WO2025200212A1 - 场景生成方法、装置、介质、设备和程序产品 - Google Patents
场景生成方法、装置、介质、设备和程序产品Info
- Publication number
- WO2025200212A1 WO2025200212A1 PCT/CN2024/108808 CN2024108808W WO2025200212A1 WO 2025200212 A1 WO2025200212 A1 WO 2025200212A1 CN 2024108808 W CN2024108808 W CN 2024108808W WO 2025200212 A1 WO2025200212 A1 WO 2025200212A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- scene
- graph
- candidate
- target
- display
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- A—HUMAN NECESSITIES
- A63—SPORTS; GAMES; AMUSEMENTS
- A63F—CARD, BOARD, OR ROULETTE GAMES; INDOOR GAMES USING SMALL MOVING PLAYING BODIES; VIDEO GAMES; GAMES NOT OTHERWISE PROVIDED FOR
- A63F13/00—Video games, i.e. games using an electronically generated display having two or more dimensions
- A63F13/50—Controlling the output signals based on the game progress
- A63F13/52—Controlling the output signals based on the game progress involving aspects of the displayed game scene
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/048—Interaction techniques based on graphical user interfaces [GUI]
- G06F3/0481—Interaction techniques based on graphical user interfaces [GUI] based on specific properties of the displayed interaction object or a metaphor-based environment, e.g. interaction with desktop elements like windows or icons, or assisted by a cursor's changing behaviour or appearance
- G06F3/0482—Interaction with lists of selectable items, e.g. menus
Definitions
- a receiving module is configured to receive scene description information of a target scene to be generated
- FIG4 is a schematic diagram of a candidate scene graph display provided according to an embodiment of the present disclosure.
- FIG7 is a block diagram of a scene generation device according to an embodiment of the present disclosure.
- FIG8 shows a schematic structural diagram of an electronic device suitable for implementing the embodiments of the present disclosure.
- the term “including” and its variations are open-ended, i.e., “including but not limited to.”
- the term “based on” means “based, at least in part, on.”
- the term “one embodiment” means “at least one embodiment,” the term “another embodiment” means “at least one additional embodiment,” and the term “some embodiments” means “at least some embodiments.”
- Other terms are defined in the following description.
- a prompt message is sent to the user to clearly inform the user that the operation requested will require the acquisition and use of the user's personal information. This allows the user to independently choose whether to provide personal information to the electronic device, application, server, storage medium, or other software or hardware that performs the operations of the disclosed technical solution based on the prompt message.
- the prompt information in response to receiving the user's active request, can be sent to the user in the form of a pop-up window, in which the prompt information can be presented in text form.
- the pop-up window can also carry a message for the user to choose "Agree” or “Disagree”. "Agree” to provide personal information to electronic devices.
- step 11 scene description information of a target scene to be generated is received.
- the scene description information may be information provided by the user describing the target scene they wish to build.
- the scene description information may be a natural language description text.
- an information input interface may be displayed, such as a text input box, so that the user can directly enter the description according to their needs.
- the user can enter parameters such as scene size and height in the corresponding interface based on the scene they wish to obtain, and enter keywords for the corresponding plot, such as style, color, holiday atmosphere, current events, etc.
- multiple attributes may be pre-set for the user to fill in.
- the scene generation can be performed using a pre-trained scene generation model.
- the scene generation model can be a large language model.
- the scene description information can be input into the scene generation model to obtain the output image as the candidate scene graph.
- a portion of the image output by the scene generation model can be selected as the candidate scene graph, for example, by randomly selecting a portion from the output image.
- step 13 the candidate scene graph is displayed.
- each generated candidate scene graph can be displayed separately in the scene graph display interface. If the current interface is not fully displayed, the scene graph display interface can be moved by sliding up and down or moving the scroll bar to display the candidate scene graphs, making it easier for users to select a scene graph that meets their building needs. As shown in Figure 2, the scene graph display interface can be moved by moving the scroll bar at A1.
- this step can achieve a three-dimensional representation of the selected target scene graph by generating voxel data corresponding to the scene graph, so as to subsequently generate a three-dimensional scene corresponding to the target scene graph.
- step 15 rendering is performed based on the voxel data to obtain a target scene, which is a three-dimensional scene.
- the user can select the candidate scene graph and confirm.
- the first candidate scene graph can be used as the target scene graph, and further rendered based on the voxel data corresponding to the first candidate scene graph to obtain the three-dimensional structure corresponding to the voxel data, that is, the target scene.
- the method of rendering the voxel data can be based on the common implementation method in this field, which will not be repeated here.
- an exemplary implementation manner of receiving scene description information of a target scene to be generated may include:
- a first dialogue area is displayed, wherein a plurality of candidate scene labels are displayed in the first dialogue area.
- the first dialogue area it is the first dialogue area, wherein the first dialogue area can be used to Prompt the user to input, such as "Please tell me the architectural style and characteristics you want, or you can choose from the keywords below", and accordingly, the candidate scene tags can be displayed in the first dialogue area.
- the candidate scene tags can be selected based on the tags in the tag library. For example, the topN candidate scene tags can be selected in descending order of selection frequency in the recent historical period. Alternatively, the topN candidate scene tags can be selected in descending order of similarity to the scenes corresponding to the current user.
- the specific selection method can be configured based on actual needs, and this disclosure does not limit this.
- the candidate scene label selected by the user is displayed in the second dialogue area, where the second dialogue area may be an area for the user to input information.
- the second dialog area is shown as Q2 in Figure 3.
- the three selected candidate scene labels can be automatically added and displayed in the second dialog area.
- the scene description information may be determined based on the display information in the second dialog area.
- the tag information corresponding to the candidate scene tag displayed in the second dialogue area can be used as the scene description information.
- receiving scene description information of a target scene to be generated further includes:
- the user's input information in the second dialogue area is received, where the input information includes text and/or images.
- users can also enter information in the second dialogue area, such as text or images. For example, after selecting the "Street" label, the user can further describe what the street scene in the scene they want to build is like, or they can query and upload the corresponding street image from a preset image library or the Internet.
- the label information corresponding to the input information and the candidate scene label selected by the user is used as the scene description information.
- the tag information and input information corresponding to the candidate scene tags displayed in the second dialogue area can be further used as the scene description information to facilitate subsequent accurate scene generation based on the scene description information.
- an exemplary implementation of receiving scene description information of a target scene to be generated may include:
- Receive user input information in the second dialogue area the input information including text and/or images. Further, the input information can be directly used as scene description information. In this embodiment, the user can directly enter the description information of the scene they want to build in the second dialogue area.
- an exemplary implementation of generating a candidate scene graph of multiple candidate scenes based on scene description information may include:
- the prompt text is input into a scene generation model, and a candidate scene graph of multiple candidate scenes is obtained according to the output of the scene generation model, wherein the scene generation model is implemented based on a large language model.
- the input format of the scene generation model can be pre-set.
- the user's input information and label information can be concatenated to form a prompt text.
- the scene generation model can be implemented based on a large language model (LLM).
- LLM large language model
- the LLM can be trained and learned based on the voxel data and description information of the generated scene.
- multiple candidate scene graphs can be generated based on the prompt text.
- an exemplary implementation of displaying a candidate scene graph may include:
- each candidate scene graph can be displayed, along with its corresponding edit control. components and selection controls.
- a portion of the scene graphs for display in the current round can be selected from the candidate scene graphs.
- four candidate scene graphs can be selected for display: candidate scene graphs S1-S4, corresponding to the first display areas P1-P4, respectively.
- the user can edit the candidate scene graphs using the edit control and select the candidate scene graphs using the select control.
- the control K1 can represent a select control
- the control K2 can be used to represent an edit control.
- the scene generation method further includes:
- a new candidate scene graph is generated and displayed according to the editing information and the candidate scene graph corresponding to the target editing control.
- the user when the generated candidate scene graph is displayed to the user, the user may be satisfied with the overall appearance of one of the candidate scene graphs, but only the local features thereof do not meet expectations.
- the user can edit and modify the single candidate scene graph through the editing control corresponding to the candidate scene graph.
- the user can edit the candidate scene graph S1 by clicking the editing control K2 in the first display area P1.
- the editing interface can be displayed.
- the corresponding candidate scene graph and its editable properties can be displayed in the editing interface, such as building color, single-story height, number of floors and other properties. For example, if the user wants to change the number of floors of the candidate scene graph S1 to 5 floors, the user can directly enter the editing information at the corresponding number of floors attribute. After the user submits and confirms, a new candidate scene graph can be generated and displayed based on the editing information and the candidate scene graph corresponding to the target editing control.
- the user may also select a portion of the area that he wants to adjust in the candidate scene graph.
- the area position information of the area selected by the user may be used as the editing
- the editing information is input into the scene generation model, thereby inputting the editing information and the candidate scene graph into the scene generation model to perform image adjustment and update to generate a new candidate scene graph.
- users can select a scene graph that meets their expectations from the candidate scene graphs, and can also realize secondary editing of a single candidate scene graph, further improving the matching degree between the candidate scene graph and user needs, while effectively reducing the data calculation and processing amount corresponding to the scene generation, and can provide effective and accurate data support for subsequent rendering of the target scene.
- another implementation of displaying the candidate scene graph may include:
- a presentation scene graph is determined from the candidate scene graphs.
- M scene graphs may be randomly selected as the presentation scene graph, i.e., the scene graph to be presented in the current round.
- the display scene graph is displayed. As an example, it can be displayed in the manner shown in FIG4 .
- the scene generation method may further include:
- a scene update control is displayed, such as the control shown by B1 in FIG4 .
- the scene update control is used to trigger the update of the displayed scene graph.
- the currently displayed scene graphs may not meet the user's building requirements.
- the user can click a scene update control.
- a new scene graph can be selected from the candidate scene graphs that have not yet been displayed as the display scene graph.
- a batch of candidate scene graphs can be regenerated based on the scene description information, and a scene graph can be further selected from the regenerated candidate scene graphs as the display scene graph.
- multiple new display scene graphs can be re-displayed through scene update, and the user can choose from the displayed new display scene graphs, providing the user with more options.
- a new display scene graph can be selected from the candidate scene graphs, reducing the call to the scene generation model during the scene generation process, and reducing the generation resource occupancy of the scene graph to a certain extent.
- the scene generation method may further include:
- a scene editing control is displayed, such as the control shown as B2 in FIG4 .
- the scene editing control is used to trigger the update of the scene description information.
- the user can further adjust the scene description information to further clarify and clarify his or her needs for building the target scene.
- the user can update the scene description information by clicking the scene editing control.
- a scene editing interface is displayed.
- the scene editing interface may be the interface shown in FIG3 , or may be an input box interface.
- new scene description information is obtained, and the operation of generating a candidate scene graph of multiple candidate scenes based on the scene description information is returned.
- the user may input new description information in the scene editing interface, and the existing scene description information and the new description information may be concatenated to serve as the new scene description information.
- the scene editing interface may display existing scene description information, and the user may add or modify the current scene description information.
- the information submitted in the scene editing interface may be used as the new scene description information.
- multiple candidate scene graphs corresponding to the target scene can be generated based on the new scene description information, and the candidate scene graphs can be displayed; in response to receiving the user's selection operation for the candidate scene graph, the step of rendering the target scene is performed according to the voxel data corresponding to the selected target scene graph.
