are briefly described below.

The virtual objects include virtual humans, virtual animals, virtual cartoon figures, and the like. The virtual person is a personified image constructed by CG technology and operated in a code form, and has various interaction modes such as language communication, expression, action display and the like. The technology of virtual persons has been rapidly developed in the field of artificial intelligence and has been applied in many technical fields such as video, media, games, finance, travel, education, medical treatment, etc., and not only can a virtual host, a virtual object, a virtual idol, a virtual customer service, a virtual lawyer, a virtual financial advisor, a virtual teacher, a virtual doctor, a virtual instructor, a virtual assistant, etc. be customized, but also a video can be generated through text or audio one-key. In the virtual people, the service type virtual people mainly have the functions of replacing real people to serve and provide daily accompaniment, are the virtualization of service type roles in reality, and have the industrial value of mainly reducing the cost of the existing service type industry and enhancing the cost reduction of the stock market.

Artificial intelligence (Artificial Intelligence, AI) is the theory, method, technique and application system that uses a digital computer or a machine controlled by a digital computer to simulate, extend and extend human intelligence, sense the environment, acquire knowledge and use the knowledge to obtain optimal results. In other words, artificial intelligence is an integrated technology of computer science that attempts to understand the essence of intelligence and to produce a new intelligent machine that can react in a similar way to human intelligence. The design principle and the implementation method of various intelligent machines are researched by artificial intelligence, so that the machines have the functions of perception, reasoning and decision. The artificial intelligence technology is a comprehensive subject, and relates to the technology with wide fields, namely the technology with a hardware level and the technology with a software level. Artificial intelligence infrastructure technologies generally include technologies such as sensors, dedicated artificial intelligence chips, cloud computing, distributed storage, big data processing technologies, operation/interaction systems, mechatronics, and the like. The artificial intelligence software technology mainly comprises a computer vision technology, a voice processing technology, a natural language processing technology, machine learning/deep learning, automatic driving, intelligent traffic and other directions.

Machine Learning (ML) is a multi-domain interdisciplinary, involving multiple disciplines such as probability theory, statistics, approximation theory, convex analysis, algorithm complexity theory, etc. The computer program may learn experience E given a certain class of tasks T and performance metrics P, and increase with experience E if its performance in task T happens to be measured by P. Machine learning is specialized in studying how a computer simulates or implements learning behavior of a human to acquire new knowledge or skills, reorganizing existing knowledge structures to continually improve its own performance. Machine learning is the core of artificial intelligence, a fundamental approach to letting computers have intelligence, which is applied throughout various areas of artificial intelligence.

Deep learning is a special machine learning by which the world is represented using a hierarchy of nested concepts, each defined as being associated with a simple concept, and achieving great functionality and flexibility, while a more abstract representation is computed in a less abstract way. Machine learning and deep learning typically include techniques such as artificial neural networks, belief networks, reinforcement learning, transfer learning, induction learning, teaching learning, and the like.

The virtual object interaction application is used for providing virtual object interaction functions. The virtual human interactive application may simulate human communication and behavior and interact with the user. Such software (referred to as virtual human interactive applications) is typically driven by artificial intelligence and natural language processing techniques and is capable of interacting with a user by means of text, speech or images, etc.

(personalized configuration method)

Referring to fig. 1, fig. 1 shows a flow chart of a personalized configuration method according to an embodiment of the present application.

The application provides a personalized configuration method, which is used for carrying out personalized configuration on a virtual object, and comprises the following steps:

step S101: obtaining preference information of configuration personnel for the virtual object, wherein the preference information is used for indicating at least one of the following: personality, gender, age, long-term looks, style of grooming and application scene;

step S102: inputting the preference information into a personalized configuration model to obtain an initial personalized policy of the virtual object, wherein the initial personalized policy comprises at least one of the following components: facial figures, makeup, hairstyles, languages, timbres, gestures, apparel, scenes, places, drawings, props and shots;

Step S103: and based on the initial personalized policy, utilizing a preview interface to perform personalized display on the virtual object.

