Scene generation system based on artificial intelligence

Abstract

The invention provides a scene generation system based on artificial intelligence. The system comprises a scene understanding module, a scene construction module, a rendering optimization module and a scene adjustment module. The scene understanding module is used for analyzing key elements, layout information and rendering information for scene generation from user input information; the scene construction module is used for constructing a three-dimensional scene according to the information provided by the scene understanding unit; the rendering optimization module is used for performing rendering processing on the three-dimensional scene constructed by the scene construction module to generate a final three-dimensional scene; the scene adjusting module is used for changing and adjusting the finally generated three-dimensional scene according to user requirements; according to the method, the automatic process from user input to high-quality three-dimensional scene generation is realized, the scene design efficiency and quality are greatly improved, and meanwhile, the personalized customization and interactive experience capabilities of the user are enhanced.

Description

Technical Field

[0001] The present invention relates to the field of scene generation systems, and in particular to a scene generation system based on artificial intelligence. Background Art

[0002] In the rapid development of the digital age, 3D scene generation technology has become a key driving force in the fields of virtual reality, game development, film and television production and even urban planning. With the advancement of technology, the demand for more realistic, efficient and highly personalized 3D scenes is growing. Traditional manual scene design methods are time-consuming, costly and difficult to adapt to rapidly changing needs, and have gradually failed to meet the requirements of modern society for 3D scene generation. Therefore, exploring a new method that can automatically generate accurate and realistic 3D scenes while allowing users to customize and adjust them according to their personal preferences has become an important goal of current scientific and technological development.

[0003] After consulting the relevant disclosed technical solutions, the technology with publication number CN117273006A proposes an artificial intelligence content generation method and system. By performing a semantic analysis check before content generation, multiple extensible items are obtained, and multiple versions are provided for subsequent content generation. After the content is generated, a security check is performed to ensure that the content generated by artificial intelligence meets the grammatical logic requirements. At the same time, the generated content is predicted to determine whether its possible direction contains extensible items, and the direction containing the most extensible items is selected to continue generating content. Finally, the content is output, which overcomes the problem that the content generated by the prior art is not of high quality and not diverse enough. However, the solution may lack the ability to understand and personalize user generation needs, especially in terms of finely capturing users' specific element layouts, style preferences, and dynamic adjustment needs for three-dimensional scenes. It cannot provide high depth and flexibility. Summary of the invention

[0004] The purpose of the present invention is to propose a scene generation system based on artificial intelligence to address the current deficiencies.

[0005] The present invention adopts the following technical solution:

[0006] A scene generation system based on artificial intelligence, the system comprising a scene understanding module, a scene construction module, a rendering optimization module and a scene adjustment module;

[0007] The scene understanding module is used to parse the key elements, layout information and rendering information for scene generation from the user input information; the scene construction module is used to construct a three-dimensional scene according to the information provided by the scene understanding unit; the rendering optimization module is used to render the three-dimensional scene constructed by the scene construction module to generate a final three-dimensional scene; the scene adjustment module is used to modify and adjust the final generated three-dimensional scene according to user needs;

[0008] The scene understanding module includes a data receiving unit, a key element extraction unit and a layout understanding unit; the data receiving unit is used to receive various forms of input information provided by the user; the key element extraction unit is used to analyze and process the received input information, and extract key elements, layout information and rendering information for scene generation; the layout understanding unit is used to establish the spatial relationship between the key elements to provide a blueprint for subsequent scene construction;

[0009] The scene construction module includes a three-dimensional object generation unit and a scene layout adjustment unit; the three-dimensional object generation unit is used to generate three-dimensional object models corresponding to each key element provided by the scene understanding module; the scene layout adjustment unit is used to complete the spatial position adjustment and fusion between the three-dimensional object models;

[0010] Furthermore, the key elements extracted by the key element extraction unit also include description clarity of each key element, and the description clarity of each key element is obtained by obtaining the attribute integrity and semantic clarity of each key element;

[0011] Further, the layout understanding unit includes an element layout subunit, an element connection word library and a conflict element adjustment subunit, wherein the element layout subunit is used to generate a preliminary layout of each key element according to each key element and its layout information; the element connection word library is a preset database, which contains the quantitatively expressed connection degree between different key elements; the conflict element adjustment subunit is used to identify and resolve the layout conflicts in the preliminary layout of each key element by combining the description clarity and connection degree of each key element, and generate each adjusted key element and its final layout;