- generating voxel data corresponding to the selected target scene graph may include:
- the selected target scene graph is gridded to obtain a grid display graph corresponding to the target scene graph.
- the selected target scene graph can be gridded.
- the target scene graph can be pixelated to obtain a corresponding mesh representation of the target scene graph.
- Figure 5 shows the mesh representation of the target scene graph selected by the user. Figure 5 only illustrates the meshing of the roof's sloped eaves. Other parts, such as surface B, are directly meshed.
- the grid display image is then displayed; in response to receiving a confirmation operation on the grid display image, voxel data corresponding to the confirmed target scene image is generated, and the confirmed target scene image is the target scene image corresponding to the grid display image indicated by the confirmation operation.
- the candidate scene graphs can be displayed to allow the user to have a rough preview of the scene graphs that can be generated, and the user can make a preliminary selection.
- the scene graph selected by the user can then be gridded, so that the user can further understand the overview of the corresponding three-dimensional scene graph based on the grid display graph to determine the scene graph used to generate voxel data later, thereby effectively reducing the processing volume of generating voxel data, improving the usability of the generated voxel data, and also improving the diversity of interaction with the user.
- generating voxel data corresponding to the selected target scene graph may include:
- a voxel data model can be trained based on data from existing scenes. For example, training can be performed based on voxel data corresponding to an already constructed scene and scene screenshot data of the scene. For example, the scene screenshot data can be used as the input of the model, and the voxel data corresponding to the scene can be used as the target input of the model to train the model to obtain the voxel data model.
- the voxel data model can be implemented based on a large language model or a neural network model.
- the target scene graph can be input into the voxel data model to obtain corresponding voxel data.
- the grid display graph can be input into the voxel data model. If the target scene graph does not generate a corresponding grid display graph, the target scene graph can be input into the voxel data model to obtain voxel data.
- voxel data corresponding to the scene graph can be quickly generated through the voxel data model.
- the correlation between the display graph of the existing scene and the voxel data in its construction process can be learned, thereby improving the effectiveness and accuracy of the voxel data, and improving the consistency between the image rendered based on the voxel data and the target scene graph, thereby ensuring the user experience.
- an exemplary implementation of generating voxel data corresponding to the selected target scene graph may include:
- the target scene graph has a corresponding grid display graph, it can be directly obtained. If the target scene graph has not yet generated a corresponding grid display graph, the target scene graph can be pixelated to generate the grid display graph.
- various scene parts in the target scene graph can be identified, such as door frames, walls, roofs, foundations, and the like.
- the properties of the meshes therein can then be determined based on the types of the identified parts.
- the properties of each mesh can then be matched with the types of voxel blocks in a voxel block library to determine the voxel block corresponding to the mesh.
- Voxel blocks with the same properties as the mesh or a similarity greater than a threshold can be used as the voxel blocks corresponding to the mesh, thereby obtaining display voxel data corresponding to the portion of the target scene graph displayed to the user.
- the portion of the target scene graph that is not displayed to the user can then be further filled based on the target scene and the display voxel data.
- the size information corresponding to the target scene graph can be determined based on the grid display graph.
- the size information includes the length and width of the target scene on the horizontal plane. Accordingly, an exemplary implementation method for determining the size information corresponding to the target scene graph based on the grid display graph may include:
- the length and width of the target scene on the horizontal plane are determined according to the display images of the two adjacent surfaces.
- the grid display contains two adjacent vertical surfaces of the target scene.
- the display diagram i.e., surface A and surface B in Figure 5
- the length of the target scene can be used as the length of the target scene based on the length of the intersection line of surface A and the horizontal plane, and as the width of the target scene based on the length of the intersection line of surface B and the horizontal plane.
- the length of the intersection line can be adjusted based on the tilt angle of surface B to serve as the width of the target scene.
- This adjustment method can be based on the dimensional change under the perspective angle commonly used in the art to adjust the width, which will not be further described here.
- the length corresponding to the vertical surface is determined based on the display image of the vertical surface, and the width is determined based on the length corresponding to the vertical surface.
- the length of the intersection of the display image and the horizontal surface can be used as the length of the target scene.
- a default ratio of length and width can be pre-set. After determining the length, the corresponding width can be determined based on the length and the default ratio to obtain the size information.
- the display voxel data is filled based on the grid display graph and size information to obtain the voxel data corresponding to the target scene graph.
- the width determined based on the B surface is 10
- the grid on the front of the grid display diagram can be filled in the S direction shown in FIG5 , wherein the filling can be based on the default voxel block, or based on the association between different voxel blocks.
- a variety of filling rules can be pre-set. For example, if the type of voxel block in the display voxel data is a wall, then the four voxel blocks connected in the S direction must also be of the wall type. When filling the display voxel data, the corresponding filling can be performed based on the rule.
- the internal structure of the scene that is not displayed can be matched based on the part that has been displayed on the surface, and thus filled according to the matched structure. If the display voxel data indicates that the scene is a three-story building structure, then the internal structure can be matched to the building structure, and then the internal filling can be filled with a combination of multiple voxel blocks corresponding to the staircase structure. For another example, if the display voxel data indicates that the top floor of the scene is a spire structure, then the top can be filled in the S direction corresponding to the top according to the matched spire. The top structure is filled to obtain the voxel data corresponding to the target scene graph. Default voxels can be filled when filling internally. Users can subsequently change and adjust the type of internal voxel data to achieve personalized scene building, further improve the user's interactive method for scene building, and simplify the user's scene building process.
- the scene generation method may further include:
- the scene map may be a map of areas in the virtual scene where scene construction is allowed, which may be determined based on pre-configuration.
- the user can select the area in the scene map where he wants to build the scene.
- the area selection can be performed in the form of a parallelogram.
- you can first move to the starting position A, and then move to the end position A'.
- the area information can include the starting position and the end position.
- the middle area formed by the straight line from the starting point to the end point as the diagonal line can be used as the target area.
- the target area can be generated based on the default height, or the user can further determine an auxiliary third point, with the third point as the vertex and the middle area formed by the straight line from the starting point to the end point as the diagonal line as the target area.
- the scene description information and the target area are input into the scene generation model to obtain multiple candidate scenes corresponding to the target area.
- displaying a target scene in a target area includes:
- the fourth processing submodule is configured to fill the display voxel data based on the grid display graph and the size information to obtain voxel data corresponding to the target scene graph.
- the size information includes the length and width of the target scene on the horizontal plane
- a fifth determining submodule configured to determine the length and width of the target scene on the horizontal plane based on the display images of the two adjacent vertical surfaces if the grid display image includes the display images of the two adjacent surfaces;
- a second display module is configured to display a scene map
- a first determining module is configured to determine a target area in response to area information selected by a user in the scene map
- the third display module is configured to display the target scene in the target area.
- the fourth display submodule is configured to display the target scene at the target position in response to the target position selected in the target area if the target scene includes a scene unit.
- a fifth display submodule is configured to display, for a candidate scene graph, an edit control and a selection control corresponding to the candidate scene graph in a first display area corresponding to the candidate scene graph;
- a fourth display module configured to display, in response to receiving a selection operation on the editing control, a candidate scene graph corresponding to the selected target editing control in the editing interface
- the second processing module is configured to generate and display a new candidate scene graph in response to editing information input by the user in the editing interface according to the editing information and the candidate scene graph corresponding to the target editing control.
- the first display module includes:
- a seventh determination submodule configured to determine a presentation scene graph from the candidate scene graphs
- a sixth display submodule configured to display the display scene graph
- the scene generating device also includes:
- a fifth display module configured to display a scene update control when displaying the display scene graph
- a second determining module configured to determine a new presentation scene graph in response to receiving a selection operation on the scene update control
- the first generating module includes:
- FIG8 illustrates a schematic diagram of the structure of an electronic device (e.g., a terminal device or server) 600 suitable for implementing embodiments of the present disclosure.
- the terminal device in the embodiments of the present disclosure may include, but is not limited to, mobile terminals such as mobile phones, laptop computers, digital broadcast receivers, PDAs (personal digital assistants), PADs (tablet computers), PMPs (portable multimedia players), in-vehicle terminals (e.g., in-vehicle navigation terminals), and fixed terminals such as digital TVs and desktop computers.
- the electronic device illustrated in FIG8 is merely an example and should not limit the functionality or scope of use of the embodiments of the present disclosure.
- electronic device 600 may include a processing device (e.g., a central processing unit, a graphics processing unit, etc.) 601, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 602 or a program loaded from a storage device 608 into a random access memory (RAM) 603.
- ROM read-only memory
- RAM random access memory
- Various programs and data required for the operation of electronic device 600 are also stored in RAM 603.
- Processing device 601, ROM 602, and RAM 603 are connected to each other via a bus 604.
- An input/output (I/O) interface 605 is also connected to bus 604.
- input devices 606 including, for example, a touch screen, a touchpad, a keyboard, a mouse, a camera, a microphone, an accelerometer, a gyroscope, etc.
- Output device 607 such as a liquid crystal display (LCD), speaker, or vibrator
- storage device 608 such as a magnetic tape or hard disk
- communication device 609 can allow electronic device 600 to communicate with other devices wirelessly or by wire to exchange data.
- FIG8 shows electronic device 600 with various devices, it should be understood that not all of the devices shown are required to be implemented or present. More or fewer devices may be implemented or present instead.
- an embodiment of the present disclosure includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes program code for executing the method shown in the flowchart.
- the computer program can be downloaded and installed from the network through the communication device 609, or installed from the storage device 608, or installed from the ROM 602.
- the processing device 601 the above-mentioned functions defined in the method of the embodiment of the present disclosure are performed.
- the computer-readable medium mentioned above in the present disclosure may be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two.
- a computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or component, or any combination of the above.
- Computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.
- a computer-readable storage medium may be any tangible medium that contains or stores a program that can be used by or in conjunction with an instruction execution system, device, or component.
- a computer-readable signal medium may include a data signal propagated in baseband or as part of a carrier wave, which carries computer-readable program code. Such a propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above.
- a computer-readable signal medium may also be any computer-readable medium other than a computer-readable storage medium that can transmit, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.
- the program code contained on the computer-readable medium may be transmitted using any suitable medium, including but not limited to wires, optical cables, RF (radio frequency), etc., or any suitable combination thereof.
- the client and server can communicate using any currently known or later developed network protocol, such as HTTP (HyperText Transfer Protocol), and can be interconnected with any form or medium of digital data communication (e.g., a communication network).
- HTTP HyperText Transfer Protocol
- Examples of communication networks include a local area network ("LAN”), a wide area network ("WAN”), an internet (e.g., the Internet), and a peer-to-peer network (e.g., an ad hoc peer-to-peer network), as well as any currently known or later developed network.
- the computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.
- the above-mentioned computer-readable medium carries one or more programs.
- the electronic device When the above-mentioned one or more programs are executed by the electronic device, the electronic device: receives scene description information of the target scene to be generated; generates candidate scene graphs of multiple candidate scenes based on the scene description information, wherein the candidate scene graphs are two-dimensional images; displays the candidate scene graphs; in response to receiving a user's selection operation for the candidate scene graph, generates voxel data corresponding to the selected target scene graph; and renders based on the voxel data to obtain the target scene, wherein the target scene is a three-dimensional scene.
