Method and device for generating virtual space based on user preference information and electronic equipment

Abstract

The invention discloses a method and device for generating a virtual space based on user preference information and electronic equipment, and relates to the technical field of meta universe and the field of artificial intelligence. The method comprises the following steps: acquiring style preference information of a user; according to the style preference information, external contour information of the virtual space is determined, and the external contour information at least comprises house type information of the virtual space; according to the style preference information and the external contour information of the user, style information of each independent space in the virtual space is determined, and the style information of each independent space at least comprises type information of built-in articles in each independent space and placement position information of each built-in article; and generating a target virtual space corresponding to the user according to the external contour information and the style information of each independent space. The technical problem of low generation efficiency of the personalized virtual space of the user in the prior art is solved.

Description

Technical Field

[0001] The present application relates to the field of metaverse technology and artificial intelligence, and more specifically, to a method, device, and electronic device for generating a virtual space based on user preference information. Background Art

[0002] In the current field of virtual content creation, users have an increasing demand for personalized virtual spaces, especially in the metaverse, virtual reality games, and social platforms, where users want to have a virtual environment that reflects their personal tastes and lifestyles.

[0003] However, in the existing technologies, although it is possible to preliminarily analyze the user's style preferences through the space environment photos uploaded by the user or the preferred style images actively selected by the user, these technologies have not gone deep into the level of efficiently generating and arranging personalized virtual spaces according to the user's style preferences. For example, when a user wants to have a space that matches his or her real style in the virtual world, or wants a creative space that completely matches his or her imagination, the existing technologies often require the user to manually select and adjust each virtual item, including office supplies, decorations, etc., which is not only time-consuming, but also inefficient, and it is difficult to meet the user's needs for quickly generating personalized virtual spaces.

[0004] In addition, when generating personalized virtual spaces, existing technologies lack an in-depth understanding of the user's overall style preferences and effective control of the consistency of styles in various areas within the virtual space. For example, even if furniture of a specific style can be generated according to user preferences, existing technologies often have deficiencies in overall space layout, such as the coordination of styles in different areas, resulting in the generated virtual space appearing disorganized in style, which does not meet the user's expectations for the unity of the overall style of the space.

[0005] Therefore, the technical problems in the prior art regarding the low efficiency of generating personalized virtual space for users and the lack of style coordination and consistency need to be solved urgently. Summary of the invention

[0006] The embodiments of the present application provide a method, an apparatus, and an electronic device for generating a virtual space based on user preference information, so as to at least solve the technical problem of low efficiency in generating a personalized virtual space for a user in the prior art.

[0007] According to one aspect of an embodiment of the present application, a method for generating a virtual space based on user preference information is provided, comprising: obtaining the user's style preference information; determining external contour information of the virtual space based on the style preference information, wherein the external contour information includes at least apartment type information of the virtual space; determining style information of each independent space in the virtual space based on the user's style preference information and the external contour information, wherein the style information of each independent space includes at least type information of built-in items in each independent space and placement position information of each built-in item; generating a target virtual space corresponding to the user based on the external contour information and the style information of each independent space.

[0008] Optionally, obtaining the user's style preference information includes: obtaining a target image uploaded by the user, wherein the target image includes a spatial environment provided by the user for reference; extracting spatial environment style features from the target image; and using the extracted spatial environment style features as the user's style preference information.

[0009] Optionally, obtaining the user's style preference information includes: obtaining a target file uploaded by the user, wherein the target file is used to record the user's description information of the expected spatial environment; extracting text description features from the target file, wherein the text description features include at least a first type of description features and a second type of description features, the first type of description features characterizing items that the user expects to appear in the spatial environment and the placement of the items, and the second type of description features characterizing items that the user expects to be prohibited from appearing in the spatial environment; and using the text description features as the user's style preference information.

[0010] Optionally, determining the external contour information of the virtual space based on the style preference information includes: inputting the style preference information into a room generation model, converting the room features in the style preference information into feature vectors through the room generation model, and using the feature vectors to determine the room type corresponding to the user, the number, layout and area of ​​each room type, wherein the room generation model is a model trained based on the house types actually selected by historical users and the style preference information of historical users; determining the external contour information of the virtual space based on the determined room type corresponding to the user, the number, layout and area of ​​each room type.

[0011] Optionally, the style information of each independent space in the virtual space is determined according to the user's style preference information and external contour information, including: inputting the style preference information and the external contour information into a space decoration model, converting the room features in the style preference information into a first feature vector through the space decoration model, converting the room features in the external contour information into a second feature vector, and determining the style information of each independent space in the virtual space based on the first feature vector and the second feature vector, wherein the training data of the space decoration model includes: the house type actually selected by the historical user, the style preference information of the historical user, and the actual decoration information of the historical user for each room.

[0012] Optionally, the training process of the room generation model includes the following steps: extracting features of the style preference information of historical users through a first encoder to obtain preference features of the historical users regarding the spatial environment; inputting the preference features of the historical users regarding the spatial environment into a first decoder, and predicting and generating a house type that meets the preferences of the historical users based on the preference features through the first decoder; iteratively training the first encoder and the first decoder based on the difference information between the house type predicted and generated by the first decoder and the house type actually selected by the historical users until the first encoder and the first decoder both enter a convergence state, wherein after receiving the preference features output by the first encoder in the convergence state, the first decoder in the convergence state predicts and generates a house type that has a similarity greater than a preset threshold with the house type actually selected by the historical users based on the preference features; and the first encoder and the first decoder that enter an iterative state form a room generation model.

[0013] Optionally, the training process of the space decoration model includes the following steps: extracting features of style preference information of historical users through a first encoder constituting a room generation model to obtain preference features of historical users regarding the space environment output by the room generation model, wherein the room generation model is used to determine the external contour information of the virtual space based on the style preference information; extracting room features of the house type actually selected by the historical users through a second encoder, wherein the second encoder is an encoder independent of the first encoder; predicting and generating the decoration information of each room by the historical users through a second decoder based on the preference features output by the room generation model and the room features output by the second encoder; iteratively training the second decoder, the second encoder and the first encoder of the room generation model according to the decoration information of each room output by the second decoder and the actual decoration information of each room by the historical users, until the second decoder predicts the similarity between the decoration information of each room output by the second decoder and the actual decoration information after receiving the preference features output by the first encoder and the room features output by the second encoder, thereby terminating the training; and forming the space decoration model with the first encoder, the second encoder and the second decoder that have completed the last training.

[0014] Optionally, the style information of each independent space in the virtual space is determined based on the user's style preference information and external contour information, including: obtaining item introduction information of items sold on the network; determining feature tags corresponding to the items sold on the network based on the item introduction information, wherein the feature tags are used to describe the decorative effect of the items in the space; determining the style information of each independent space in the virtual space based on the user's style preference information, external contour information and feature tags corresponding to the items.