[0012] Furthermore, the three-dimensional object generation unit dynamically generates three-dimensional object models corresponding to the key elements provided by the conflict element adjustment subunit using a generative adversarial network; in the process of generating the three-dimensional object models corresponding to the key elements, the detail level of the three-dimensional object model is adjusted according to the description clarity of each key element, so as to ensure that the generated three-dimensional object model meets the description input by the user;

[0013] Furthermore, the scene layout adjustment unit adjusts and merges the spatial positions of the generated three-dimensional object models according to the final layout provided by the conflict element adjustment subunit; this includes adjusting the relative positions and angles between the three-dimensional object models in combination with geometric constraints and simulated physical rules, thereby completing the construction of the three-dimensional scene;

[0014] Furthermore, the rendering optimization module renders the three-dimensional scene constructed by the scene layout adjustment unit through ray tracing and physically based rendering technology based on the rendering information provided by the scene understanding module, thereby generating a final three-dimensional scene;

[0015] Furthermore, the scene adjustment module includes a scene feedback update unit and a rendering feedback update unit; the scene feedback update unit is used to receive user feedback information on the generated scene and perform modification operations on the generated scene; the rendering update unit is used to receive user rendering feedback information and perform rendering operations on the modified scene to generate a modified three-dimensional scene.

[0016] The beneficial effects achieved by the present invention are:

[0017] The present invention accurately parses the user's input information to capture the key elements, layout information and rendering information used for scene generation, thereby ensuring a seamless transition from user needs to scene design; by refining the description clarity and connection degree of each key element to resolve the layout conflicts of the key elements in the scene, the generation of the three-dimensional scene can be completed in more detail based on the user's understanding; by receiving user feedback and adjusting the scene based on the user's feedback information, the user's sense of participation and satisfaction in the scene design is enhanced, thereby ensuring that the final three-dimensional scene fully meets the user's expectations. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] The present invention can be further understood from the following description in conjunction with the accompanying drawings. The components in the figures are not necessarily drawn to scale, but the emphasis is placed on illustrating the principles of the embodiments. In different views, the same reference numerals designate corresponding parts.

[0019] Figure 1 It is a schematic diagram of the overall module of the present invention.

[0020] Figure 2 The figure is a flow chart of a method for obtaining the description clarity of the key elements of the present invention.

[0021] Figure 3 Schematic diagram of the conflict element adjustment subunit workflow of the present invention.

[0022] Figure 4 It is a schematic diagram of the workflow of the layout update unit of the present invention. DETAILED DESCRIPTION

[0023] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with its embodiments; it should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention; for those skilled in the art, other systems, methods and / or features of the present embodiment will become apparent after reviewing the following detailed description; it is intended that all such additional systems, methods, features and advantages are included in this specification; included within the scope of the present invention and protected by the appended claims; additional features of the disclosed embodiments are described in the following detailed description, and these features will be apparent from the following detailed description.

[0024] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right" and the like indicate directions or positional relationships based on the directions or positional relationships shown in the drawings, it is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or component referred to must have a specific direction, be constructed and operated in a specific direction. Therefore, the terms describing the positional relationship in the drawings are only used for illustrative purposes and cannot be understood as limitations on this patent. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0025] Embodiment 1:

[0026] like Figure 1 As shown, this embodiment provides a scene generation system based on artificial intelligence, the system includes a scene understanding module, a scene construction module, a rendering optimization module and a scene adjustment module;

[0027] The scene understanding module is used to parse the key elements, layout information and rendering information for scene generation from the user input information; the scene construction module is used to construct a three-dimensional scene according to the information provided by the scene understanding unit; the rendering optimization module is used to render the three-dimensional scene constructed by the scene construction module to generate a final three-dimensional scene; the scene adjustment module is used to modify and adjust the final generated three-dimensional scene according to user needs;

[0028] The scene understanding module includes a data receiving unit, a key element extraction unit and a layout understanding unit; the data receiving unit is used to receive various forms of input information provided by the user; the key element extraction unit is used to analyze and process the received input information, and extract key elements, layout information and rendering information for scene generation; the layout understanding unit is used to establish the spatial relationship between the key elements to provide a blueprint for subsequent scene construction;

[0029] The scene construction module includes a three-dimensional object generation unit and a scene layout adjustment unit; the three-dimensional object generation unit is used to generate three-dimensional object models corresponding to each key element provided by the scene understanding module; the scene layout adjustment unit is used to complete the spatial position adjustment and fusion between the three-dimensional object models;