- Computer program code for performing the operations of the present disclosure may be written in one or more programming languages, or a combination thereof, including, but not limited to, object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" or similar programming languages.
- the program code may be executed entirely on the user's computer, partially on the user's computer, as a stand-alone software package, partially on the user's computer and partially on a remote computer, or entirely on the remote computer or server.
- the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).
- LAN local area network
- WAN wide area network
- Internet service provider e.g., AT&T, MCI, Sprint, EarthLink, MSN, GTE, etc.
- each box in the flowchart or block diagram may represent a module, program segment, or a portion of code, which contains one or more executable instructions for implementing the specified logical functions.
- the functions marked in the boxes may also occur in an order different from that marked in the accompanying drawings. For example, two boxes represented in succession may actually be executed substantially in parallel, and they may sometimes be executed in the opposite order, depending on the functions involved.
- each block in the block diagrams and/or flowcharts, and combinations of blocks in the block diagrams and/or flowcharts can be implemented by a dedicated hardware-based system that performs the specified functions or operations, or can be implemented by a combination of dedicated hardware and computer instructions.
- modules described in the embodiments of the present disclosure may be implemented in software or hardware.
- the name of a module does not limit the module itself.
- a receiving module may also be described as a "module that receives scene description information of a target scene to be generated.”
- exemplary types of hardware logic components include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chip (SOCs), complex programmable logic devices (CPLDs), and the like.
- FPGAs field programmable gate arrays
- ASICs application specific integrated circuits
- ASSPs application specific standard products
- SOCs systems on chip
- CPLDs complex programmable logic devices
- a machine-readable medium can be a tangible medium that can contain or store a program for use by or in conjunction with an instruction execution system, device or equipment.
- a machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium.
- a machine-readable medium can include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or equipment, or any suitable combination of the foregoing.
- a more specific example of a machine-readable storage medium can include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
- RAM random access memory
- ROM read-only memory
- EPROM or flash memory erasable programmable read-only memory
- CD-ROM portable compact disk read-only memory
- CD-ROM compact disk read-only memory
- magnetic storage device or any suitable combination of the foregoing.
- Example 1 provides a scene generation method, the scene generation method including:
- generating a candidate scene graph for a plurality of candidate scenes wherein the candidate scene graph is a two-dimensional image generated based on voxel data corresponding to the candidate scenes;
- rendering is performed according to voxel data corresponding to the selected target scene graph to obtain a target scene, where the target scene is a three-dimensional scene.
- Example 2 provides the method of Example 1, wherein receiving scene description information of a target scene to be generated includes:
- the scene description information is determined according to the display information in the second dialog area.
- Example 3 provides the method of Example 2, wherein receiving scene description information of a target scene to be generated further includes:
- the label information corresponding to the input information and the candidate scene label selected by the user is used as the scene description information.
- voxel data corresponding to the confirmed target scene graph is generated, and the confirmed target scene graph is the target scene graph corresponding to the grid presentation graph indicated by the confirmation operation.
- Example 5 provides the method of Example 1, wherein, in response to receiving a user selection operation for a candidate scene graph, generating voxel data corresponding to the selected target scene graph includes:
- the target scene graph is input into a voxel data model to obtain voxel data corresponding to the target scene graph, wherein the voxel data model is obtained by training based on voxel data corresponding to the three-dimensional scene and scene screenshot data corresponding to the three-dimensional scene.
- the display voxel data is filled based on the grid display image and size information to obtain the voxel data corresponding to the target scene image.
- Example 7 provides the method of Example 6, wherein the size information includes a length and a width corresponding to the target scene on a horizontal plane;
- Determine the size information corresponding to the target scene graph based on the grid display graph including:
- Example 8 provides the method of Example 1, wherein the scene generation method further includes:
- Example 9 provides the method of Example 8, wherein presenting the target scene in the target area includes:
- the target scene contains multiple scene units, then the multiple scene units are displayed in the target area;
- the target scene includes a scene unit
- the target scene is displayed at the target position in response to the target position being selected in the target area.
- Example 10 provides the method of Example 1, wherein presenting the candidate scene graph includes:
- a new candidate scene graph is generated and displayed according to the editing information and the candidate scene graph corresponding to the target editing control.