[0015] According to another aspect of the present application, a device for generating a virtual space based on user preference information is also provided, which includes: an acquisition unit for acquiring the user's style preference information; a first determination unit for determining the external contour information of the virtual space based on the style preference information, wherein the external contour information includes at least the apartment type information of the virtual space; a second determination unit for determining the style information of each independent space in the virtual space based on the user's style preference information and the external contour information, wherein the style information of each independent space includes at least the type information of built-in items in each independent space and the placement position information of each built-in item; a virtual space generation unit for generating a target virtual space corresponding to the user based on the external contour information and the style information of each independent space.

[0016] According to another aspect of the present application, a computer-readable storage medium is also provided, wherein the computer-readable storage medium includes a stored executable program, wherein when the executable program is running, the device where the computer-readable storage medium is located is controlled to execute the above-mentioned method of generating a virtual space based on user preference information.

[0017] According to another aspect of the present application, an electronic device is provided, which includes: a memory storing an executable program; and a processor for running the program, wherein the program executes the above-mentioned method of generating a virtual space based on user preference information when running.

[0018] According to another aspect of the present application, a computer program product is also provided, including computer instructions, which implement the above-mentioned method of generating a virtual space based on user preference information when executed by a processor.

[0019] In the present application, the user's style preference information is first obtained, and then the external contour information of the virtual space is determined based on the style preference information, wherein the external contour information at least includes the apartment type information of the virtual space. Subsequently, the style information of each independent space in the virtual space is determined based on the user's style preference information and the external contour information, wherein the style information of each independent space at least includes the items required in each independent space and the placement position information of the items. Finally, a target virtual space corresponding to the user is generated based on the external contour information and the style information of each independent space.

[0020] From the above content, it can be seen that based on the user's style preferences, this application can intelligently determine the external contour information of the virtual space, including the apartment area, type, and room allocation, etc. This step takes into account the user's spatial needs and aesthetic tendencies. Secondly, after determining the external contour information, this application further analyzes the style information of each independent space (such as living room, bedroom, etc.) to ensure the consistency of the internal style of the space. Through the pre-trained style evaluation model, the style preference of the initially generated independent space is comprehensively evaluated, and the style elements are adjusted to maintain the coordination and unity of the overall style. At the same time, network resources are used to automatically select and match items that meet style preferences for each independent space, realize the automation of space layout, and greatly improve the layout efficiency.

[0021] In summary, the technical solution of this application solves the problems of low efficiency and inconsistent styles in the generation of personalized virtual spaces in the prior art by comprehensively acquiring the user's style preference information, intelligently determining the external contour of the virtual space, coordinating and unifying the space style, and automatically matching items that match the style preferences. This method can not only quickly generate personalized virtual spaces, but also ensure the uniformity and aesthetics of the space style, greatly improving the user experience and meeting the user's needs for efficient generation and layout of personalized virtual spaces. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation on the present application. In the drawings:

[0023] Figure 1 A hardware structure block diagram of a computer terminal for implementing a method for generating a virtual space based on user preference information is shown;

[0024] Figure 2 is a flowchart of an optional method for generating a virtual space based on user preference information according to an embodiment of the present application;

[0025] Figure 3 is an architectural diagram of an optional room generation model according to an embodiment of the present application;

[0026] Figure 4 is a structural diagram of an optional space decoration model according to an embodiment of the present application;

[0027] Figure 5 is a schematic diagram of an optional device for generating a virtual space based on user preference information according to an embodiment of the present application;

[0028] Figure 6 It is a structural block diagram of an electronic device according to an embodiment of the present application. DETAILED DESCRIPTION

[0029] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present application.

[0030] It should be noted that the terms "first", "second", etc. in the specification and claims of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] It should also be noted that the collected information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for display, data for analysis, etc.) involved in this application are information and data authorized by the user or fully authorized by all parties, and the collection, storage, use, processing, transmission, provision, disclosure and application of relevant data are in compliance with relevant laws, regulations and standards, necessary confidentiality measures are taken, and public order and good customs are not violated, and corresponding operation entrances are provided for users to choose to authorize or refuse. For example, an interface is set up between this system and relevant users or institutions to provide users with corresponding operation entrances for users to choose to agree or refuse the results of automated decision-making; if the user chooses to refuse, the expert decision-making process will be entered.

[0032] According to an embodiment of the present application, an embodiment of a method for generating a virtual space based on user preference information is provided. It should be noted that the steps shown in the flowchart in the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0033] The method embodiments provided in the embodiments of the present application can be executed in a mobile terminal, a computer terminal or a similar computing device. Figure 1 The hardware structure block diagram of a computer terminal (or mobile device) for implementing a method for generating a virtual space based on user preference information is shown. Figure 1 As shown, the computer terminal 10 (or mobile device) may include one or more (102a, 102b, ..., 102n are used to illustrate) processors 102 (the processor 102 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), a memory 104 for storing data, and a transmission device 106 for communication functions. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the BUS bus), a network interface, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 1 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 1 More or fewer components as shown, or with Figure 1 Different configurations shown.

[0034] It should be noted that the above-mentioned one or more processors 102 and / or other data processing circuits may generally be referred to as "data processing circuits" in this article. The data processing circuit may be embodied in whole or in part as software, hardware, firmware or any other combination. In addition, the data processing circuit may be a single independent processing module, or may be fully or partially integrated into any of the other components in the computer terminal 10 (or mobile device). As in the method for generating a virtual space based on user preference information involved in the embodiments of the present application, the data processing circuit acts as a processor control (e.g., selection of a variable resistor terminal path connected to an interface).

[0035] The memory 104 can be used to store software programs and modules of application software, such as the program instructions / data storage device corresponding to the method for generating a virtual space based on user preference information in the embodiment of the present application. The processor 102 executes various functional applications and data processing by running the software programs and modules stored in the memory 104, that is, the method for generating a virtual space based on user preference information is realized. The memory 104 may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory 104 may further include a memory remotely arranged relative to the processor 102, and these remote memories may be connected to the computer terminal 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0036] The transmission device 106 is used to receive or send data via a network. The specific example of the above network may include a wireless network provided by a communication provider of the computer terminal 10. In one example, the transmission device 106 includes a network adapter (Network Interface Controller, NIC), which can be connected to other network devices through a base station so as to communicate with the Internet. In one example, the transmission device 106 can be a radio frequency (RF) module, which is used to communicate with the Internet wirelessly.

[0037] The display may be, for example, a touch screen liquid crystal display (LCD) that enables a user to interact with a user interface of the computer terminal 10 (or mobile device).

[0038] Under the above operating environment, this application provides Figure 2 The method of generating a virtual space based on user preference information is shown. Figure 2 FIG. 1 is a flow chart of an optional method for generating a virtual space based on user preference information according to an embodiment of the present application. Figure 2 As shown, the method comprises the following steps:

[0039] Step S201, obtaining the user's style preference information.

[0040] Optionally, in order to obtain the user's style preference information, the system can first analyze the virtual human image created or selected by the user, and extract visual elements such as clothing, physical features, etc. For example, if a user's virtual human wears traditional clothing, it may imply that the user prefers traditional or ethnic style; and if the virtual human image tends to be futuristic or technological, then the user's style preference may tend to be modern or science fiction style. These elements will become an important reference for generating a virtual space that suits the user's style.