In an embodiment of the present application, the virtual object includes one or more of a virtual person, a virtual animal, and a virtual cartoon character. As one example, the virtual object is a virtual person "JING" (chinese name: mirror).

The personalized configuration method can be operated on the electronic equipment, configuration personnel can use the corresponding terminal equipment to realize a series of user operations, the electronic equipment and the terminal equipment can be independent, and the electronic equipment and the terminal equipment can be integrated.

The terminal equipment is provided with a virtual object interactive application, and the virtual object interactive application is provided with a personalized configuration interface and a preview interface, wherein the personalized configuration interface and the preview interface can be independent interfaces or a unified interface. The personalized configuration interface is used for executing step S101 and step S102, and the preview interface is used for executing step S103.

The electronic device may be a computer, a server (including a cloud server), or the like having computing power. The terminal device is not limited in the embodiment of the application, and may be, for example, an intelligent terminal device with a display screen and a microphone, such as a mobile phone, a tablet computer, a notebook computer, a desktop computer, an intelligent wearable device, or the terminal device may be a workstation or a console. The display screen may be a touch display screen or a non-touch display screen.

Therefore, the configurator is not required to configure each item and check the corresponding effect, and only needs to provide general preference information, and the preference information is input into the personalized configuration model to automatically generate the corresponding initial personalized policy, so that the complicated process of personalized configuration is simplified, and the configuration efficiency is improved. The configurator is typically a customer's staff, whose work content includes personalized configuration of the virtual objects. The clients refer to clients of virtual object interactive application, typically B-end clients of enterprises, institutions, schools, hospitals and the like, and a small number of C-end clients; the display content of the virtual object is typically the display content provided by the client to its user, so that the user (of the client) can know the relevant information that the client wants to present to its user, such as enterprise history, enterprise culture, types of services and service offers that the enterprise can provide, office business services guides, educational and tutorial courses, health and life knowledge, etc.

In some embodiments, the personalized configuration model is a semantic extraction model, and the obtaining preference information of the configurator for the virtual object (step S101) includes:

The personalized configuration interface is provided with a text input box, an audio input button and an image uploading control, and a configurator can input corresponding text information by using the text input box or input corresponding voice information by using a microphone after clicking the audio input button. The configurator can also upload corresponding image information by using the image uploading control.

As one example, a configurator enters a piece of text information (i.e., preference information) through a text entry box, the content of which is: "Mild Zhixian sister with long hair like seaweed and double eyes with sapphire blue always show the calm and maturity of more than usual people, and there is a great chance of fashion trend.

As another example, after the configurator clicks the audio input button, a piece of voice information (i.e., preference information) is input by using the microphone, the contents being: "fashion cool boy, quite deep in understanding from media operation and food, get up to the hall and get down to the kitchen, and enjoy sharing own food menu with people.

As yet another example, a configurator uploaded a picture of his or her hand-drawn with an image upload control, the contents being: "a lovely cartoon character blinks the eyes, and" biyer ".

In some embodiments, when the input information is text information, the semantic extraction model to which the input information corresponds is a pre-trained language model based on deep learning; when the input information is voice information, the semantic extraction model corresponding to the input information comprises a voice-to-text model based on deep learning and a pre-training language model based on deep learning; when the input information is image information, the semantic extraction model corresponding to the input information is a semantic segmentation model based on deep learning.

Thus, a deep learning-based model is used to extract semantic features of the input information. For different types of input information, different deep learning models are adopted for semantic extraction, including a pre-training language model, a voice-to-text model, a semantic segmentation model and the like.

For text information, pre-trained language models, such as BERT, GPT, etc., are employed to extract semantic features of the input text. The models can learn the structure and semantic information of the language by pre-training a large amount of texts, so that the semantic information of the input texts can be effectively extracted.