- the scene generation method further includes:
- Example 12 provides the method of Example 1, wherein the scene generation method further includes:
- new scene description information is obtained, and the operation of generating a candidate scene graph of multiple candidate scenes based on the scene description information is returned.
- Example 13 provides the method of Example 1, wherein generating a candidate scene graph of multiple candidate scenes based on scene description information includes:
- the prompt text is input into a scene generation model, and a candidate scene graph of multiple candidate scenes is obtained according to the output of the scene generation model, wherein the scene generation model is implemented based on a large language model.
- Example 14 provides a scene generation device, which includes: a receiving module, configured to receive scene description information of a target scene to be generated; a first generation module, configured to generate a candidate scene graph of multiple candidate scenes based on the scene description information, wherein the candidate scene graph is a two-dimensional image; a first display module, configured to display the candidate scene graph; a second generation module, configured to generate voxel data corresponding to the selected target scene graph in response to receiving a user's selection operation for the candidate scene graph; and a first processing module, configured to render based on the voxel data to obtain a target scene, wherein the target scene is a three-dimensional scene.
- a receiving module configured to receive scene description information of a target scene to be generated
- a first generation module configured to generate a candidate scene graph of multiple candidate scenes based on the scene description information, wherein the candidate scene graph is a two-dimensional image
- a first display module configured to display the candidate scene graph
- a second generation module configured to generate voxe
- Example 15 provides a computer-readable medium having a computer program stored thereon, which, when executed by a processing device, implements the steps of the scene generation method involved in any one of Examples 1-13.
- Example 16 provides an electronic device, comprising: a storage device having a computer program stored thereon; a processing device configured to execute the computer program stored in the storage device; A computer program for implementing the steps of the scene generation method involved in any one of Examples 1-13.
- Example 17 provides a computer program product, including a computer program, which implements the steps of the method of any one of Examples 1-13 when executed by a processor.
Landscapes
- Engineering & Computer Science (AREA)
- General Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Multimedia (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Processing Or Creating Images (AREA)
Abstract
一种场景生成方法、装置、介质、设备和程序产品,场景生成方法包括:接收待生成的目标场景的场景描述信息(11);基于场景描述信息,生成多个候选场景的候选场景图,其中,候选场景图为二维图像(12);展示候选场景图(13);响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据(14);基于体素数据进行渲染获得目标场景,目标场景为三维场景(15)。当用户想要进行场景生成时,用户只需要对其想要搭建的场景进行描述,无需其进行手动设计和搭建,简化场景生成操作,有效降低场景搭建周期。
Description
本申请要求于2024年3月28日递交的中国专利申请第202410371010.X号的优先权,在此全文引用上述中国专利申请公开的内容以作为本申请的一部分。
本公开的实施例涉及一种场景生成方法、装置、介质、设备和程序产品。
目前体素游戏的技术下,通常需要用户搭建场景以进行交互操作。通常需要玩家手动搭建场景以获得其所需的场景环境,或者玩家可以通过复制和共享其他玩家的场景进行搭建或改造。上述搭建方式中均需要依靠玩家完全投入的设计能力与动手制作能力,搭建技术要求高,且搭建周期长。
发明内容
第一方面,本公开提供一种场景生成方法,所述方法包括:
接收待生成的目标场景的场景描述信息;
基于所述场景描述信息,生成多个候选场景的候选场景图,其中,所述候选场景图为二维图像;
展示所述候选场景图;
响应于接收到用户针对所述候选场景图的选择操作,生成所选择的目标场景图对应的体素数据;
基于所述体素数据进行渲染获得所述目标场景,所述目标场景为三维场景。
第二方面,本公开提供一种场景生成装置,所述装置包括:
接收模块,被配置为接收待生成的目标场景的场景描述信息;
第一生成模块,被配置为基于所述场景描述信息,生成多个候选场景的候选场景图,其中,所述候选场景图为二维图像;
第一展示模块,被配置为展示所述候选场景图;
第二生成模块,被配置为响应于接收到用户针对所述候选场景图的选择操作,生成所选择的目标场景图对应的体素数据;
第一处理模块,被配置为基于所述体素数据进行渲染获得所述目标场景,所述目标场景为三维场景。
第三方面,本公开提供一种计算机可读介质,其上存储有计算机程序,该程序被处理装置执行时实现第一方面所述方法的步骤。
第四方面,本公开提供一种电子设备,包括:
存储装置,其上存储有计算机程序;
处理装置,被配置为执行所述存储装置中的所述计算机程序,以实现第一方面所述方法的步骤。
第五方面,本公开提供一种计算机程序产品,包括计算机程序,该计算机程序被处理器执行时实现第一方面所述的方法的步骤。
结合附图并参考以下具体实施方式,本公开各实施例的上述和其他特征、优点及方面将变得更加明显。贯穿附图中,相同或相似的附图标记表示相同或相似的元素。应当理解附图是示意性的,原件和元素不一定按照比例绘制。在附图中:
图1是根据本公开的一种实施方式提供的场景生成方法的流程图。
图2是根据本公开的一种实施方式提供的场景图展示界面的示意图。
图3是根据本公开的一种实施方式提供的接收待生成的目标场景的场景描述信息的示意图。
图4是根据本公开的一种实施方式提供的候选场景图展示的示意图。
图5是根据本公开的一种实施方式提供的网格展示图的示意图。
图6是根据本公开的一种实施方式提供的目标区域选择的示意图。
图7是根据本公开的一种实施方式提供的场景生成装置的框图。
图8示出了适于用来实现本公开实施例的电子设备的结构示意图。
下面将参照附图更详细地描述本公开的实施例。虽然附图中显示了本公开的某些实施例,然而应当理解的是,本公开可以通过各种形式来实现,而且不应该被解释为限于这里阐述的实施例,相反提供这些实施例是为了更加透彻和完整地理解本公开。应当理解的是,本公开的附图及实施例仅用于示例性作用,并非用于限制本公开的保护范围。
应当理解,本公开的方法实施方式中记载的各个步骤可以按照不同的顺序执行,和/或并行执行。此外,方法实施方式可以包括附加的步骤和/或省略执行示出的步骤。本公开的范围在此方面不受限制。
本文使用的术语“包括”及其变形是开放性包括,即“包括但不限于”。术语“基于”是“至少部分地基于”。术语“一个实施例”表示“至少一个实施例”;术语“另一实施例”表示“至少一个另外的实施例”;术语“一些实施例”表示“至少一些实施例”。其他术语的相关定义将在下文描述中给出。
需要注意,本公开中提及的“第一”、“第二”等概念仅用于对不同的装置、模块或单元进行区分,并非用于限定这些装置、模块或单元所执行的功能的顺序或者相互依存关系。
需要注意,本公开中提及的“一个”、“多个”的修饰是示意性而非限制性的,本领域技术人员应当理解,除非在上下文另有明确指出,否则应该理解为“一个或多个”。
本公开实施方式中的多个装置之间所交互的消息或者信息的名称仅用于说明性的目的,而并不是用于对这些消息或信息的范围进行限制。
可以理解的是,在使用本公开各实施例公开的技术方案之前,均应当依据相关法律法规通过恰当的方式对本公开所涉及个人信息的类型、使用范围、使用场景等告知用户并获得用户的授权。
例如,在响应于接收到用户的主动请求时,向用户发送提示信息,以明确地提示用户,其请求执行的操作将需要获取和使用到用户的个人信息。从而,使得用户可以根据提示信息来自主地选择是否向执行本公开技术方案的操作的电子设备、应用程序、服务器或存储介质等软件或硬件提供个人信息。
作为一种可选的但非限定性的实现方式,响应于接收到用户的主动请求,向用户发送提示信息的方式例如可以是弹窗的方式,弹窗中可以以文字的方式呈现提示信息。此外,弹窗中还可以承载供用户选择“同意”或者“不同
意”向电子设备提供个人信息的选择控件。
可以理解的是,上述通知和获取用户授权过程仅是示意性的,不对本公开的实现方式构成限定,其它满足相关法律法规的方式也可应用于本公开的实现方式中。
同时,可以理解的是,本技术方案所涉及的数据(包括但不限于数据本身、数据的获取或使用)应当遵循相应法律法规及相关规定的要求。
图1所示,为根据本公开的一种实施方式提供的场景生成方法的流程图,如图1所示,该场景生成方法可以包括:
在步骤11中,接收待生成的目标场景的场景描述信息。
其中,该场景描述信息可以是用户对其想要搭建的目标场景的说明信息,例如,该场景描述信息可以为自然语言描述文本,例如可以展示信息输入界面,如可以显示一个文本输入框,则用户可以直接根据自己的需求进行描述输入,如用户可以根据想要获得的场景,在对应界面输入场景大小、高度等参数,并输入对应剧情,如风格、色彩、节日氛围、当前事件等的关键词。作为另一示例,为了保证场景描述信息的全面性,也可以预先设置多个属性以由用户进行填写。
在步骤12中,基于场景描述信息,生成多个候选场景的候选场景图,其中,候选场景图为二维图像。