[0041] Secondly, users can upload photos that reflect their personal preferences and real living environment. The system analyzes these photos and identifies the characteristics of the space environment style, color matching, etc. For example, from the family photos uploaded by users, the system can identify the user's preference for natural wooden furniture, green plant decorations and other elements, and then infer his or her inclination towards environmental protection and natural style. This information can be used to generate a virtual space that is closer to the user's actual life.

[0042] Optionally, the system can also show the user a series of images of different styles, from which the user can select the style image they like. Based on the user's selection, the system can identify the user's preferred style elements such as color, pattern, layout, etc., and these data will be used in the subsequent generation of virtual space to ensure that the space style matches the user's aesthetic preferences.

[0043] In addition to direct style preference data, the system also indirectly evaluates users' style preferences by analyzing their comments and operational behaviors on the platform. For example, if users frequently discuss environmental issues on social platforms, or often browse and purchase furniture made of environmentally friendly materials in virtual spaces, the system can determine that users have a high preference for environmental protection and sustainability. This analysis based on user behavior can capture style tendencies that users may not even be aware of, thereby generating a virtual space that better meets the user's potential needs.

[0044] The system processes the style preference information collected above and adds corresponding labels and indicators to the items generated or selected for each style. For example, a piece of furniture may be labeled as "modern simplicity", "environmentally friendly materials", "Nordic style", etc. These labels and indicators will serve as the basis for subsequent matching and generation of virtual space layout.

[0045] Through the above-mentioned multi-angle and multi-level style preference information acquisition and analysis, the present application can more accurately understand the user's personalized needs for virtual space style, provide users with fast and accurate virtual space generation services, avoid the tedious process of users needing to repeatedly modify and adjust the virtual space in the prior art, and significantly improve the generation efficiency and satisfaction of personalized virtual space.

[0046] Step S202: determining the external contour information of the virtual space according to the style preference information, wherein the external contour information at least includes the apartment type information of the virtual space.

[0047] Optionally, first, the system establishes a mapping rule based on the style preference information obtained from the user (such as preference for nature, love of food, pursuit of simplicity, etc.), and associates these preferences with specific apartment designs and spatial layouts. For example, if the system recognizes that the user has a strong interest in cooking, it will tend to generate a virtual space with a larger kitchen area and a layout that focuses more on practicality and operational processes; while for users who prefer a minimalist style, the system may design an apartment with a wider space and fewer partitions and decorative elements. After clarifying the association between style preferences and apartment designs, the system can intelligently generate external contour information of the virtual space. This includes determining the total building area of ​​the virtual space, the number of rooms, the size and shape of each room, and the relative position and connection method between the rooms. For example, for users who pursue a minimalist style, the system may reduce the number of rooms, increase the size of the main living space, optimize the spatial flow, and reduce unnecessary aisles or partitions, thereby improving the overall sense and efficiency of the space.

[0048] Optionally, in addition to the apartment type information, the external outline also involves details of the space layout. The system will adjust the layout of each independent space and the division of functional areas according to the user's style preferences. For example, for users who like natural style, the system may add more natural light introduction, indoor green space, and open balcony or terrace areas to the design to enhance the sense of connection with nature. In addition, each apartment type or space layout comes with parameters related to style preferences. The system will use these parameters to ensure that the generated external outline accurately reflects the user's style preferences. For example, the "environmental protection" style parameters may affect the selection of materials and energy use; the "foodie" style parameters may affect the size and configuration of the kitchen.

[0049] Through the above steps, the technical solution of this application can transform abstract style preferences into specific virtual space design elements, such as apartment size, shape, layout, etc., thereby automatically generating a virtual space external outline that is highly matched with the user's style preferences. This method not only greatly improves the efficiency of generating personalized virtual spaces, but also ensures that the design results can better meet the personalized needs of users, reduces the need for subsequent users to modify the layout, and improves user experience and satisfaction.

[0050] Step S203, determining the style information of each independent space in the virtual space according to the user's style preference information and the external contour information, wherein the style information of each independent space at least includes the type information of the built-in items in each independent space and the placement position information of each built-in item.

[0051] Optionally, after determining the outer contour of the virtual space, the system further refines the user's style preference information into each independent space. Independent spaces include but are not limited to living rooms, bedrooms, kitchens, bathrooms, etc. The system will consider the functional requirements of each independent space and the user's specific preferences for the space. For example, if the user prefers nature and leisure, the style information of the living room may include furniture made of natural materials, comfortable leisure area layout, etc.

[0052] Next, the system will select items that match the style from a database containing a large number of items based on the style information of each independent space. These items include not only furniture, but also electrical appliances, decorations, lighting equipment, etc. Each item is given a style label in advance (such as simple, luxurious, traditional, modern, etc.) to facilitate the system to make accurate matches. For example, for the kitchen space, if the user's style preference is modern and simple, the system will select kitchen appliances and furniture with simple lines and low-key colors.

[0053] Optionally, the system not only determines the items in each independent space, but also intelligently plans the placement of the items to ensure that the layout meets both style requirements and space utilization efficiency and the rationality of human-computer interaction. This process may involve the use of machine learning algorithms, such as scene understanding models based on deep learning, to predict the best layout of items in space. For example, considering the cooking process, the stove, sink, and refrigerator in the kitchen will be laid out according to the "work triangle" principle to improve the convenience and efficiency of kitchen use.

[0054] After initially determining the items and layout of each independent space, the system will also coordinate and optimize the overall style. This is because, although each independent space may satisfy the user's style preferences, the entire virtual space needs to maintain consistency and harmony in style. By inputting the preliminarily arranged independent space images into the pre-trained style evaluation model, the system can identify and adjust those areas that are stylistically inconsistent. For example, if the style of the living room is modern and simple, and the adjacent restaurant is too traditional, the system will adjust the layout of the restaurant to make its style more consistent with the living room.

[0055] Finally, the system will present the preliminary generated virtual space layout to the user, who can give feedback and choose to replace some items recommended by the system or adjust their placement. Based on the user's feedback, the system will provide multiple items with similar style tags for the user to choose from, allowing the user to easily customize the virtual space until he is completely satisfied.

[0056] Step S204: generating a target virtual space corresponding to the user according to the external contour information and the style information of each independent space.

[0057] From the above content, it can be seen that the technical solution of this application realizes the automation of the whole process from user style preference to virtual space external outline design, and then to internal layout and item selection. The virtual space finally generated is not only highly consistent with the user's preference in visual style, but also takes into account the functionality and practicality of the space, greatly improving the user's immersion and comfort in the virtual environment. This method not only solves the problem of low efficiency in generating personalized virtual space in the prior art, but also provides a more personalized and coordinated virtual space design, satisfying the user's pursuit of an ideal living environment.