[0030] The scene adjustment module includes a user feedback receiving unit, a scene fine-tuning unit and a rendering update unit; the user feedback receiving unit is used to receive user feedback information on the generated scene, the scene fine-tuning unit is used to perform specific scene fine-tuning operations in combination with the feedback information; the rendering update unit is used to complete the rendering operation of the fine-tuned scene;

[0031] Further, such as Figure 2 As shown, the key elements extracted by the key element extraction unit also include the description clarity of each key element; the description clarity of each key element is obtained in the following manner:

[0032] S11: Acquire attribute information of each key element, wherein the attribute information includes but is not limited to attribute descriptions such as size, color, shape, etc. of the key element;

[0033] S12: Calculate the attribute completeness score of each key element:

[0034]

[0035]

[0036] Among them, P i is the attribute completeness score of the i-th key element, a ik is the existence determination coefficient of attribute k in the ith key element, m is the number of attributes of the ith key element, which can be pre-set according to the type of the key element;

[0037] S13: Acquire semantic description information of each key element; the semantic description information is all lexical descriptions of the corresponding key element;

[0038] S14: Calculate the semantic clarity score of each key element:

[0039]

[0040] Among them, S i is the semantic clarity score of the i-th key element, M i is the number of unclear or polysemous words in the semantic description information of the i-th key element, W i is the total number of words in the semantic description information of the i-th key element;

[0041] S15: Evaluate the clarity of description of each key element:

[0042] D i =ω1·P i +ω2·S i ;

[0043] Among them, D i is the description clarity of the i-th key element, ω1 and ω2 are the pre-set attribute completeness and semantic clarity weight coefficients, respectively, satisfying ω1+ω2=1;

[0044] Further, the layout understanding unit includes an element layout subunit, an element connection word library and a conflict element adjustment subunit, wherein the element layout subunit is used to generate a preliminary layout of each key element according to each key element and its layout information; the element connection word library is a preset database containing the quantitatively expressed connection degree between different key elements; the conflict element adjustment subunit is used to identify and resolve the layout conflicts occurring in the preliminary layout of each key element, and generate each adjusted key element and its final layout;

[0045] Further, such as Figure 3 As shown, the specific workflow of the conflict element adjustment subunit is as follows:

[0046] S21: identifying layout conflicts of key elements in the scene by using spatial analysis technology; the conflicts include but are not limited to position overlaps between key elements and layouts that do not conform to physical laws; the spatial analysis technology includes but is not limited to collision detection and physical rule verification;

[0047] S22: For key elements that have layout conflicts with each other, evaluate their priorities:

[0048]

[0049] Among them, F i is the priority coefficient of the i-th key element with layout conflict, E i,j is the degree of connection between the i-th key element with layout conflict and the j-th key element without layout conflict in the scene, satisfying 0≤E i,j ≤1; n is the number of key elements in the scene without layout conflicts;

[0050] S23: comparing the priorities of key elements that have layout conflicts with each other, and selecting the key element with the highest priority, keeping its current attributes and position unchanged, as the basic point for layout adjustment;

[0051] S24: for other key elements with lower priorities, performing attribute reduction or position adjustment operations on these key elements;

[0052] S25: Repeat the previous step until there is no layout conflict in the scene, and use the final adjusted key elements and their final layout as a blueprint for subsequent scene construction;

[0053] Furthermore, the three-dimensional object generation unit dynamically generates three-dimensional object models corresponding to the key elements provided by the conflict element adjustment subunit using a generative adversarial network; in the process of generating the three-dimensional object models corresponding to the key elements, the detail level of the three-dimensional object model is adjusted according to the description clarity of each key element, so as to ensure that the generated three-dimensional object model meets the description input by the user;

[0054] Furthermore, the scene layout adjustment unit adjusts and merges the spatial positions of the generated three-dimensional object models according to the final layout provided by the conflict element adjustment subunit; this includes adjusting the relative positions and angles between the three-dimensional object models in combination with geometric constraints and simulated physical rules, thereby completing the construction of the three-dimensional scene;

[0055] Furthermore, the scene adjustment module includes a scene feedback update unit and a rendering feedback update unit; the scene feedback update unit is used to receive user feedback information on the generated scene and perform modification operations on the generated scene; the rendering update unit is used to receive user rendering feedback information and perform rendering operations on the modified scene to generate a modified three-dimensional scene.