其中,该步骤中可以通过预先训练的场景生成模型进行场景生成,如该场景生成模型可以为大语言模型等,则可以将该场景描述信息输入该场景生成模型后获得其输出的图像作为候选场景图。作为另一示例,也可以从场景生成模型输出的图像中选择部分作为该候选场景图,如可以通过随机选择的方式从输出的图像中选择部分作为候选场景图。
在步骤13中,展示候选场景图。
作为示例,可以在场景图展示界面中分别展示生成的每一候选场景图,若当前界面中未完全展示,则可以通过上下滑动操作或者移动滚动条的方式等移动场景图展示界面,以对候选场景图进行展示,便于用户从中选择满足其搭建需求的场景图。如图2所示可以通过移动A1处的滚动条实现场景图展示界面的移动。
在步骤14中,响应于接收到用户针对候选场景图的选择操作,生成所选
择的目标场景图对应的体素数据。
其中,体素是体积元素(Volume Pixel)的简称,包含体素的立体可以通过立体渲染或者提取给定阈值轮廓的多边形等值面表现出来。该实施例中,候选场景对应的体素数据中包含用于搭建该候选场景的各个体素块以及该体素块的属性信息。例如候选场景为别墅,其中可以包含墙壁体素块、天花板体素块以及门框体素块等。
示例地,基于步骤12中确定出的候选场景图为二维图像,则该步骤中可以通过生成场景图对应的体素数据,实现所选择的目标场景图的三维表示,以便后续生成目标场景图对应的三维场景。
在步骤15中,基于体素数据进行渲染获得目标场景,目标场景为三维场景。
其中,若当前展示的候选场景图满足用户的搭建需求,则用户可以选中该候选场景图并确认。如图2中,用户选择第一个候选场景图,则可以将该第一个候选场景图作为目标场景图,并进一步基于该第一个候选场景图对应的体素数据进行渲染以获得该体素数据对应的三维结构,即该目标场景。其中,对体素数据进行渲染的方式可以基于本领域中的通用实现方式,在此不再赘述。
由此,通过上述技术方案,当用户想要进行场景生成时,用户只需要对其想要搭建的场景进行描述,无需其进行手动设计和搭建,有效降低用户进行场景生成搭建的技术要求,简化场景生成操作,有效降低场景搭建周期。并且,可以通过用户对目标场景的场景描述信息以生成候选场景图对用户进行展示,则用户可以基于该候选场景图对多种可能的场景进行预览,以从中确定满足其搭建需求的场景,在用户选择后再进行场景搭建,一方面可以有效提高生成的目标场景与用户需求进行的匹配度,同时可以有效避免对大量候选场景进行场景生成所造成的资源浪费,从而提高场景生成的效率,提高用户之间交互的多样性。
在一种可能的实施例中,接收待生成的目标场景的场景描述信息的示例性实现方式可以包括:
展示第一对话区域,第一对话区域中显示有多个候选场景标签。
如图3中Q1所示,为第一对话区域,其中,该第一对话区域可以用于
对用户进行输入提示,如“请告诉我想要的建筑风格和特点,也可以从下面的关键词中选择”,相应的,可以在第一对话区域中展示候选场景标签。其中,该候选场景标签可以基于标签库中的标签进行选择。例如可以将最近的历史时段内选择频次由高至低的顺序选择topN作为候选场景标签。也可以是将与当前用户对应的场景的相似度由高至低的顺序选择topN作为候选场景标签,其具体选择方式可以基于实际需求进行配置,本公开对此不作限定。
响应于用户对候选场景标签的选择,将用户选择的候选场景标签显示在第二对话区域,其中该第二对话区域可以为用于用户进行信息输入的区域。
如图3中Q2所示为第二对话区域。用户选择“欧式”、“城市”、“街道”三个候选场景标签,则可以将选中的三个候选场景标签自动添加显示到第二对话区域中。
之后,可以根据第二对话区域中的显示信息确定场景描述信息。
作为示例,若用户仅选择候选场景标签,则可以将第二对话区域中的显示的候选场景标签对应的标签信息作为该场景描述信息。
由此,通过上述技术方案,可以通过第一对话区域中展示多个候选场景标签为用户进行提示,以为用户确定场景描述信息提供多个可选标签,以在一定程度上细化用户要生成的场景的提示,简化场景描述信息的生成流程,进一步简化用户操作流程。
作为另一示例,接收待生成的目标场景的场景描述信息,还包括:
接收用户在第二对话区域中的输入信息,输入信息包含文本和/或图像。
为了进一步提高场景描述信息的全面性和多样性,用户也可以在第二对话区域中进行信息输入,如可以输入文本或图像,例如用户在选择“街道”标签后可以进一步描述其想要搭建的场景中街道的场景是什么样的,或者也可以从预置的图像库或者网络上查询对应的街道图像并上传。
相应地,将输入信息和用户选择的候选场景标签对应的标签信息,作为场景描述信息。
则该实施例中,可以进一步地将第二对话区域中的显示的候选场景标签对应的标签信息和输入信息作为该场景描述信息,以便于后续基于该场景描述信息进行准确的场景生成。
由此,通过上述技术方案,可以结合候选场景标签和用户的输入信息确
定场景描述信息,则用户可以通过文本描述或者图像描述的方式对其想要进行搭建的场景进行直观描述,从而可以自动化生成对应的场景,无需用户对场景搭建过程进行设计和手动搭建,提高场景生成的效率。
作为示例,接收待生成的目标场景的场景描述信息的示例性实现方式可以包括:
接收用户在第二对话区域中的输入信息,输入信息包含文本和/或图像。进一步地,可以直接将该输入信息作为场景描述信息。该实施例中,用户可以直接在第二对话区域中输入其想要搭建的场景的描述信息。
在一种可能的实施例中,基于场景描述信息,生成多个候选场景的候选场景图的示例性实现方式可以包括:
根据场景描述信息构建提示文本。
将提示文本输入场景生成模型,根据场景生成模型的输出获得多个候选场景的候选场景图,其中,场景生成模型基于大语言模型实现。
作为示例,可以预先设置场景生成模型的输入对应的格式。如可以将用户的输入信息和标签信息进行拼接作为提示文本prompt。其中,场景生成模型可以基于大语言模型(LLM,Large Language Model)实现,例如可以基于已生成的场景的体素数据和其描述信息进行大语言模型训练和学习。则该场景生成过程中,可以基于提示文本生成多个候选场景图。
作为示例,可以直接基于场景生成模型生成候选场景图对应的体素数据,在候选场景图中仅展示二维表面图,则可以将体素数据对应的场景图像进行截图处理获得候选场景图,以使得用户可以基于该候选场景图进行场景预览,在场景渲染生成之前可以对场景的外在展示有直观的了解。作为另一示例,可以通过场景生成模型直接生成二维的候选场景图,以便基于候选场景图对用户进行提示,待用户确认后再进一步生成对应的三维场景,从而可以提升用户使用体验。基于大语言模型进行场景的体素数据的生成和推荐,能在一定程度上突破人工设计的范围局限,拓宽本公开方法的使用范围。
在一种可能的实施例中,展示候选场景图的示例性实现方式可以包括:
针对候选场景图,在该候选场景图对应的第一展示区域中展示该候选场景图对应的编辑控件和选择控件。
作为示例,可以对每一候选场景图进行展示,同时展示其对应的编辑控
件和选择控件。
作为另一示例,可以从候选场景图中选择部分当前轮次进行展示的场景图,如图4所示,可以从中选择4个候选场景图进行展示,候选场景图S1-S4,其分别对应的第一展示区域为P1-P4。用户可以通过编辑控件实现对候选场景图的编辑,通过选择控件实现对候选场景图的选择。如图4中所示,该控件K1可以表示选择控件,控件K2可以用于表示编辑控件。
相应地,该场景生成方法还包括:
响应于接收到对编辑控件的选择操作,在编辑界面中展示选择的目标编辑控件对应的候选场景图;
响应于用户在编辑界面中输入的编辑信息,根据编辑信息和目标编辑控件对应的候选场景图,生成新的候选场景图并展示。
在该实施例中,在为用户展示生成的候选场景图时,用户可能对其中的某一候选场景图的整体是符合预期的,只是其中的局部特征不符合预期,则该场景中用户可以通过候选场景图对应的编辑控件实现对单一候选场景图的编辑和修改。
作为示例,用户可以通过点击第一展示区域P1中的编辑控件K2而实现对候选场景图S1的编辑。在用户选择该编辑控件后可以展示编辑界面,作为示例,可以在编辑界面中展示对应的候选场景图以及其可编辑的属性,例如可以是建筑颜色、单层高度、层数等属性。例如,用户想要将候选场景图S1的层数修改为5层,则可以直接在对应的层数属性处输入编辑信息。则该用户提交确认后,可以根据编辑信息和目标编辑控件对应的候选场景图,生成新的候选场景图并展示。
作为示例,在对候选场景图进行展示时,可以将新的候选场景图S1’替换候选场景图S1进行展示,即对S1’、S2、S3和S4进行展示。作为另一示例,当用户对候选场景图S1进行编辑时,可以认为用户对当前展示的候选场景图中的S1的感兴趣程度最高,则在生成新的候选展示图S1’后,可以仅展示该新生成的候选场景图S1’,以便于用户快速确认新的候选场景图对应的场景是否符合其预期。
作为另一示例,用户也可以对候选场景图中的想要调整的部分区域进行选择,则响应于用户的选择,可以将用户选中的区域的区域位置信息作为编
辑信息,从而将该编辑信息和该候选场景图输入场景生成模型,以进行图像调整和更新,生成新的候选场景图。
由此,通过上述技术方案,用户可以从候选场景图中选择符合其预期的场景图,同时也可以实现对单个候选场景图的二次编辑,进一步提高候选场景图与用户需求的匹配度的同时,有效降低场景生成对应的数据计算量和处理量,并且可以为后续进一步渲染目标场景提高有效且准确的数据支持。
在一种可能的实施例中,展示候选场景图的另一种实现方式可以包括:
从候选场景图中确定出展示场景图。作为示例,可以从中随机选择M个场景图作为展示场景图,即当前轮次进行展示的场景图。
对展示场景图进行展示。作为示例,可以如图4所示方式进行展示。
相应地,该场景生成方法还可以包括:
在对展示场景图进行展示时,展示场景更新控件,如图4中B1所示控件,场景更新控件用于触发更新展示的场景图。
响应于接收到对场景更新控件的选择操作,确定新的展示场景图;
展示新的展示场景图。
一种可能的实施例中,当前展示的各个展示场景图可能均未满足用户的搭建需求,则该实施例中,用户可以点击场景更新控件。响应于接收到对场景更新控件的选择操作,作为示例,可以从候选场景图中尚未展示的场景图中选择新的场景图作为展示场景图。作为另一示例,若候选场景图的场景图均被展示过,则可以基于场景描述信息重新生成一批候选场景图而进一步从重新生成的候选场景图中选择场景图作为该展示场景图。
由此,该实施例中,当前展示的场景图中没有满足用户搭建需求的场景图时,可以用过场景更新重新展示多个新的展示场景图,则用户可以从展示的新的展示场景图中进行选择,为用户提供更多的可选项,同时可以从候选场景图中选择新的展示场景图,降低场景生成过程中对场景生成模型的调用,在一定程度上降低场景图的生成资源占用。
在一种可能的实施例中,该场景生成方法还可以包括:
在展示候选场景图时展示场景编辑控件,如图4中B2所示控件,场景编辑控件用于触发更新场景描述信息。
作为示例,当展示的候选场景图中没有符合用户搭建需求的场景图时,
用户可以进一步对场景描述信息进行调整,以进一步明确和清晰自己对目标场景的搭建需求,则该实施例中,用户可以通过点击该场景编辑控件实现对场景描述信息的更新。
响应于接收到对场景编辑控件的选择操作,展示场景编辑界面。
作为示例,该场景编辑界面可以如图3所示的界面,也可以是输入框界面。
响应于用户在场景编辑界面中的编辑操作,获得新的场景描述信息,并返回执行基于场景描述信息,生成多个候选场景的候选场景图的操作。
作为示例,用户可以在场景编辑界面中输入新增的描述信息,则可以将已有的场景描述信息和新增的描述信息的拼接作为该新的场景描述信息。
作为另一示例,场景编辑界面中可以展示已有的场景描述信息,则用户可以在当前的场景描述信息的基础上进行新增或修改,则该示例中可以将场景编辑界面中的提交的信息作为该新的场景描述信息。
在获得新的场景描述信息后,则可以基于该新场景描述信息,生成目标场景对应的多个候选场景图,以及展示候选场景图;响应于接收到用户针对候选场景图的选择操作,根据选择的目标场景图对应的体素数据进行渲染获得目标场景的步骤。
由此,通过上述技术方案,在对用户展示候选场景图的过程中,用户可以多次编辑和调整场景描述信息,以提高场景描述信息中的信息量,从而获得满足用户搭建需求的候选场景图,该过程中生成二维的预览图像即可,无需对三维场景进行多次渲染,在保证生成的目标场景满足用户的搭建需求的同时,提高场景生成的效率。同时提高UGC(User Generated Content,用户生成内容)的丰富度,以对用户生成内容进行数据参考和激发,使得用户可以更简单地获得自己想要的目标场景。
在一种可能的实施例中,响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据的实现方式可以包括:
响应于选择操作,将所选择的目标场景图进行网格化处理,获得目标场景图对应的网格展示图。
如图2中所示,若用户认为第一个图满足其需求,则可以对该第一个图进行选择。在用户确定选择后,则可以对选择的目标场景图进行网格化处理,
作为示例,可以对目标场景图进行像素化处理,以获取该目标场景图对应的网格展示图。如图5所示,为用户选择的目标场景图对应的网格展示图,其中在图5中仅对屋顶斜线屋檐部分的网格化处理进行展示以示例说明,其他部分直接网格切分,如B面的网格处理未体现。
之后展示网格展示图;响应于接收到对网格展示图的确认操作,生成所确认的目标场景图对应的体素数据,所确认的目标场景图为确认操作指示的网格展示图对应的目标场景图。
其中,场景图在进行网格化处理后其展示与二维图像略有差异,则在对场景图进行网格化处理时,用户可以选择一个或多个候选场景图进行网格化处理。当用户选择多个时,可以分别对用户选择的每一个场景图进行网格化处理并输出展示,以使得用户能够从网格展示图中进一步确定符合其需求的场景图。用户可以从展示的网格展示图中选择其所需的网格展示图,响应于确认操作,则可以确定用户最终确定出的符合其需求的场景图。