[0058] Based on the contents of the above steps S201 to S204, it can be known that in the present application, the user's style preference information is first obtained, and then the external contour information of the virtual space is determined based on the style preference information, wherein the external contour information at least includes the apartment type information of the virtual space. Subsequently, the style information of each independent space in the virtual space is determined based on the user's style preference information and the external contour information, wherein the style information of each independent space at least includes the items required in each independent space and the placement position information of the items. Finally, a target virtual space corresponding to the user is generated based on the external contour information and the style information of each independent space.

[0059] It should be noted that, based on the user's style preferences, this application can intelligently determine the external contour information of the virtual space, including the apartment area, type, and room allocation, etc. This step takes into account the user's spatial needs and aesthetic tendencies. Secondly, after determining the external contour information, this application further analyzes the style information of each independent space (such as living room, bedroom, etc.) to ensure the consistency of the internal style of the space. Through the pre-trained style evaluation model, the style preference of the initially generated independent space is comprehensively evaluated, and the style elements are adjusted to maintain the coordination and unity of the overall style. At the same time, network resources are used to automatically select and match items that meet style preferences for each independent space, realize the automation of space layout, and greatly improve the layout efficiency.

[0060] In an optional embodiment, in order to obtain the user's style preference information, the system may obtain a target image uploaded by the user, wherein the target image includes a spatial environment provided by the user for reference. The system may also extract spatial environment style features from the target image, and then use the extracted spatial environment style features as the user's style preference information.

[0061] Optionally, the system allows users to upload images of space environments that they think match their personal aesthetic and style preferences. These target images can be inspiration images found by users from social media, home magazines, online platforms, etc., or images of the user's own home, with the purpose of providing the system with intuitive visual references to help the system understand the user's style preferences. The uploaded target images are then analyzed by the system to extract the style features of the space environment. This process may use computer vision technology, especially convolutional neural networks (CNNs) in deep learning, to identify elements in the image, such as color, texture, shape, layout, furniture type, and decorations. These elements will be further classified and quantified to form style features. For example, if the uploaded image uses a lot of wood materials, has plenty of natural light, tends to be warm in color, and has an open layout, the system may extract style features such as "natural warmth", "modern simplicity", or "Nordic style".

[0062] Optionally, the spatial environment style features that the system can extract from the target image will be integrated by the system to form a database of user style preference information. This information will be quantified and labeled, so that the system can match and adjust it when generating the virtual space according to the user's style preferences. For example, the system will label different style features with labels such as "environmental protection", "simple", "luxury", etc., so that when selecting furniture and decorations in the virtual space, options that meet the user's style preferences can be quickly located. Based on the style preference information extracted from the target image, the system begins to generate or adjust the design of the virtual space. This includes selecting items that match the user's style preferences, determining the placement of items, and optimizing the space layout and design details, such as color matching, material selection, etc., to ensure that every element of the virtual space can reflect the user's style preferences.

[0063] It should be noted that the above process uses image recognition and machine learning technology. By deeply analyzing the spatial environment pictures provided by users, the system can accurately capture the user's specific style preferences, so as to make more personalized choices when generating virtual space, and finally present a virtual environment that meets both the user's aesthetic and functional needs. This method not only simplifies the process of users expressing their style preferences, but also greatly improves the efficiency and accuracy of virtual space generation.

[0064] In an optional embodiment, in order to obtain the user's style preference information, the system can also obtain a target file uploaded by the user, wherein the target file is used to record the user's description information of the desired spatial environment. Then, the system extracts text description features from the target file, wherein the text description features include at least first-class description features and second-class description features, wherein the first-class description features represent the items that the user expects to appear in the spatial environment and the placement of the items, and the second-class description features represent the items that the user expects to be prohibited from appearing in the spatial environment. Finally, the system uses the text description features as the user's style preference information.

[0065] Optionally, the user can upload a text file that describes in detail the user's desired spatial environment characteristics. This can be a written description of the user's ideal spatial environment, which may include what furniture and decorations the user hopes to include in the space, as well as a description of the specific placement of these items in the room. At the same time, the user can also clearly indicate elements that should be avoided in the design, such as specific types of furniture or decorative styles. Secondly, the system uses natural language processing (NLP) technology to extract text description features from the target file uploaded by the user. NLP technology can recognize and understand the semantics in the text, and distinguish between positive desired descriptions (first-class descriptive features) and negative prohibited descriptions (second-class descriptive features).

[0066] The first type of descriptive features (expected items and locations): For example, a user may write "I hope to have a large bookshelf on the north wall of the study, with a comfortable reading chair next to it." The system will recognize information such as people (bookshelf, reading chair), location (north wall of the study), and form features of the expected items and their placement.

[0067] The second type of descriptive features (forbidden items): such as "I absolutely do not want to see any TV cabinets in my living room, I want it to be simple." Here, the system will identify "TV cabinet" as an item to be avoided, and "simple" as the desired overall style preference.

[0068] In addition, the system integrates the extracted first and second category descriptive features into a complete set of style preference information. This information not only includes the furniture and decorations that users want to see, but also covers their layout and position in the space, as well as the style elements that users do not want to appear, providing specific and clear guidance for the design of the virtual space. Based on the text description features extracted from the target file, the system begins to generate or adjust the design of the virtual space. The system will give priority to the first category of descriptive features to ensure that the items expected by the user appear in the specified location and follow the described layout. At the same time, the system will avoid the items and styles mentioned in the second category of descriptive features to ensure that the final virtual space environment not only meets the user's aesthetic preferences, but also avoids elements that the user does not want to see.

[0069] From the above content, we can see that by allowing users to directly express their spatial environment vision through text, the system can understand the user's style preferences more accurately and avoid the possible understanding bias of image analysis. This method not only provides users with a direct channel to express their personal needs, but also ensures the personalization and accuracy of the virtual space generation process, thereby improving user satisfaction and the efficiency of virtual space design.

[0070] In an optional embodiment, in the process of determining the external contour information of the virtual space based on the style preference information, the system can input the style preference information into the room generation model, convert the room features in the style preference information into feature vectors through the room generation model, and use the feature vectors to determine the room type corresponding to the user, the number, layout and area of ​​each room type, wherein the room generation model is a model trained based on the house type actually selected by historical users and the style preference information of historical users. Then, the system can determine the external contour information of the virtual space based on the determined room type corresponding to the user, the number, layout and area of ​​each room type.

[0071] Optionally, the system first receives the user's style preference information, which may include images uploaded by the user, text descriptions, and preference tags directly provided by the user. In order to make this information understandable and processed by the room generation model, the system uses natural language processing (NLP) and image recognition technology (such as deep learning convolutional neural network, CNN) to convert the style preference information into a set of feature vectors. Feature vectors are a numerical representation that can accurately capture the key points in the preference information, such as the user's preferred color, material, space layout, furniture type, etc.