For voice information, a voice-to-text model based on deep learning, such as CTC, transformer, is adopted to convert the voice information into text information, and a pre-training language model is adopted to extract semantic features of the text information. Thus, semantic information related to the input information can be extracted from the voice information.

For image information, a semantic segmentation model based on deep learning, such as UNet, deep Lab and the like, is adopted to segment the image, and semantic information corresponding to each pixel point is extracted. Thus, semantic information related to the input information can be accurately extracted from the image.

The method has the advantages that semantic features of input information can be extracted more accurately by using a semantic extraction model based on deep learning, so that accuracy and efficiency of semantic understanding are improved; the pre-training language model can improve the natural language processing capacity, including text classification, emotion analysis, machine translation and other aspects; by adopting different deep learning models, the multi-mode information can be processed, including text information, voice information, image information and the like, so that more comprehensive information processing is realized.

Therefore, the semantic extraction model can carry out deep analysis according to preference information input by the configurator, and the intention of the configurator can be accurately understood, so that a personalized strategy which meets the requirements of the configurator better is generated. By adopting an automatic semantic extraction model, the intervention of manpower in the personalized configuration process can be reduced, and the configuration cost is reduced. By responding to the text input operation or the voice input operation, preference information of configuration personnel for the virtual object can be obtained more quickly, and configuration efficiency is improved.

Referring to fig. 2, fig. 2 is a schematic flow chart of another personalized configuration method according to an embodiment of the present application.

In some embodiments, the method further comprises:

step S104: responding to the adjustment operation of the configurator for the initial personalized policy, acquiring the final personalized policy of the virtual object, updating the virtual object, and displaying the updated virtual object in real time by utilizing the preview interface;

The adjustment operation can be performed on a personalized configuration interface or on a preview interface.

For example, the adjustment operation may be to input adjustment text information into the personalized configuration model in the personalized configuration interface.

The configuration personnel can see the personalized collocation of the virtual object in the preview interface as follows: "Mushroom head image, lovely and even, standing and living broadcast scene", the configurator can return to the personalized configuration interface, input and adjust the text information to the personalized configuration model, the content is as follows: the mushroom head image is replaced by a sister image, and the lovely beatifying sound is replaced by a cool female sound.

For another example, the adjustment operation may be clicking a corresponding personalized radio frame in the preview interface to replace. The personalized radio comprises at least one of the following items: facial figures, makeup, hairstyles, languages, timbres, gestures, apparel, scenes, places, drawings, props and shots.

Therefore, configuration personnel can timely adjust the personalized strategy according to the feedback of the real-time preview interface, timely discover and correct errors in personalized configuration, and improve the accuracy and efficiency of configuration. Through real-time adjustment and updating of the personalized strategies, configuration personnel can more deeply understand the design and characteristics of the virtual object, and strengthen the sense of participation and the use wish of the virtual object. The personalized strategy of the virtual object can be iterated rapidly, the adjustment cost is reduced, the iteration speed is improved, and the user requirements are met and the user experience is optimized rapidly. The configuration personnel can adjust and customize the personalized strategy of the virtual object according to the self requirements, and the customization and flexibility of the virtual object are improved.

In some embodiments, the method further comprises:

Therefore, the model parameters of the personalized configuration model are one of important factors influencing the personalized effect, the accuracy and precision of the personalized configuration can be improved by updating the model parameters, the working principle of the personalized configuration model can be better understood, and the interpretability of the personalized configuration is improved. By updating the model parameters based on the preference information and the final personalization policy, the risk of overfitting can be better reduced. Model parameters of the personalized configuration model are optimized for different configuration personnel and different scenes, and personalized optimization can be better performed according to the requirements of specific configuration personnel when the model parameters are updated, so that a better personalized effect is achieved.