由此,通过上述技术方案,可以通过展示候选场景图使得用户对能够生成的场景图进行大概预览,用户则可以初步进行选择,之后可以对用户选择的场景图进行网格化处理,以由用户进一步基于网格展示图使得用户了解相应的三维场景图的概况,以确定之后用于生成体素数据的场景图,从而可以有效降低生成体素数据的处理量,提升生成的体素数据的可用性,同时也可以提高与用户交互的多样性。
在一种可能的实施例中,响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据可以包括:
将目标场景图输入体素数据模型,获得目标场景图对应的体素数据,其中,体素数据模型是基于三维场景对应的体素数据和三维场景对应的场景截图数据进行训练获得的。
作为示例,可以基于已有场景的数据对体素数据模型进行训练。如可以基于已经搭建的场景对应的体素数据和该场景的场景截图数据作为训练数据进行训练。如可以将场景截图数据作为模型的输入,将场景对应的体素数据作为模型的目标输入对模型进行训练,以获得该体素数据模型。其中,该体素数据模型可以基于大语言模型或者神经网络模型实现。
相应地,可以将目标场景图输入体素数据模型,获得相应的体素数据。
其中,若目标场景图存在对应的网格展示图,则可以将网格展示图输入体素数据模型,若目标场景图未生成对应的网格展示图,则可以将该目标场景图输入体素数据模型,以获得体素数据。
由此,通过上述技术方案,可以通过体素数据模型快速生成场景图对应的体素数据,该过程中可以学习到已有场景的展示图和其搭建过程中的体素数据的关联性,提高体素数据的有效性和准确性,提高基于体素数据进行渲染所得的图像和目标场景图之间的一致性,保证用户使用体验。
在一种可能的实施例中,响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据的示例性实现方式可以包括:
获取目标场景图对应的网格展示图。
其中,若该目标场景图存在对应的网格展示图则可以直接获取,若该目标场景图尚未生成对应的网格展示图,则可以对该目标场景图进行像素化处理,以生成该网格展示图。
针对网格展示图中的每一网格,确定网格对应的体素块,获得展示体素数据。
作为示例,可以对目标场景图中的各个场景部分进行识别,如可以识别出其中的门框、墙壁、屋顶、地基等部分,则可以基于识别出的各个部分的类型确定其中的网格的属性。之后,可以将每一网格的属性与体素块库中的体素块的类型进行匹配,以确定该网格对应的体素块,其中可以将与该网格的属性相同或者相似度大于阈值的体素块作为该网格对应的体素块,从而可以获得该目标场景图中对用户展示的部分对应的展示体素数据。
之后针对目标场景图中未对用户展示的部分,可以基于该目标场景和展示体素数据进行进一步的填充。例如可以基于网格展示图确定目标场景图对应的尺寸信息。
作为示例,尺寸信息包含目标场景在水平面上对应的长度和宽度。相应地,基于网格展示图确定目标场景图对应的尺寸信息的示例性实现方式可以包括:
若网格展示图中包含目标场景的相邻的两个竖直面的展示图,则分别根据相邻的两个面的展示图确定目标场景在水平面上对应的长度和宽度。
如图5所示的图像,该网格展示图中包含目标场景相邻的两个竖直面的
展示图,即图5中的A面和B面。如可以将A面与水平面的交线的长度作为目标场景的长度,基于B面与水平面的交线的长度作为目标场景的宽度,其中,可以基于B面中的倾斜角度对交线的长度进行调整后作为目标场景的宽度,其中该调整方式可以基于本领域中通用的透视角度下的尺寸变化对宽度进行调整,在此不再赘述。
若网格展示图中只包含目标场景的一个竖直面的展示图,则基于竖直面的展示图确定竖直面对应的长度,并基于该竖直面对应的长度确定宽度。
若网格展示图中只包含目标场景的一个竖直面的展示图,即目标场景图正对显示屏幕进行展示,此时可以将展示图与水平面的交线的长度作为目标场景的长度。作为示例可以预先设置长度和宽度的默认比例,则在确定出长度之后,可以基于该长度和默认比例,确定对应的宽度,从而获得尺寸信息。
作为另一示例,用户可以在场景地图中预先设置用于进行场景生成的区域,则该长度和宽度需要处于用户选择的区域内,则可以基于该用户选择的区域对尺寸信息进行调整,即以用户选择的区域作为尺寸信息的最大范围对目标场景图对应的尺寸信息进行约束,如确定出的宽度超过用户选择的区域的宽度,则以用户选择的区域的宽度作为目标场景图的宽度。由此,通过上述技术方案,可以通过生成的目标场景图进一步地确定出其对应的尺寸信息,为后续对其进行三维结构的生成提供可靠的数据支持。
之后,基于网格展示图和尺寸信息对展示体素数据进行填充,获得目标场景图对应的体素数据。
作为示例,如图5所示,基于B面确定出的宽度为10,则可以基于该网格展示图正面的网格向图5所示S方向进行填充,其中可以基于默认体素块进行填充,也可以基于不同体素块之间的关联关系进行填充,其可以预先设置多种填充规则,如展示体素数据中体素块的类型为墙壁,则其S方向上相连的4个体素块也需要是墙壁类型,则对展示体素数据进行填充时,可以基于该规则进行相应的填充,场景的内部未展示的结构可以基于表面中已经展示的部分进行匹配,从而按照匹配到的结构进行填充。如展示体素数据中表示该场景为三层楼体结构,则内部结构中则可以匹配到楼体结构,则可以内部填充时填充楼梯结构对应的多个体素块的组合。又如,如展示体素数据中表示该场景顶层为尖顶结构,则可以在顶部对应的S方向上按照匹配到的尖
顶结构进行填充,从而获得目标场景图对应的体素数据。其中,可以内部填充时填充默认体素,后续用户可以对内部体素数据的类型进行更改和调整,实现个性化场景搭建,也可以进一步提高用户进行场景搭建的交互方式,简化用户进行场景搭建的流程。
在一种可能的实施例中,该场景生成方法还可以包括:
展示场景地图。其中,该场景地图可以为虚拟场景中允许进行场景搭建的区域地图,其可以基于预先配置确定。
响应于用户在场景地图中选择的区域信息,确定目标区域;
在目标区域中展示目标场景。
作为示例,用户可以在场景地图中选择其想要进行场景搭建的区域,如图6所示,可以通过平行四边形的方式进行区域选择。作为示例,可以首先移动到起点位置A,之后,移动到终点位置A’。其中,该区域信息可以包含该起点位置和终点位置,在确定起点和终点位置后,可以以起点至终点的直线作为对角线形成的中间区域作为目标区域。其中,可以基于默认的高生成该目标区域,也可以由用户进一步确定辅助的第三点,以第三点为顶点,以起点至终点的直线作为对角线形成的中间区域作为该目标区域。
由此,可以进一步在场景地图中确定目标区域,以便快速确定搭建的场景所处的环境,同时也可以基于目标区域所处的环境为用户进行场景描述信息的确定提高参考。
在一种可能的实施例中,基于场景描述信息,生成多个候选场景的候选场景图的实现方式可以包括:
将场景描述信息和目标区域输入场景生成模型,以获得与该目标区域对应的多个候选场景。
其中,区域的不同也可能会对场景的搭建有所影响,如目标场景为街道,则目标区域越小该街道场景中包含的建筑通常越少的,则在生成候选场景的候选场景图进一步结合该目标区域,以提高候选场景与目标区域之间的一致性。
在一种可能的实施例中,在目标区域中展示目标场景,包括:
若目标场景中包含多个场景单元,则在目标区域中显示多个场景单元。
其中,场景单元可以是根据实际应用场景进行预先设置的不可拆分的场
景建筑,如一个房屋、一栋楼。
若目标场景为街道,则可以由多个房屋和楼宇组合而成,其可以包含多个场景单元。针对该街道场景,在生成该街道场景时,是基于其选择中的目标区域内对场景单元进行组合以生成该目标场景,其组合中的场景单元是在目标区域进行生成的组合,则该实施例中可以直接在该目标区域内展示该多个场景单元。
若目标场景中包含一个场景单元,则响应于在目标区域中选择的目标位置,在目标位置显示目标场景。
作为示例,该目标场景为一栋楼,则其包含一个场景单元,其可以为独立的场景,则其可以在目标区域中的任意位置。该示例中,可以由用户在目标区域中指定该目标场景对应的目标位置,进而将其展示在该目标位置处。
由此,通过上述技术方案,针对建筑群或组合场景等类型的目标场景,可以直接目标区域中进行展示,无需用户进行位置配置,提高目标场景与目标区域之间的适配,针对单一场景单元,可以由用户对其展示位置进行调整,也可以在一定程度上提高场景展示位置的多样性,进一步提高场景生成的内容丰富度和交互度,提升用户使用体验。
基于同样的发明构思,本公开还提供一种场景生成装置,如图7所示,该场景生成装置10包括:接收模块100、第一生成模块200、第一展示模块300、第二生成模块400以及第一处理模块500。
接收模块100,被配置为接收待生成的目标场景的场景描述信息;
第一生成模块200,被配置为基于场景描述信息,生成多个候选场景的候选场景图,其中,候选场景图为二维图像;
第一展示模块300,被配置为展示候选场景图;
第二生成模块400,被配置为响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据;
第一处理模块500,被配置为基于体素数据进行渲染获得目标场景,目标场景为三维场景。
可选地,接收模块包括:
第一展示子模块,被配置为展示第一对话区域,第一对话区域中显示有多个候选场景标签;
第一处理子模块,被配置为响应于用户对候选场景标签的选择,将用户选择的候选场景标签显示在第二对话区域;
第一确定子模块,被配置为根据第二对话区域中的显示信息确定场景描述信息。
可选地,接收模块还包括:
接收子模块,被配置为接收用户在第二对话区域中的输入信息,输入信息包含文本和/或图像;
第二确定子模块,被配置为将输入信息和用户选择的候选场景标签对应的标签信息,作为场景描述信息。
可选地,第二生成模块包括:
第二处理子模块,被配置为响应于选择操作,将所选择的目标场景图进行网格化处理,获得目标场景图对应的网格展示图;
第二展示子模块,被配置为展示网格展示图;
响应于接收到对网格展示图的确认操作,生成所确认的目标场景图对应的体素数据,所确认的目标场景图为确认操作指示的网格展示图对应的目标场景图。
可选地,第二生成模块包括:
第三处理子模块,被配置为将目标场景图输入体素数据模型,获得目标场景图对应的体素数据,其中,体素数据模型是基于三维场景对应的体素数据和三维场景对应的场景截图数据进行训练获得的。
可选地,第二生成模块包括:
获取子模块,被配置为获取目标场景图对应的网格展示图;
第三确定子模块,被配置为针对网格展示图中的每一网格,确定网格对应的体素块,获得展示体素数据;
第四确定子模块,被配置为基于网格展示图确定目标场景图对应的尺寸信息;
第四处理子模块,被配置为基于网格展示图和尺寸信息对展示体素数据进行填充,获得目标场景图对应的体素数据。
可选地,尺寸信息包含目标场景在水平面上对应的长度和宽度;
第四确定子模块包括:
第五确定子模块,被配置为若网格展示图中包含目标场景的相邻的两个竖直面的展示图,则分别根据相邻的两个面的展示图确定目标场景在水平面上对应的长度和宽度;
第六确定子模块,被配置为若网格展示图中只包含目标场景的一个竖直面的展示图,则基于竖直面的展示图确定竖直面对应的长度,并基于该竖直面对应的长度确定宽度。
可选地,该场景生成装置还包括:
第二展示模块,被配置为展示场景地图;
第一确定模块,被配置为响应于用户在场景地图中选择的区域信息,确定目标区域;
第三展示模块,被配置为在目标区域中展示目标场景。
可选地,第三展示模块包括:
第三展示子模块,被配置为若目标场景中包含多个场景单元,则在目标区域中显示多个场景单元;
第四展示子模块,被配置为若目标场景中包含一个场景单元,则响应于在目标区域中选择的目标位置,在目标位置显示目标场景。
可选地,第一展示模块包括:
第五展示子模块,被配置为针对候选场景图,在该候选场景图对应的第一展示区域中展示该候选场景图对应的编辑控件和选择控件;
该场景生成装置还包括:
第四展示模块,被配置为响应于接收到对编辑控件的选择操作,在编辑界面中展示选择的目标编辑控件对应的候选场景图;
第二处理模块,被配置为响应于用户在编辑界面中输入的编辑信息,根据编辑信息和目标编辑控件对应的候选场景图,生成新的候选场景图并展示。
可选地,第一展示模块包括:
第七确定子模块,被配置为从候选场景图中确定出展示场景图;
第六展示子模块,被配置为对展示场景图进行展示;
该场景生成装置还包括:
第五展示模块,被配置为在对展示场景图进行展示时,展示场景更新控件;
第二确定模块,被配置为响应于接收到对场景更新控件的选择操作,确定新的展示场景图;
第六展示模块,被配置为展示新的展示场景图。
可选地,该场景生成装置还包括:
第七展示模块,被配置为在展示候选场景图时展示场景编辑控件;
第八展示模块,被配置为响应于接收到对场景编辑控件的选择操作,展示场景编辑界面;
第三处理模块,被配置为响应于用户在场景编辑界面中的编辑操作,获得新的场景描述信息,并触发第一生成模块执行基于场景描述信息,生成多个候选场景的候选场景图的操作。
可选地,第一生成模块包括:
第五处理子模块,被配置为根据场景描述信息构建提示文本;
生成子模块,被配置为将提示文本输入场景生成模型,根据场景生成模型的输出获得多个候选场景的候选场景图,其中,场景生成模型基于大语言模型实现。
下面参考图8,其示出了适于用来实现本公开实施例的电子设备(例如终端设备或服务器)600的结构示意图。本公开实施例中的终端设备可以包括但不限于诸如移动电话、笔记本电脑、数字广播接收器、PDA(个人数字助理)、PAD(平板电脑)、PMP(便携式多媒体播放器)、车载终端(例如车载导航终端)等等的移动终端以及诸如数字TV、台式计算机等等的固定终端。图8示出的电子设备仅仅是一个示例,不应对本公开实施例的功能和使用范围带来任何限制。
如图8所示,电子设备600可以包括处理装置(例如中央处理器、图形处理器等)601,其可以根据存储在只读存储器(ROM)602中的程序或者从存储装置608加载到随机访问存储器(RAM)603中的程序而执行各种适当的动作和处理。在RAM 603中,还存储有电子设备600操作所需的各种程序和数据。处理装置601、ROM 602以及RAM 603通过总线604彼此相连。输入/输出(I/O)接口605也连接至总线604。