[0072] Optionally, the above-mentioned room generation model is trained by analyzing a large amount of historical data on users' actual choice of house type and style preferences. These historical data include the house type (such as apartment, villa, duplex, etc.) finally selected by the user, the number and layout of rooms, and the style preference information related to these choices. By learning these data, the model can establish a mapping relationship between the style preference feature vector and the room design parameters. After inputting the user's style preference feature vector, the room generation model can predict the parameters such as room type, number, layout and area that are most suitable for the user based on the learned mapping relationship. For example, if the user prefers an open modern style, the model may generate external contour information with a large living room, an open kitchen and fewer partitions; if the user prefers a traditional style and privacy, the model may generate external contour information with more independent rooms and a traditional layout.

[0073] Optionally, once the room generation model outputs room type information that matches the user's style preferences, the system will further determine the external outline of the virtual space based on this information. This includes the size of the specific room, the relative position between the rooms (i.e., layout), corridor design, door and window positions, and space streamlines, etc., to ensure that the virtual space not only meets the user's aesthetic preferences in appearance, but also meets the user's needs in terms of functionality, such as appropriate privacy protection, sufficient living space, and efficient traffic planning.

[0074] Through the above process, the system can automatically convert the user's abstract style preferences into specific parameters in virtual space design, such as room type, quantity, layout and area, thereby generating a virtual space outline that is both personalized and practical. This method not only greatly reduces the user's workload in space design, but also ensures that the generated virtual space is highly matched with the user's preferences, improving the user experience satisfaction and immersion. At the same time, the room generation model trained based on historical user data can continuously learn and optimize. As user data accumulates, the virtual space it generates will be more in line with user expectations, further improving the accuracy and innovation of the design.

[0075] In an optional embodiment, in order to determine the style information of each independent space in the virtual space based on the user's style preference information and external contour information, the system can input the style preference information and external contour information into the space decoration model, and convert the room features in the style preference information into a first feature vector and the room features in the external contour information into a second feature vector through the space decoration model, and determine the style information of each independent space in the virtual space based on the first feature vector and the second feature vector, wherein the training data of the space decoration model includes: the house type actually selected by the historical user, the style preference information of the historical user, and the actual decoration information of the historical user for each room.

[0076] Optionally, the system processes the received user style preference information and virtual space external contour information separately and converts them into numerical feature vectors. Among them, the first feature vector (style preference information vector) contains the user's aesthetic tendencies, color preferences, furniture preferences, etc., while the second feature vector (external contour information vector) contains specific design parameters such as room type, quantity, size, layout, etc. This vectorization processing is a necessary step for machine learning and deep learning models to understand and operate data.

[0077] Optionally, the above-mentioned space decoration model is trained on a large amount of historical user data, which includes the house type actually selected by the user, style preference information, and their actual decoration information for each room. For example, the model may learn that when the user prefers "Nordic style" and selects "open floor plan", the living room and dining room are usually decorated with light colors, wooden furniture and simple lines. Through deep learning algorithms, the space decoration model can understand the complex relationship between style preferences and external contour design, and how to apply these preferences specifically to the design of each independent space.

[0078] Optionally, the system can input the first feature vector (user style preference) and the second feature vector (external contour information) into the space decoration model, and the model can comprehensively consider these two sets of information to predict the best style decoration plan for each independent space in the virtual space. This plan not only includes the types of items that should be included in the room, but also the placement of items, color matching, material selection and other specific details to ensure that the design of each independent space can fully reflect the user's style preferences while being consistent and coordinated with the external contour design.

[0079] In addition, for each independent space (such as living room, bedroom, kitchen, etc.), the space decoration model will generate a detailed design plan, including but not limited to the style, size, color, material, location, etc. of furniture and decorations. For example, for users who prefer "modern minimalist style", the model may recommend using a minimalist sofa in the living room, matching it with artistic decorative paintings, and designing the TV wall to be hidden to keep the space clean and open.

[0080] Through the above content, the system can automatically and accurately apply the user's style preferences to the virtual space design, which not only improves the personalization of the design, but also ensures the practicality and aesthetics of the design. As the core of this process, the space dressing model can continuously optimize its prediction ability by learning the actual choices and preferences of historical users, and provide users with virtual space designs that better meet their expectations. This space design method based on user data and deep learning not only improves the efficiency of design, but also enhances the satisfaction and immersion of user experience.

[0081] In an optional embodiment, the training process of the room generation model includes the following steps:

[0082] The style preference information of historical users is feature extracted by the first encoder to obtain the preference features of the historical users regarding the spatial environment; the preference features of the historical users regarding the spatial environment are input into the first decoder, and the first decoder predicts and generates a house type that satisfies the preferences of the historical users based on the preference features; the first encoder and the first decoder are iteratively trained based on the difference information between the house type predicted and generated by the first decoder and the house type actually selected by the historical users until both the first encoder and the first decoder enter a convergence state, wherein after receiving the preference features output by the first encoder in the convergence state, the first decoder in the convergence state predicts and generates a house type based on the preference features with a similarity greater than a preset threshold with the house type actually selected by the historical users; the first encoder and the first decoder that enter the iterative state form a room generation model.

[0083] Optionally, Figure 3 is an architectural diagram of an optional room generation model according to an embodiment of the present application, such as Figure 3 As shown, the room generation model includes a first encoder and a first decoder.

[0084] During the training process of the room generation model, the first encoder is used to process the historical user style preference information. This encoder is a deep learning model that may be based on a recurrent neural network (RNN), a long short-term memory network (LSTM), or a self-attention mechanism (such as Transformer) to extract core preference features from multiple sources of information such as user-provided descriptions, pictures, and selected historical records. These features are digital representations of user style preferences that can capture the user's aesthetic tendencies, functional requirements, and personalized preferences for the spatial environment.

[0085] The preference features extracted from the first encoder are then input into the first decoder, which is also a deep learning model whose goal is to predict and generate a house type that meets the user's preferences based on these features. The decoder uses the mapping relationship between the learned preference features and house types to generate a series of possible house design solutions, including room layout, size, type, etc.

[0086] By comparing the difference information between the house type predicted by the first decoder and the house type actually selected by historical users, the system can calculate an indicator of prediction accuracy or similarity. If the similarity between the predicted result and the actual selection is lower than the preset threshold, the system will adjust the parameters of the first encoder and the first decoder, and then predict again until the similarity index reaches or exceeds the threshold, indicating that the accuracy and stability of the model prediction have reached an acceptable level, that is, it enters the "convergence state". This process is called iterative training, and its goal is to continuously optimize the model performance so that the model can more accurately predict the house type that users may prefer from their style preference information.

[0087] Finally, once the first encoder and the first decoder reach convergence, that is, they are able to stably extract preference features and predict house types, the two models can be combined into a complete room generation model. In the combined application, when a new user uploads style preference information, the first encoder will extract the preference features, which are then used by the first decoder to predict and generate a house type that meets the user's preferences, including specific room layouts, number of rooms, area and other design details.

[0088] Through this iterative training method based on deep learning, the room generation model can be continuously optimized and learn the complex relationship between user style preferences and house types, so that when facing new users, it can more accurately predict and generate house design solutions that meet their style preferences. This method not only improves the personalization and customization capabilities of virtual space design, but also greatly improves design efficiency, providing users with a fast and accurate virtual space generation experience.