In some embodiments, the training process of the personalized configuration model comprises:

Therefore, the model parameters of the personalized configuration model are continuously optimized by using the annotation data, and the method has strong self-adaptability, so that the user requirements are better met. And inputting sample preference information in the training set into a preset deep learning model to obtain prediction data of a final personalized strategy, and combining the prediction data with labeling data to update model parameters of the deep learning model. The method can effectively improve the accuracy of personalized configuration. And performing repeated iterative optimization on the deep learning model through a plurality of training data in the training set to finally obtain an accurate and efficient personalized configuration model. And training by using the deep learning model, and deeply analyzing the relation between the sample preference information and the final personalized policy, so that the interpretability of the personalized configuration model is enhanced.

The method for obtaining the personalized configuration model is not limited in the embodiment of the application, in some embodiments, the models can be obtained through training, in other embodiments, the models can be obtained through training in advance.

When the personalized configuration model is obtained through training in a deep learning mode, a proper amount of neuron computing nodes and a multi-layer operation hierarchical structure are established through design, a proper input layer and a proper output layer are selected, a preset deep learning model corresponding to the personalized configuration model (namely an initial model corresponding to the personalized configuration model) can be obtained, a functional relation from input to output is established through learning and tuning of the deep learning model, although the functional relation between the input and the output cannot be found 100%, the functional relation between the input and the output can be approximated as close as possible, and accordingly the obtained personalized configuration model can obtain corresponding output data based on the input data.

Training a deep learning model by using a training set corresponding to the personalized configuration model, and quickly modeling by learning a small number of samples, wherein training errors of the deep learning model can be gradually reduced in the continuous training process, and the optimal weight is stored and read; recording the accuracy of the training set and the verification set, and facilitating parameter adjustment (adjustment of model parameters); the model parameters of the deep learning model are updated, so that the model can be better fitted with data, the generalization capability is effectively achieved, and the robustness and the fitting precision are improved.

In some alternative embodiments, the historical data may be data mined to obtain sample data in the training set. That is, the sample data may be collected during the real interaction. In addition, the sample data may be automatically generated by using a GAN model generation network.

The GAN model generates an countermeasure network (Generative Adversarial Network) composed of a generation network and a discrimination network. The generation network samples randomly from the potential space (latency space) as input, the output of which needs to mimic as much as possible the real samples in the training set. The input of the discrimination network is then the real sample or the output of the generation network, the purpose of which is to distinguish the output of the generation network as far as possible from the real sample. And the generation of the network should be as fraudulent as possible to discriminate the network. The two networks are mutually opposed and continuously adjust parameters, and the final purpose is that the judging network can not judge whether the output result of the generated network is real or not. The GAN model can be used for generating a large amount of sample data for the training process of the personalized configuration model, so that the data volume of original data acquisition can be effectively reduced, and the cost of data acquisition and labeling is greatly reduced.

The training process of the personalized configuration model is not limited, and for example, a training mode of supervised learning, a training mode of semi-supervised learning or a training mode of unsupervised learning can be adopted.

When a training mode of supervised learning or semi-supervised learning is adopted, the method for acquiring the annotation data is not limited, and for example, a manual annotation mode or an automatic annotation or semi-automatic annotation mode can be adopted. When the sample data is acquired in the real interaction process, the real data can be acquired from the historical data in a keyword extraction mode to serve as the annotation data.

The training ending condition in the training process of the personalized configuration model is not limited, for example, the training times can reach the preset times (the preset times are, for example, 1 time, 3 times, 10 times, 100 times, 1000 times, 10000 times, etc.), or the training data in the training set can be all trained once or a plurality of times, or the total loss value obtained in the training is not more than the preset loss value.

Referring to fig. 3, fig. 3 is a schematic flow chart of another personalized configuration method according to an embodiment of the present application.

In some embodiments, the virtual object is used for explaining and introducing a preset product, and the method further includes:

step S105: obtaining product information of the preset product;

the inputting the preference information into a personalized configuration model to obtain an initial personalized policy of the virtual object (step S102), including:

step S102a: and inputting the preference information and the product information into the personalized configuration model to obtain an initial personalized policy of the virtual object.