通常,以下装置可以连接至I/O接口605:包括例如触摸屏、触摸板、键盘、鼠标、摄像头、麦克风、加速度计、陀螺仪等的输入装置606;包括例如
液晶显示器(LCD)、扬声器、振动器等的输出装置607;包括例如磁带、硬盘等的存储装置608;以及通信装置609。通信装置609可以允许电子设备600与其他设备进行无线或有线通信以交换数据。虽然图8示出了具有各种装置的电子设备600,但是应理解的是,并不要求实施或具备所有示出的装置。可以替代地实施或具备更多或更少的装置。
特别地,根据本公开的实施例,上文参考流程图描述的过程可以被实现为计算机软件程序。例如,本公开的实施例包括一种计算机程序产品,其包括承载在非暂态计算机可读介质上的计算机程序,该计算机程序包含用于执行流程图所示的方法的程序代码。在这样的实施例中,该计算机程序可以通过通信装置609从网络上被下载和安装,或者从存储装置608被安装,或者从ROM 602被安装。在该计算机程序被处理装置601执行时,执行本公开实施例的方法中限定的上述功能。
需要说明的是,本公开上述的计算机可读介质可以是计算机可读信号介质或者计算机可读存储介质或者是上述两者的任意组合。计算机可读存储介质例如可以是——但不限于——电、磁、光、电磁、红外线、或半导体的系统、装置或器件,或者任意以上的组合。计算机可读存储介质的更具体的例子可以包括但不限于:具有一个或多个导线的电连接、便携式计算机磁盘、硬盘、随机访问存储器(RAM)、只读存储器(ROM)、可擦式可编程只读存储器(EPROM或闪存)、光纤、便携式紧凑磁盘只读存储器(CD-ROM)、光存储器件、磁存储器件、或者上述的任意合适的组合。在本公开中,计算机可读存储介质可以是任何包含或存储程序的有形介质,该程序可以被指令执行系统、装置或者器件使用或者与其结合使用。而在本公开中,计算机可读信号介质可以包括在基带中或者作为载波一部分传播的数据信号,其中承载了计算机可读的程序代码。这种传播的数据信号可以采用多种形式,包括但不限于电磁信号、光信号或上述的任意合适的组合。计算机可读信号介质还可以是计算机可读存储介质以外的任何计算机可读介质,该计算机可读信号介质可以发送、传播或者传输用于由指令执行系统、装置或者器件使用或者与其结合使用的程序。计算机可读介质上包含的程序代码可以用任何适当的介质传输,包括但不限于:电线、光缆、RF(射频)等等,或者上述的任意合适的组合。
在一些实施方式中,客户端、服务器可以利用诸如HTTP(HyperText Transfer Protocol,超文本传输协议)之类的任何当前已知或未来研发的网络协议进行通信,并且可以与任意形式或介质的数字数据通信(例如,通信网络)互连。通信网络的示例包括局域网(“LAN”),广域网(“WAN”),网际网(例如,互联网)以及端对端网络(例如,ad hoc端对端网络),以及任何当前已知或未来研发的网络。
上述计算机可读介质可以是上述电子设备中所包含的;也可以是单独存在,而未装配入该电子设备中。
上述计算机可读介质承载有一个或者多个程序,当上述一个或者多个程序被该电子设备执行时,使得该电子设备:接收待生成的目标场景的场景描述信息;基于场景描述信息,生成多个候选场景的候选场景图,其中,候选场景图为二维图像;展示候选场景图;响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据;基于体素数据进行渲染获得目标场景,目标场景为三维场景。
可以以一种或多种程序设计语言或其组合来编写用于执行本公开的操作的计算机程序代码,上述程序设计语言包括但不限于面向对象的程序设计语言—诸如Java、Smalltalk、C++,还包括常规的过程式程序设计语言——诸如“C”语言或类似的程序设计语言。程序代码可以完全地在用户计算机上执行、部分地在用户计算机上执行、作为一个独立的软件包执行、部分在用户计算机上部分在远程计算机上执行、或者完全在远程计算机或服务器上执行。在涉及远程计算机的情形中,远程计算机可以通过任意种类的网络——包括局域网(LAN)或广域网(WAN)——连接到用户计算机,或者,可以连接到外部计算机(例如利用因特网服务提供商来通过因特网连接)。
附图中的流程图和框图,图示了按照本公开各种实施例的系统、方法和计算机程序产品的可能实现的体系架构、功能和操作。在这点上,流程图或框图中的每个方框可以代表一个模块、程序段、或代码的一部分,该模块、程序段、或代码的一部分包含一个或多个用于实现规定的逻辑功能的可执行指令。也应当注意,在有些作为替换的实现中,方框中所标注的功能也可以以不同于附图中所标注的顺序发生。例如,两个接连地表示的方框实际上可以基本并行地执行,它们有时也可以按相反的顺序执行,这依所涉及的功能
而定。也要注意的是,框图和/或流程图中的每个方框、以及框图和/或流程图中的方框的组合,可以用执行规定的功能或操作的专用的基于硬件的系统来实现,或者可以用专用硬件与计算机指令的组合来实现。
描述于本公开实施例中所涉及到的模块可以通过软件的方式实现,也可以通过硬件的方式来实现。其中,模块的名称在某种情况下并不构成对该模块本身的限定,例如,接收模块还可以被描述为“接收待生成的目标场景的场景描述信息的模块”。
本文中以上描述的功能可以至少部分地由一个或多个硬件逻辑部件来执行。例如,非限制性地,可以使用的示范类型的硬件逻辑部件包括:现场可编程门阵列(FPGA)、专用集成电路(ASIC)、专用标准产品(ASSP)、片上系统(SOC)、复杂可编程逻辑设备(CPLD)等等。
在本公开的上下文中,机器可读介质可以是有形的介质,其可以包含或存储以供指令执行系统、装置或设备使用或与指令执行系统、装置或设备结合地使用的程序。机器可读介质可以是机器可读信号介质或机器可读储存介质。机器可读介质可以包括但不限于电子的、磁性的、光学的、电磁的、红外的、或半导体系统、装置或设备,或者上述内容的任何合适组合。机器可读存储介质的更具体示例会包括基于一个或多个线的电气连接、便携式计算机盘、硬盘、随机存取存储器(RAM)、只读存储器(ROM)、可擦除可编程只读存储器(EPROM或快闪存储器)、光纤、便捷式紧凑盘只读存储器(CD-ROM)、光学储存设备、磁储存设备、或上述内容的任何合适组合。
根据本公开的一个或多个实施例,示例1提供了一种场景生成方法,该场景生成方法包括:
接收待生成的目标场景的场景描述信息;
基于场景描述信息,生成多个候选场景的候选场景图,其中,候选场景图为基于候选场景对应的体素数据所生成的二维图像;
展示候选场景图;
响应于接收到用户针对候选场景图的选择操作,根据选择的目标场景图对应的体素数据进行渲染获得目标场景,目标场景为三维场景。
根据本公开的一个或多个实施例,示例2提供了示例1的方法,其中,接收待生成的目标场景的场景描述信息,包括:
展示第一对话区域,第一对话区域中显示有多个候选场景标签;
响应于用户对候选场景标签的选择,将用户选择的候选场景标签显示在第二对话区域;
根据第二对话区域中的显示信息确定场景描述信息。
根据本公开的一个或多个实施例,示例3提供了示例2的方法,其中,接收待生成的目标场景的场景描述信息,还包括:
接收用户在第二对话区域中的输入信息,输入信息包含文本和/或图像;
将输入信息和用户选择的候选场景标签对应的标签信息,作为场景描述信息。
根据本公开的一个或多个实施例,示例4提供了示例1的方法,其中,响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据,包括:
响应于选择操作,将所选择的目标场景图进行网格化处理,获得目标场景图对应的网格展示图;
展示网格展示图;
响应于接收到对网格展示图的确认操作,生成所确认的目标场景图对应的体素数据,所确认的目标场景图为确认操作指示的网格展示图对应的目标场景图。
根据本公开的一个或多个实施例,示例5提供了示例1的方法,其中,响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据,包括:
将目标场景图输入体素数据模型,获得目标场景图对应的体素数据,其中,体素数据模型是基于三维场景对应的体素数据和三维场景对应的场景截图数据进行训练获得的。
根据本公开的一个或多个实施例,示例6提供了示例1的方法,其中,响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据,包括:
获取目标场景图对应的网格展示图;
针对网格展示图中的每一网格,确定网格对应的体素块,获得展示体素数据;
基于网格展示图确定目标场景图对应的尺寸信息;
基于网格展示图和尺寸信息对展示体素数据进行填充,获得目标场景图对应的体素数据。
根据本公开的一个或多个实施例,示例7提供了示例6的方法,其中,尺寸信息包含目标场景在水平面上对应的长度和宽度;
基于网格展示图确定目标场景图对应的尺寸信息,包括:
若网格展示图中包含目标场景的相邻的两个竖直面的展示图,则分别根据相邻的两个面的展示图确定目标场景在水平面上对应的长度和宽度;
若网格展示图中只包含目标场景的一个竖直面的展示图,则基于竖直面的展示图确定竖直面对应的长度,并基于该竖直面对应的长度确定宽度。
根据本公开的一个或多个实施例,示例8提供了示例1的方法,其中,该场景生成方法还包括:
展示场景地图;
响应于用户在场景地图中选择的区域信息,确定目标区域;
在目标区域中展示目标场景。
根据本公开的一个或多个实施例,示例9提供了示例8的方法,其中,在目标区域中展示目标场景,包括:
若目标场景中包含多个场景单元,则在目标区域中显示多个场景单元;
若目标场景中包含一个场景单元,则响应于在目标区域中选择的目标位置,在目标位置显示目标场景。
根据本公开的一个或多个实施例,示例10提供了示例1的方法,其中,展示候选场景图,包括:
针对候选场景图,在该候选场景图对应的第一展示区域中展示该候选场景图对应的编辑控件和选择控件;
该场景生成方法还包括:
响应于接收到对编辑控件的选择操作,在编辑界面中展示选择的目标编辑控件对应的候选场景图;
响应于用户在编辑界面中输入的编辑信息,根据编辑信息和目标编辑控件对应的候选场景图,生成新的候选场景图并展示。
根据本公开的一个或多个实施例,示例11提供了示例1的方法,其中,
展示候选场景图,包括:
从候选场景图中确定出展示场景图;
对展示场景图进行展示;
该场景生成方法还包括:
在对展示场景图进行展示时,展示场景更新控件;
响应于接收到对场景更新控件的选择操作,确定新的展示场景图;
展示新的展示场景图。
根据本公开的一个或多个实施例,示例12提供了示例1的方法,其中,该场景生成方法还包括:
在展示候选场景图时展示场景编辑控件;
响应于接收到对场景编辑控件的选择操作,展示场景编辑界面;
响应于用户在场景编辑界面中的编辑操作,获得新的场景描述信息,并返回执行基于场景描述信息,生成多个候选场景的候选场景图的操作。
根据本公开的一个或多个实施例,示例13提供了示例1的方法,其中,基于场景描述信息,生成多个候选场景的候选场景图,包括:
根据场景描述信息构建提示文本;
将提示文本输入场景生成模型,根据场景生成模型的输出获得多个候选场景的候选场景图,其中,场景生成模型基于大语言模型实现。
根据本公开的一个或多个实施例,示例14提供了一种场景生成装置,该场景生成装置包括:接收模块,被配置为接收待生成的目标场景的场景描述信息;第一生成模块,被配置为基于场景描述信息,生成多个候选场景的候选场景图,其中,候选场景图为二维图像;第一展示模块,被配置为展示候选场景图;第二生成模块,被配置为响应于接收到用户针对候选场景图的选择操作,生成所选择的目标场景图对应的体素数据;第一处理模块,被配置为基于体素数据进行渲染获得目标场景,目标场景为三维场景。
根据本公开的一个或多个实施例,示例15提供了一种计算机可读介质,其上存储有计算机程序,该程序被处理装置执行时实现示例1-13中任一项涉及的场景生成方法的步骤。
根据本公开的一个或多个实施例,示例16提供了一种电子设备,包括:存储装置,其上存储有计算机程序;处理装置,被配置为执行存储装置中的
计算机程序,以实现示例1-13中任一项涉及的场景生成方法的步骤。
根据本公开的一个或多个实施例,示例17提供了一种计算机程序产品,包括计算机程序,该计算机程序被处理器执行时实现示例1-13中任一项的方法的步骤。
以上描述仅为本公开的较佳实施例以及对所运用技术原理的说明。本领域技术人员应当理解,本公开中所涉及的公开范围,并不限于上述技术特征的特定组合而成的技术方案,同时也应涵盖在不脱离上述公开构思的情况下,由上述技术特征或其等同特征进行任意组合而形成的其它技术方案。例如上述特征与本公开中公开的(但不限于)具有类似功能的技术特征进行互相替换而形成的技术方案。
此外,虽然采用特定次序描绘了各操作,但是这不应当理解为要求这些操作以所示出的特定次序或以顺序次序执行来执行。在一定环境下,多任务和并行处理可能是有利的。同样地,虽然在上面论述中包含了若干具体实现细节,但是这些不应当被解释为对本公开的范围的限制。在单独的实施例的上下文中描述的某些特征还可以组合地实现在单个实施例中。相反地,在单个实施例的上下文中描述的各种特征也可以单独地或以任何合适的子组合的方式实现在多个实施例中。
尽管已经采用特定于结构特征和/或方法逻辑动作的语言描述了本主题,但是应当理解所附权利要求书中所限定的主题未必局限于上面描述的特定特征或动作。相反,上面所描述的特定特征和动作仅仅是实现权利要求书的示例形式。关于上述实施例中的装置,其中各个模块执行操作的具体方式已经在有关该方法的实施例中进行了详细描述,此处将不做详细阐述说明。
Claims (17)
- 一种场景生成方法,包括:接收待生成的目标场景的场景描述信息;基于所述场景描述信息,生成多个候选场景的候选场景图,其中,所述候选场景图为二维图像;展示所述候选场景图;响应于接收到用户针对所述候选场景图的选择操作,生成所选择的目标场景图对应的体素数据;基于所述体素数据进行渲染获得所述目标场景,其中,所述目标场景为三维场景。