[0089] In an optional embodiment, the training process of the space decoration model includes the following steps:

[0090] The style preference information of historical users is feature extracted by a first encoder constituting a room generation model, and the preference features of historical users regarding the spatial environment are obtained by the room generation model, wherein the room generation model is used to determine the external contour information of the virtual space based on the style preference information; the room features of the house type actually selected by the historical users are extracted by a second encoder, wherein the second encoder is an encoder independent of the first encoder; the dressing information of each room by the historical users is predicted and generated by a second decoder based on the preference features output by the room generation model and the room features output by the second encoder; the second decoder, the second encoder and the first encoder of the room generation model are iteratively trained according to the dressing information of each room output by the second decoder and the actual dressing information of each room by the historical users, until the second decoder predicts that the dressing information output by the second decoder is greater than a preset similarity with the actual dressing information after receiving the preference features output by the first encoder and the room features output by the second encoder, and the training is terminated; the first encoder, the second encoder and the second decoder that have completed the last training form a space dressing model.

[0091] Optionally, Figure 4 is an architecture diagram of an optional space decoration model according to an embodiment of the present application, such as Figure 4 As shown, the space decoration model includes a first encoder, a second encoder and a second decoder.

[0092] Optionally, the first encoder can also process historical user style preference information, which may include pictures uploaded by users, text descriptions, or selected style tags, etc. The first encoder converts these style preference information into preference feature vectors, which contain user preference information on color, layout, furniture style, etc., and are used to guide the design of virtual space.

[0093] Optionally, the second encoder is independent of the first encoder, and is used to analyze the house types actually selected by historical users and extract room feature vectors of each room, such as room size, shape, location, functional requirements, etc. The room feature vector provides information about the actual house structure, which, combined with the preference feature vector, can provide a basis for subsequent virtual dressing.

[0094] Optionally, the second decoder receives the preference feature vector output from the first encoder and the room feature vector output from the second encoder as input. The second decoder attempts to predict the decoration information that the historical user may choose for each room based on the input information, including furniture type, decoration style, color matching, layout, etc.

[0095] Optionally, the dressing information of each room predicted by the second decoder is compared with the actual dressing information of the historical user to calculate the similarity index. If the similarity between the predicted dressing information and the actual information is lower than a preset threshold, the system will adjust the parameters of the second decoder, the second encoder and the first encoder to perform iterative training. The training process continues until the prediction result of the second decoder is very close to the actual choice of the historical user, that is, the similarity exceeds the preset threshold, at which time the model reaches a convergence state and the training ends. After the last iterative training is completed, the first encoder, the second encoder and the second decoder are combined into a space dressing model. This model has the ability to predict dressing information based on user style preferences and specific room features, and is used to generate personalized and coordinated virtual space designs.

[0096] From the above content, we can see that by integrating multiple deep learning models (encoders and decoders), the space decoration model can fully understand user needs and make creative designs based on these needs. The model not only takes into account user style preferences, but also combines the actual layout and structure of the house to ensure that the virtual space not only meets the user's aesthetics, but also meets their actual usage needs. Through iterative training, the model can continuously learn and adjust to improve the accuracy of predictions and the level of personalization of designs. Through deep learning and comprehensive analysis of user style preferences and actual house characteristics, a highly personalized and coordinated virtual space design is achieved, greatly improving user experience and design efficiency.

[0097] In an optional embodiment, the system can obtain the item introduction information of the items sold on the network, and then determine the feature tags corresponding to the items sold on the network based on the item introduction information, wherein the feature tags are used to describe the decorative function of the items in the space. Finally, the system determines the style information of each independent space in the virtual space based on the user's style preference information, external contour information, and the feature tags corresponding to the items.

[0098] Optionally, the system obtains item description information of items currently on sale from the Internet. This information may include product titles, descriptions, user reviews, keywords provided by merchants, etc. The system uses natural language processing (NLP) technologies, such as text mining and semantic analysis, to deeply interpret this information, identify and extract key information that can describe the function, style, material, color and other characteristics of the item, and then convert this information into feature tags. Feature tags are digital representations of item attributes, which facilitate system understanding and operation. For example, tags such as "retro style", "solid wood material", "large storage space", and "suitable for small apartments" describe the aesthetic style, physical properties, and applicable scenarios of items.

[0099] Optionally, after obtaining the feature tags of the items, the system will match these feature tags with the user's desired design scheme based on the user's style preference information and the external contour information of the virtual space. The user's style preference information, such as "modern simplicity", "Nordic style", "dark colors", etc., guides the system in selecting items. External contour information, including the number, size, and layout of rooms, limits the selection range of items to ensure that the size and function of the items are suitable for a specific room. The system compares the feature tags of the items with the user's preferences and the actual needs of the room to select the most suitable items for the decoration of the virtual space.

[0100] Finally, the system will consider the selected item feature tags, the user's style preference information, and the external contour information of the virtual space to determine the style information of each independent space (such as living room, bedroom, kitchen, etc.). This process may involve multiple aspects of decision-making, such as whether the style of the item is consistent with the user's preference, whether the size of the item matches the size of the space, whether the color of the item is coordinated with the overall tone of the room, and whether the functionality of the item meets the user's usage needs. Based on these comprehensive factors, the system will design a complete style dressing plan for each independent space, including the selection and layout of furniture and decorations, to ensure that the virtual space not only meets the user's aesthetic preferences, but also has practicality and comfort.

[0101] It should be noted that through the above content, the system can intelligently filter out the items that best suit the user's style preferences and virtual space needs from the vast amount of items on the Internet, and then design a coordinated and personalized virtual space style. This method not only improves the efficiency and accuracy of virtual space design, but also provides users with a more convenient and personalized space environment design experience. Through automatic feature extraction and matching of deep learning, the system can understand complex user needs and item attributes, thereby making smarter and more user-friendly design decisions.

[0102] According to an embodiment of the present application, a device for generating a virtual space based on user preference information is also provided, wherein: Figure 5 is a schematic diagram of an optional device for generating a virtual space based on user preference information according to an embodiment of the present application, such as Figure 5 As shown, the apparatus for generating a virtual space based on user preference information includes: an acquisition unit 501 , a first determination unit 502 , a second determination unit 503 , and a virtual space generation unit 504 .

[0103] Optionally, an acquisition unit 501 is used to acquire style preference information of a user; a first determination unit 502 is used to determine external contour information of a virtual space based on the style preference information, wherein the external contour information includes at least apartment type information of the virtual space; a second determination unit 503 is used to determine style information of each independent space in the virtual space based on the style preference information and the external contour information of the user, wherein the style information of each independent space includes at least type information of built-in items in each independent space and placement position information of each built-in item; a virtual space generation unit 504 generates a target virtual space corresponding to the user based on the external contour information and the style information of each independent space.