In the embodiment of the application, the product information can be an introduction of a product.

For example, for a foundation, the product information is: the foundation liquid is a foundation liquid with long-term effects of moistening and moisturizing and making up. The novel durable make-up effect technology is adopted, so that make-up can be kept for more than 12 hours without removing make-up or being dull, and frequent make-up can not be carried out. Meanwhile, the cosmetic also contains various moisturizing components, so that sufficient moisture and moisture are provided for the skin, the cosmetic is more fit to the skin, and the cosmetic is more natural. It also has added SPF15 sun-screening component, which can help to resist ultraviolet injury. In addition, the foundation liquid also provides a plurality of choices of different complexion, and meets the requirements of different complexion. The selling points of the powder base liquid are as follows: durable no makeup, moistening and moisturizing, strong sun-screening capability and selectable various complexion.

Aiming at a Bluetooth headset, the product information is: the Bluetooth earphone adopts a high-quality audio technology, can provide clearer and vivid sound effects, and enables you to enjoy entertainment contents such as music, movies, games and the like. The earphone is connected by using the Bluetooth technology, does not need to use any cable, is convenient and fast, and can effectively reduce complicated operation steps and improve the use comfort. Meanwhile, the system also supports multi-equipment connection and can be easily switched for use. The earphone adopts a comfortable ear-hook design, and is matched with a soft earmuff material, so that the pressure during long-time wearing can be effectively reduced, and the earphone is not easy to fatigue. In addition, the waterproof sweat-resistant and waterproof cloth also has sweat-resistant and waterproof functions, and is suitable for wearing in outdoor exercises or fierce exercises. The earphone has light volume and portability, can be carried at any time in travel, work or daily travel, and can be used in various environments. The battery has long service life, can stand by for tens of hours and can be used for a plurality of hours, frequent charging is not needed, and the service time is effectively prolonged. The selling point of the Bluetooth headset is high-quality tone quality, wireless connection, comfort, portability and long standby time. These features can meet consumer demand for a good and convenient earphone.

Therefore, the personalized configuration of the virtual object can be combined with the product, the product selling point is applied to the personalized display of the virtual object, for example, when the preset product is a makeup product, the makeup of the virtual object can be configured into a corresponding makeup; when the preset product is a mobile phone, the clothing of the virtual object can be configured as a shirt and/or a skirt printed with images corresponding to the mobile phone product. By combining the preference information for the virtual object with the product information, the initial personalized policy more meeting the requirements of the configurator can be provided in a customized manner, so that the user experience is improved, and the products most suitable for the configurator can be recommended to the user more accurately, so that the sales opportunity is increased. The personalized strategy of the virtual object is adjusted and optimized by utilizing the personalized configuration model, so that the user requirements and preferences can be better met, and the sales conversion rate is improved.

Referring to fig. 4, fig. 4 is a schematic flow chart of acquiring an initial personalized policy according to an embodiment of the present application.

In some embodiments, the method further comprises:

step S201: inputting the preference information into the personalized configuration model to obtain a plurality of alternative personalized strategies of the virtual object;

step S202: and selecting one of a plurality of alternative personalized policies from the plurality of alternative personalized policies as the initial personalized policy based on the historical configuration data.

In some embodiments, the step S202 may include:

matching each alternative personalized policy with personalized policies in the historical configuration data one by one to obtain corresponding policy similarity;

and taking the alternative personalized strategy with the highest strategy similarity as an initial personalized strategy.