- 根据权利要求1所述的方法,其中,所述接收待生成的目标场景的场景描述信息,包括:展示第一对话区域,所述第一对话区域中显示有多个候选场景标签;响应于所述用户对所述候选场景标签的选择,将用户选择的候选场景标签显示在第二对话区域;根据所述第二对话区域中的显示信息确定所述场景描述信息。
- 根据权利要求2所述的方法,其中,所述接收待生成的目标场景的场景描述信息,还包括:接收所述用户在所述第二对话区域中的输入信息,所述输入信息包含文本和/或图像;将所述输入信息和所述用户选择的候选场景标签对应的标签信息,作为所述场景描述信息。
- 根据权利要求1所述的方法,其中,所述响应于接收到用户针对所述候选场景图的选择操作,生成所选择的目标场景图对应的体素数据,包括:响应于所述选择操作,将所选择的目标场景图进行网格化处理,获得所述目标场景图对应的网格展示图;展示所述网格展示图;响应于接收到对所述网格展示图的确认操作,生成所确认的目标场景图对应的体素数据,其中,所确认的目标场景图为所述确认操作指示的网格展示图对应的目标场景图。
- 根据权利要求1所述的方法,其中,所述响应于接收到用户针对所述候选场景图的选择操作,生成所选择的目标场景图对应的体素数据,包括:将所述目标场景图输入体素数据模型,获得所述目标场景图对应的体素数据,其中,所述体素数据模型是基于三维场景对应的体素数据和所述三维场景对应的场景截图数据进行训练获得的。
- 根据权利要求1所述的方法,其中,所述响应于接收到用户针对所述候选场景图的选择操作,生成所选择的目标场景图对应的体素数据,包括:获取所述目标场景图对应的网格展示图;针对所述网格展示图中的每一网格,确定所述网格对应的体素块,获得展示体素数据;基于所述网格展示图确定所述目标场景图对应的尺寸信息;基于网格展示图和所述尺寸信息对所述展示体素数据进行填充,获得目标场景图对应的体素数据。
- 根据权利要求6所述的方法,其中,所述尺寸信息包含目标场景在水平面上对应的长度和宽度;所述基于所述网格展示图确定所述目标场景图对应的尺寸信息,包括:若所述网格展示图中包含所述目标场景的相邻的两个竖直面的展示图,则分别根据所述相邻的两个面的展示图确定所述目标场景在水平面上对应的长度和宽度;若所述网格展示图中只包含所述目标场景的一个竖直面的展示图,则基于所述竖直面的展示图确定所述竖直面对应的长度,并基于该竖直面对应的长度确定宽度。
- 根据权利要求1所述的方法,还包括:展示场景地图;响应于所述用户在所述场景地图中选择的区域信息,确定目标区域;在所述目标区域中展示所述目标场景。
- 根据权利要求8所述的方法,其中,所述在所述目标区域中展示所述目标场景,包括:若所述目标场景中包含多个场景单元,则在所述目标区域中显示所述多个场景单元;若所述目标场景中包含一个场景单元,则响应于在所述目标区域中选择的目标位置,在所述目标位置显示所述目标场景。
- 根据权利要求1所述的方法,其中,所述展示所述候选场景图,包括:针对所述候选场景图,在该候选场景图对应的第一展示区域中展示该候选场景图对应的编辑控件和选择控件;所述方法还包括:响应于接收到对所述编辑控件的选择操作,在编辑界面中展示选择的目标编辑控件对应的候选场景图;响应于所述用户在所述编辑界面中输入的编辑信息,根据所述编辑信息和所述目标编辑控件对应的候选场景图,生成新的候选场景图并展示。
- 根据权利要求1所述的方法,其中,所述展示所述候选场景图,包括:从所述候选场景图中确定出展示场景图;对所述展示场景图进行展示;所述方法还包括:在对所述展示场景图进行展示时,展示场景更新控件;响应于接收到对所述场景更新控件的选择操作,确定新的展示场景图;展示所述新的展示场景图。
- 根据权利要求1所述的方法,还包括:在展示所述候选场景图时展示场景编辑控件;响应于接收到对所述场景编辑控件的选择操作,展示场景编辑界面;响应于所述用户在所述场景编辑界面中的编辑操作,获得新的场景描述信息,并返回执行所述基于所述场景描述信息,生成多个候选场景的候选场景图的操作。
- 根据权利要求1所述的方法,其中,所述基于所述场景描述信息,生成多个候选场景的候选场景图,包括:根据所述场景描述信息构建提示文本;将所述提示文本输入场景生成模型,根据所述场景生成模型的输出获得所述多个候选场景的候选场景图,其中,所述场景生成模型基于大语言模型实现。
- 一种场景生成装置,包括:接收模块,被配置为接收待生成的目标场景的场景描述信息;第一生成模块,被配置为基于所述场景描述信息,生成多个候选场景的候选场景图,其中,所述候选场景图为二维图像;第一展示模块,被配置为展示所述候选场景图;第二生成模块,被配置为响应于接收到用户针对所述候选场景图的选择操作,生成所选择的目标场景图对应的体素数据;以及第一处理模块,被配置为基于所述体素数据进行渲染获得所述目标场景,其中,所述目标场景为三维场景。
- 一种计算机可读介质,其上存储有计算机程序,其中,该程序被处理装置执行时实现权利要求1-13中任一项所述方法的步骤。
- 一种电子设备,包括:存储装置,其上存储有计算机程序;以及处理装置,被配置为执行所述存储装置中的所述计算机程序,以实现权利要求1-13中任一项所述方法的步骤。
- 一种计算机程序产品,包括计算机程序,其中,该计算机程序被处理器执行时实现权利要求1-13中任一项所述的方法的步骤。
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410371010.X | 2024-03-28 | ||
| CN202410371010.XA CN118079378A (zh) | 2024-03-28 | 2024-03-28 | 场景生成方法、装置、介质、设备和程序产品 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| WO2025200212A1 true WO2025200212A1 (zh) | 2025-10-02 |
Family
ID=91145650
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| PCT/CN2024/108808 Pending WO2025200212A1 (zh) | 2024-03-28 | 2024-07-31 | 场景生成方法、装置、介质、设备和程序产品 |
Country Status (2)
| Country | Link |
|---|---|
| CN (1) | CN118079378A (zh) |
| WO (1) | WO2025200212A1 (zh) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118079378A (zh) * | 2024-03-28 | 2024-05-28 | 北京字跳网络技术有限公司 | 场景生成方法、装置、介质、设备和程序产品 |
| CN119937879A (zh) * | 2025-01-10 | 2025-05-06 | 优酷文化科技(北京)有限公司 | 针对资产的使用预览方法及装置 |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116416404A (zh) * | 2023-02-23 | 2023-07-11 | 抖音视界有限公司 | 虚拟场景的生成方法、装置、可读介质及电子设备 |
| CN117008795A (zh) * | 2023-01-12 | 2023-11-07 | 杭州群核信息技术有限公司 | 建筑场景渲染方法、装置和存储介质 |
| CN117274489A (zh) * | 2023-09-21 | 2023-12-22 | 北京字跳网络技术有限公司 | 一种场景生成方法、装置、计算机设备及存储介质 |
| CN118079378A (zh) * | 2024-03-28 | 2024-05-28 | 北京字跳网络技术有限公司 | 场景生成方法、装置、介质、设备和程序产品 |
-
2024
- 2024-03-28 CN CN202410371010.XA patent/CN118079378A/zh active Pending
- 2024-07-31 WO PCT/CN2024/108808 patent/WO2025200212A1/zh active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN117008795A (zh) * | 2023-01-12 | 2023-11-07 | 杭州群核信息技术有限公司 | 建筑场景渲染方法、装置和存储介质 |
| CN116416404A (zh) * | 2023-02-23 | 2023-07-11 | 抖音视界有限公司 | 虚拟场景的生成方法、装置、可读介质及电子设备 |
| CN117274489A (zh) * | 2023-09-21 | 2023-12-22 | 北京字跳网络技术有限公司 | 一种场景生成方法、装置、计算机设备及存储介质 |
| CN118079378A (zh) * | 2024-03-28 | 2024-05-28 | 北京字跳网络技术有限公司 | 场景生成方法、装置、介质、设备和程序产品 |
Also Published As
| Publication number | Publication date |
|---|---|
| CN118079378A (zh) | 2024-05-28 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2025200212A1 (zh) | 场景生成方法、装置、介质、设备和程序产品 | |
| US8345043B2 (en) | Solving networks of geometric constraints | |
| JP2014504384A (ja) | 2次元画像からの3次元バーチャルツアーの生成 | |
| CN111079215B (zh) | 坡道工程量的确定方法、装置和电子设备 | |
| CN114742981B (zh) | 后处理特效的制作系统、方法、ar特效渲染方法及装置 | |
| CN113205601B (zh) | 漫游路径生成方法、装置、存储介质及电子设备 | |
| US20220358258A1 (en) | Computer-aided design methods and systems | |
| JP2025516368A (ja) | ビュー生成方法、装置、電子機器、記憶媒体、及びコンピュータプログラム | |
| JP2025537558A (ja) | 仮想ギフト生成方法、装置、機器、媒体及びコンピュータプログラム | |
| WO2023221962A1 (zh) | 视图生成方法、装置、电子设备及存储介质 | |
| CN120372745A (zh) | 基于三维分析的建筑模块化组合生成方法、系统及介质 | |
| CN118627486A (zh) | 表格数据处理方法、装置、设备及存储介质 | |
| CN114265583A (zh) | Vr内容开发优化方法、装置、电子设备及可读存储介质 | |
| CN114020977A (zh) | 建筑信息模型交互方法及相关装置 | |
| WO2024051639A1 (zh) | 图像处理方法、装置、设备及存储介质和产品 | |
| US12197843B2 (en) | Automatic layout of elements in a process flow on a 2-D canvas based on representations of flow logic | |
| US20250199661A1 (en) | Information display method, medium, and electronic device | |
| US20240412436A1 (en) | Effect edition method and apparatus, electronic device and storage medium | |
| CN110490926B (zh) | 标高确定方法、装置和电子设备 | |
| CN119668763A (zh) | 虚拟对象布局方法、装置、介质、电子设备及程序产品 | |
| WO2024152924A1 (zh) | 虚拟对象的处理方法、装置、设备及存储介质 | |
| US20240073460A1 (en) | Method and apparatus for displaying information, electronic device and computer readable medium | |
| EP4671965A1 (en) | METHOD AND APPARATUS FOR MANAGING NAVIGATION TREE VIEW, DEVICE AND STORAGE MEDIA | |
| WO2024188089A1 (zh) | 特效信息显示方法、装置、电子设备及存储介质 | |
| CN114022601A (zh) | 一种体积元素渲染方法、装置及设备 |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| 121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 24932491 Country of ref document: EP Kind code of ref document: A1 |