[0104] Optionally, the acquisition unit 501 includes: a first acquisition subunit, a feature extraction subunit, and a first processing subunit. The first acquisition subunit is used to acquire a target image uploaded by a user, wherein the target image includes a spatial environment provided by the user for reference; the feature extraction subunit is used to extract spatial environment style features from the target image; and the first processing subunit is used to use the extracted spatial environment style features as the user's style preference information.

[0105] Optionally, the acquisition unit 501 includes: a second acquisition subunit, a second processing subunit, and a third processing subunit. The second acquisition subunit is used to acquire a target file uploaded by a user, wherein the target file is used to record the user's description information of the desired spatial environment; the second processing subunit is used to extract text description features from the target file, wherein the text description features include at least a first type of description features and a second type of description features, wherein the first type of description features represent the items that the user expects to appear in the spatial environment and the placement of the items, and the second type of description features represent the items that the user expects to be prohibited from appearing in the spatial environment; and the third processing subunit is used to use the text description features as the user's style preference information.

[0106] Optionally, the first determination unit 502 includes: a first determination subunit and a second determination subunit. The first determination subunit is used to input the style preference information into the room generation model, convert the room features in the style preference information into feature vectors through the room generation model, and use the feature vectors to determine the room type corresponding to the user, the number, layout and area of ​​each room type, wherein the room generation model is a model trained based on the house type actually selected by historical users and the style preference information of historical users; the second determination subunit is used to determine the external contour information of the virtual space according to the determined room type corresponding to the user, the number, layout and area of ​​each room type.

[0107] Optionally, the second determination unit 503 includes: a third determination subunit, used to input the style preference information and the external contour information into the space decoration model, convert the room features in the style preference information into a first feature vector through the space decoration model, and convert the room features in the external contour information into a second feature vector, and determine the style information of each independent space in the virtual space based on the first feature vector and the second feature vector, wherein the training data of the space decoration model includes: the house type actually selected by the historical user, the style preference information of the historical user, and the actual decoration information of the historical user for each room.

[0108] Optionally, the device for generating a virtual space based on user preference information further includes: a first processing unit, a second processing unit, a third processing unit, and a fourth processing unit. The first processing unit is used to extract features of the style preference information of historical users through a first encoder to obtain the preference features of the historical users regarding the spatial environment; the second processing unit is used to input the preference features of the historical users regarding the spatial environment into the first decoder, and the first decoder is used to predict and generate a house type that meets the preferences of the historical users based on the preference features; the third processing unit is used to iteratively train the first encoder and the first decoder based on the difference information between the house type predicted and generated by the first decoder and the house type actually selected by the historical user, until both the first encoder and the first decoder enter a convergence state, wherein the first decoder in the convergence state, after receiving the preference features output by the first encoder in the convergence state, predicts and generates a house type based on the preference features and a similarity with the house type actually selected by the historical user greater than a preset threshold; the fourth processing unit is used to form a room generation model with the first encoder and the first decoder that enter an iterative state.

[0109] Optionally, the device for generating a virtual space based on user preference information also includes: a fifth processing unit, which is used to extract features of the style preference information of historical users through a first encoder constituting a room generation model, and obtain the preference features of the historical users about the spatial environment output by the room generation model, wherein the room generation model is used to determine the external contour information of the virtual space based on the style preference information; a sixth processing unit, which is used to extract room features of the house type actually selected by the historical users through a second encoder, wherein the second encoder is an encoder independent of the first encoder; a seventh processing unit, which is used to predict and generate the dressing information of each room of the historical users through a second decoder based on the preference features output by the room generation model and the room features output by the second encoder; an eighth processing unit, which is used to iteratively train the second decoder, the second encoder and the first encoder of the room generation model according to the dressing information of each room output by the second decoder and the actual dressing information of each room by the historical users, until the second decoder, after receiving the preference features output by the first encoder and the room features output by the second encoder, the similarity between the dressing information of each room predicted and output by the second decoder and the actual dressing information is greater than a preset similarity, and the training is terminated; a ninth processing unit, which is used to form a space dressing model with the first encoder, the second encoder and the second decoder that have completed the last training.

[0110] Optionally, the second determination unit 503 includes: a third acquisition subunit, a fourth determination subunit, and a fifth determination subunit. The third acquisition subunit is used to acquire the item introduction information of the items sold on the network; the fourth determination subunit is used to determine the feature tags corresponding to the items sold on the network according to the item introduction information, wherein the feature tags are used to describe the dressing function of the items in the space; the fifth determination subunit is used to determine the style information of each independent space in the virtual space according to the user's style preference information, the external contour information, and the feature tags corresponding to the items.

[0111] An embodiment of the present application may provide an electronic device, Figure 6 is a structural block diagram of an electronic device according to an embodiment of the present application. Figure 6 As shown, the electronic device may include: one or more ( Figure 6 Only one is shown) a processor, a memory, a storage controller, and a peripheral interface, wherein the peripheral interface is connected to a radio frequency module, an audio module, and a display.

[0112] Among them, the memory can be used to store software programs and modules, such as the program instructions / modules corresponding to the method and device for generating a virtual space based on user preference information in the embodiment of the present application. The processor executes various functional applications and data processing by running the software programs and modules stored in the memory, that is, realizing the above-mentioned method for generating a virtual space based on user preference information. The memory may include a high-speed random access memory, and may also include a non-volatile memory, such as one or more magnetic storage devices, flash memory, or other non-volatile solid-state memory. In some instances, the memory may further include a memory remotely arranged relative to the processor, and these remote memories may be connected to the terminal via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0113] The processor can call the information and application stored in the memory through the transmission device to perform the following steps: obtain the user's style preference information; determine the external contour information of the virtual space based on the style preference information, wherein the external contour information at least includes the apartment type information of the virtual space; determine the style information of each independent space in the virtual space based on the user's style preference information and the external contour information, wherein the style information of each independent space at least includes the items required in each independent space and the placement position information of the items; generate a target virtual space corresponding to the user based on the external contour information and the style information of each independent space.

[0114] It can be understood by those skilled in the art that Figure 6The structure shown is for illustration only, and the electronic device may also be a smart phone (such as an Android phone, an iOS phone, etc.), a tablet computer, a PDA, a mobile Internet device (Mobile Internet Devices, MID), a PAD, or other terminal devices. Figure 6 The structure of the electronic device is not limited. Figure 6 More or fewer components (such as network interfaces, display devices, etc.) shown in, or having Figure 6 Different configurations shown.

[0115] A person of ordinary skill in the art can understand that all or part of the steps in the various methods of the above embodiments can be completed by instructing the hardware related to the terminal device through a program, and the program can be stored in a computer-readable storage medium, and the storage medium may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, etc.

[0116] According to another aspect of the present application, a computer-readable storage medium is provided, wherein a computer program is stored in the computer-readable storage medium, wherein when the computer program is executed, the device where the computer-readable storage medium is located executes the above-mentioned method of generating a virtual space based on user preference information.

[0117] According to another aspect of the present application, a computer program product is also provided, wherein the computer program product includes computer instructions, wherein when the computer instructions are executed, the device where the computer program product is located executes the above-mentioned method of generating a virtual space based on user preference information.