Therefore, by considering the historical configuration data, the use habit of the configurator can be better understood, an initial personalized strategy which meets the requirements of the configurator is provided for the configurator, and the user satisfaction is improved. Specifically, a plurality of candidate personalized strategies can be generated according to preference information provided by configuration personnel, the personalized strategy with the highest matching degree with the historical configuration data is selected from the plurality of candidate personalized strategies to serve as an initial personalized strategy, the obtained initial personalized strategy accords with the use habit of the configuration personnel, the requirements of the configuration personnel are met more easily, the configuration personnel are prevented from adjusting and modifying for many times, and the configuration efficiency is provided.

In a specific application scenario, the embodiment of the application further provides a personalized configuration method, which is used for personalized configuration of the virtual object, and the method comprises the following steps:

The method further comprises the steps of:

The training process of the personalized configuration model comprises the following steps:

The virtual object is used for explaining and introducing a preset product, and the method further comprises the following steps:

obtaining product information of the preset product;

The personalized configuration model is a semantic extraction model, and the obtaining the preference information of the configuration personnel for the virtual object comprises the following steps:

The method further comprises the steps of:

(personalized configuration device)

Referring to fig. 5, fig. 5 is a schematic structural diagram of a personalized configuration device according to an embodiment of the present application.

The application provides a personalized configuration device, which is used for personalized configuration of a virtual object, and comprises:

a preference obtaining module 101, configured to obtain preference information of a configurator for the virtual object, where the preference information is used to indicate at least one of the following: personality, gender, age, long-term looks, style of grooming and application scene;

an initial policy obtaining module 102, configured to input the preference information into a personalized configuration model, so as to obtain an initial personalized policy of the virtual object, where the initial personalized policy includes at least one of the following: facial figures, makeup, hairstyles, languages, timbres, gestures, apparel, scenes, places, drawings, props and shots;

and the personalized display module 103 is used for performing personalized display on the virtual object by utilizing a preview interface based on the initial personalized policy.

In some alternative embodiments, the apparatus further comprises:

the initial policy acquisition module 102 is configured to:

In some optional embodiments, the personalized configuration model is a semantic extraction model, and the preference obtaining module 101 is configured to:

In some alternative embodiments, the apparatus further comprises:

the initial policy acquisition module 102 includes:

(electronic device)

The embodiment of the application also provides an electronic device, the specific embodiment of which is consistent with the embodiment described in the method embodiment and the achieved technical effect, and part of the contents are not repeated.

The application provides an electronic device comprising a memory storing a computer program and at least one processor configured to implement the following steps when executing the computer program:

obtaining product information of the preset product;

Referring to fig. 6, fig. 6 is a block diagram of an electronic device 10 according to an embodiment of the present application.

The electronic device 10 may for example comprise at least one memory 11, at least one processor 12 and a bus 13 connecting the different platform systems.

Memory 11 may include readable media in the form of volatile memory, such as Random Access Memory (RAM) 111 and/or cache memory 112, and may further include Read Only Memory (ROM) 113.

The memory 11 also stores a computer program executable by the processor 12 to cause the processor 12 to implement the steps of any of the methods described above.

Memory 11 may also include utility 114 having at least one program module 115, such program modules 115 include, but are not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of which may include an implementation of a network environment.

Accordingly, the processor 12 may execute the computer programs described above, as well as may execute the utility 114.

The processor 12 may employ one or more application specific integrated circuits (ASICs, application Specific Integrated Circui t), DSPs, programmable logic devices (PLD, programmableLogic devices), complex programmable logic devices (CPLDs, complex Programmable Logic Device), field programmable gate arrays (FPGAs, fields-Programmable Gate Array), or other electronic components.

Bus 13 may be a local bus representing one or more of several types of bus structures including a memory bus or memory controller, a peripheral bus, an accelerated graphics port, a processor, or any of a variety of bus architectures.