[0118] The above-mentioned embodiments or examples disclosed in the present application are not exhaustive, but are only illustrative of some embodiments or examples, and are not intended to be specific limitations on the scope of protection disclosed in the present application. In the absence of contradiction, each step in a certain embodiment or example in the present application can be implemented as an independent example, and the steps can be combined arbitrarily. For example, the scheme after removing some steps in a certain embodiment or example can also be implemented as an independent example, and the order of the steps in a certain embodiment or example can be arbitrarily exchanged. In addition, the optional methods or optional examples in a certain embodiment or example can be combined arbitrarily; in addition, the various embodiments or examples can be combined arbitrarily, for example, some or all steps of different embodiments or examples can be combined arbitrarily, and a certain embodiment or example can be combined arbitrarily with the optional methods or optional examples of other embodiments or examples.

[0119] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0120] In the above embodiments of the present application, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, please refer to the relevant description of other embodiments.

[0121] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of the units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0122] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0123] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0124] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions to enable a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, disk or optical disk and other media that can store program codes.

[0125] The above is only a preferred implementation of the present application. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principles of the present application. These improvements and modifications should also be regarded as the scope of protection of the present application.

Claims

1. A method for generating a virtual space based on user preference information, characterized in that: include: Get the user's style preference information; Determining external contour information of the virtual space according to the style preference information, wherein the external contour information at least includes apartment type information of the virtual space; Determining style information of each independent space in the virtual space according to the user's style preference information and the external contour information, wherein the style information of each independent space at least includes type information of built-in items in each independent space and placement position information of each built-in item; A target virtual space corresponding to the user is generated according to the external contour information and the style information of each independent space.

2. The method for generating a virtual space based on user preference information according to claim 1, characterized in that: Get the user's style preference information, including: Acquire a target image uploaded by the user, wherein the target image includes a spatial environment provided by the user for reference; Extracting spatial environment style features from the target image; The extracted spatial environment style features are used as the style preference information of the user.

3. The method for generating a virtual space based on user preference information according to claim 1, characterized in that: Get the user's style preference information, including: Acquire a target file uploaded by the user, wherein the target file is used to record the user's description information of the desired spatial environment; Extracting text description features from the target file, wherein the text description features include at least a first type of description features and a second type of description features, the first type of description features characterizing the items that the user expects to appear in the spatial environment and the placement of the items, and the second type of description features characterizing the items that the user expects to be prohibited from appearing in the spatial environment; The text description features are used as the style preference information of the user.

4. The method for generating a virtual space based on user preference information according to claim 1, characterized in that: Determining the external contour information of the virtual space according to the style preference information includes: Inputting the style preference information into a room generation model, converting room features in the style preference information into feature vectors through the room generation model, and using the feature vectors to determine the room type corresponding to the user, the number, layout, and area of ​​each room type, wherein the room generation model is a model trained based on the house types actually selected by historical users and the style preference information of the historical users; The external contour information of the virtual space is determined according to the determined room type corresponding to the user, the quantity, layout and area of ​​each room type.

5. The method for generating a virtual space based on user preference information according to claim 1, characterized in that: Determining style information of each independent space in the virtual space according to the user's style preference information and the external contour information includes: The style preference information and the external contour information are input into a space decoration model, and the room features in the style preference information are converted into a first feature vector through the space decoration model, and the room features in the external contour information are converted into a second feature vector, and the style information of each independent space in the virtual space is determined according to the first feature vector and the second feature vector, wherein the training data of the space decoration model includes: the house type actually selected by the historical user, the style preference information of the historical user, and the actual decoration information of the historical user for each room.

6. The method for generating a virtual space based on user preference information according to claim 4, characterized in that: The training process of the room generation model includes the following steps: Extracting features of the style preference information of the historical user by a first encoder to obtain preference features of the historical user with respect to the spatial environment; Inputting the preference characteristics of the historical user regarding the spatial environment into a first decoder, and generating a house type that satisfies the preference of the historical user through prediction by the first decoder based on the preference characteristics; Iteratively training the first encoder and the first decoder according to difference information between the house type predicted by the first decoder and the house type actually selected by the historical user until the first encoder and the first decoder both enter a converged state, wherein after receiving the preference feature output by the first encoder in the converged state, the first decoder in the converged state predicts and generates a house type according to the preference feature, and the similarity between the house type actually selected by the historical user is greater than a preset threshold; The first encoder and the first decoder entering the iteration state constitute the room generation model.

7. The method for generating a virtual space based on user preference information according to claim 5, characterized in that: The training process of the space decoration model includes the following steps: Extracting features of the style preference information of the historical user through a first encoder constituting a room generation model, thereby obtaining preference features of the historical user regarding the space environment output by the room generation model, wherein the room generation model is used to determine the external contour information of the virtual space according to the style preference information; Extracting room features of the house type actually selected by the historical user through a second encoder, wherein the second encoder is an encoder independent of the first encoder; Predicting and generating dressing information of each room by the historical user according to the preference features output by the room generation model and the room features output by the second encoder by a second decoder; According to the dressing information of each room output by the second decoder and the actual dressing information of each room by the historical user, the second decoder, the second encoder and the first encoder of the room generation model are iteratively trained until the second decoder receives the preference features output by the first encoder and the room features output by the second encoder, and the similarity between the dressing information of each room predicted and output by the second decoder and the actual dressing information is greater than a preset similarity, and the training is terminated; The first encoder, the second encoder and the second decoder that have completed the last training form the space decoration model.

8. The method for generating a virtual space based on user preference information according to claim 1, characterized in that: Determining style information of each independent space in the virtual space according to the user's style preference information and the external contour information includes: Obtain information about items sold online; Determining a feature tag corresponding to the item sold in the network according to the item introduction information, wherein the feature tag is used to describe the decorative function of the item sold in the network in the space; The style information of each independent space in the virtual space is determined according to the style preference information of the user, the external contour information, and the feature tags corresponding to the items sold in the network.

9. A device for generating a virtual space based on user preference information, characterized in that: include: An acquisition unit, used to acquire the user's style preference information; A first determining unit, configured to determine external contour information of the virtual space according to the style preference information, wherein the external contour information at least includes apartment type information of the virtual space; a second determining unit, configured to determine style information of each independent space in the virtual space according to the user's style preference information and the external contour information, wherein the style information of each independent space at least includes type information of built-in items in each independent space and placement position information of each built-in item; The virtual space generating unit generates a target virtual space corresponding to the user according to the external contour information and the style information of each independent space.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored executable program, wherein when the executable program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for generating a virtual space based on user preference information as described in any one of claims 1 to 8.

11. An electronic device, characterized in that: include: A memory storing an executable program; A processor is used to run the program, wherein the program, when running, executes the method for generating a virtual space based on user preference information as described in any one of claims 1 to 8.

12. A computer program product comprising computer instructions, characterized in that: When the computer instructions are executed by a processor, the steps of the method for generating a virtual space based on user preference information as described in any one of claims 1 to 8 are implemented.

Images