The electronic device 10 may also communicate with one or more external devices such as a keyboard, pointing device, bluetooth device, etc., as well as one or more devices capable of interacting with the electronic device 10 and/or with any device (e.g., router, modem, etc.) that enables the electronic device 10 to communicate with one or more other computing devices. Such communication may be via the input-output interface 14. Also, the electronic device 10 may communicate with one or more networks such as a Local Area Network (LAN), a Wide Area Network (WAN) and/or a public network, such as the Internet, through a network adapter 15. The network adapter 15 may communicate with other modules of the electronic device 10 via the bus 13. It should be appreciated that although not shown, other hardware and/or software modules may be used in connection with the electronic device 10 in actual applications, including, but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, data backup storage platforms, and the like.

(computer-readable storage Medium)

The embodiment of the application also provides a computer readable storage medium, and the specific embodiment of the computer readable storage medium is consistent with the embodiment recorded in the method embodiment and the achieved technical effect, and part of the contents are not repeated.

The computer readable storage medium stores a computer program which, when executed by at least one processor, performs the steps of any of the methods or performs the functions of any of the electronic devices described above.

Referring to fig. 7, fig. 7 is a schematic structural diagram of a program product according to an embodiment of the present application.

The program product is for implementing the steps of any of the methods described above or for implementing the functions of any of the electronic devices described above. The program product may take the form of a portable compact disc read-only memory (CD-ROM) and comprises program code and may be run on a terminal device, such as a personal computer. However, the program product of the present application is not limited thereto, and in the embodiments of the present application, the readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device. The program product may employ any combination of one or more readable media. The readable medium may be a readable signal medium or a readable storage medium. The readable storage medium can be, for example, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the readable storage medium would include the following: an electrical connection having one or more wires, a portable disk, a hard disk, random Access Memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

The computer readable storage medium may include a data signal propagated in baseband or as part of a carrier wave, with readable program code embodied therein. Such a propagated data signal may take any of a variety of forms, including, but not limited to, electro-magnetic, optical, or any suitable combination of the foregoing. A readable storage medium may also be any readable medium that can transmit, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device. Program code embodied on a readable storage medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing. Program code for carrying out operations of the present invention may be written in any combination of one or more programming languages, including an object oriented programming language such as Java, C++ or the like and conventional procedural programming languages, such as the C programming language or similar programming languages. The program code may execute entirely on the user's computing device, partly on the user's device, as a stand-alone software package, partly on the user's computing device, partly on a remote computing device, or entirely on the remote computing device or server. In the case of remote computing devices, the remote computing device may be connected to the user computing device through any kind of network, including a Local Area Network (LAN) or a Wide Area Network (WAN), or may be connected to an external computing device (e.g., connected via the Internet using an Internet service provider).

The present application has been described in terms of its purpose, performance, advancement, and novelty, and the like, and is thus adapted to the functional enhancement and use requirements highlighted by the patent statutes, but the description and drawings are not limited to the preferred embodiments of the present application, and therefore, all equivalents and modifications that are included in the construction, apparatus, features, etc. of the present application shall fall within the scope of the present application.

Claims

1. A method for personalized configuration of a virtual object, the method comprising:

2. The personalized configuration method according to claim 1, wherein the method further comprises:

3. The personalized configuration method according to claim 2, wherein the method further comprises:

4. A personalized configuration method according to claim 3, wherein the training process of the personalized configuration model comprises:

5. The personalized configuration method according to claim 1, wherein the virtual object is used for explaining and introducing a preset product, and the method further comprises:

obtaining product information of the preset product;

6. The personalized configuration method according to claim 5, wherein the personalized configuration model is a semantic extraction model, and the obtaining preference information of the configurator for the virtual object comprises:

7. The personalized configuration method according to claim 1, wherein the method further comprises:

8. A personalized configuration apparatus for personalized configuration of a virtual object, the apparatus comprising:

9. An electronic device comprising a memory and at least one processor, the memory storing a computer program, the at least one processor being configured to implement the following steps when executing the computer program:

10. A computer-readable storage medium, characterized in that it stores a computer program which, when executed by at least one processor, implements the steps of the method of any of claims 1-7 or implements the functionality of the electronic device of claim 9.