[0056] Embodiment 2:

[0057] This embodiment should be understood to include at least all the features of any of the above embodiments, and further improve upon them;

[0058] Although the automatic generation technology can greatly improve the efficiency and creativity of scene design, it is still a challenge to ensure that the generated scene fully meets the specific needs and detailed expectations of the user. To make up for the gaps that may occur during the automatic generation process, this embodiment provides a flexible post-adjustment mechanism for the user in combination with a scene adjustment module. The scene adjustment module includes a scene feedback update unit and a rendering feedback update unit. The scene feedback update unit is used to receive user feedback information on the generated scene and perform modification operations on the generated scene. The rendering update unit is used to receive user rendering feedback information and perform rendering operations on the modified scene to generate a modified final three-dimensional scene.

[0059] Furthermore, the scene feedback update unit includes a key element adjustment subunit, a key element replacement unit and a layout update unit; the key element adjustment subunit is used to receive user feedback on the adjustment information of the key elements already in the generated scene, and the key element replacement unit is used to receive user feedback on the addition of new key elements and the deletion of key elements already in the generated scene; the layout update unit is used to receive the layout adjustment information fed back by the user, generate the key elements adjusted by the user feedback and their final layout as the user modified blueprint, and send the modified blueprint to the scene construction module to construct the modified three-dimensional scene;

[0060] Further, such as Figure 4 As shown, the specific workflow of the layout updating unit is as follows:

[0061] S31: Obtaining layout adjustment information fed back by the user, generating adjustment information of key elements existing in the scene, adding new key elements and generating deletion information of key elements existing in the scene, and generating a preliminary layout of each key element after adjustment based on all the above information;

[0062] S32: Combining spatial analysis technology to identify layout conflicts of key elements after adjustment in the preliminary layout scene;

[0063] S33: For key elements that have layout conflicts with each other in the adjusted preliminary layout scene, evaluate their modification priorities:

[0064]

[0065] Among them, G k is the priority coefficient of the key element with layout conflict after the kth adjustment, D k is the description clarity of the key element with layout conflict after the kth adjustment, α k is the adjustment weight of the key element with layout conflict after the kth adjustment, E k,m is the degree of connection between the kth key element with layout conflict after adjustment and the mth key element without layout conflict in the scene after adjustment, satisfying 0≤E k,m ≤1; l is the number of key elements without layout conflicts in the adjusted scene, max(E k,m ) for all E k,m The maximum value in ;

[0066] For adjusting the weight α k satisfy:

[0067]

[0068] S34: comparing the priorities of the key elements that have layout conflicts after the adjustment, and selecting the key element with the highest priority, keeping its current attributes and position unchanged, as the basic point for layout adjustment;

[0069] S35: For other key elements with lower priorities, perform attribute reduction or position adjustment operations on these key elements;

[0070] S36: Repeat the previous step until there is no layout conflict in the modified scene, and use the key elements and their final layouts that have been finally adjusted as the blueprint modified by the user;

[0071] S37: Sending the modified blueprint to the scene construction module to construct the modified three-dimensional scene;

[0072] This embodiment improves the functionality and flexibility of the scene generation system by executing user feedback modifications on the generated scene through the scene adjustment module; allows the user to carefully customize the automatically generated scene according to personal preferences, and the personalized customization capability greatly enhances the user experience; the user can directly provide feedback and adjust the generated scene, greatly improving the design and development efficiency of the overall scene.

[0073] The contents disclosed above are only preferred feasible embodiments of the present invention, and do not limit the protection scope of the present invention. Therefore, all equivalent technical changes made using the contents of the present invention specification and drawings are included in the protection scope of the present invention. In addition, the elements therein can be updated as technology develops.

Claims

1. A scene generation system based on artificial intelligence, characterized in that: The system includes a scene understanding module, a scene construction module, a rendering optimization module and a scene adjustment module; The scene understanding module is used to parse the key elements, layout information and rendering information for scene generation from the user input information; the scene construction module is used to construct a three-dimensional scene according to the information provided by the scene understanding unit; The rendering optimization module is used to render the three-dimensional scene constructed by the scene construction module to generate a final three-dimensional scene; The scene adjustment module is used to modify and adjust the finally generated three-dimensional scene according to user needs; The scene understanding module includes a data receiving unit, a key element extraction unit and a layout understanding unit; the data receiving unit is used to receive various forms of input information provided by the user; the key element extraction unit is used to analyze and process the received input information, and extract key elements, layout information and rendering information for scene generation; the layout understanding unit is used to establish the spatial relationship between the key elements to provide a blueprint for subsequent scene construction; The scene construction module includes a three-dimensional object generation unit and a scene layout adjustment unit; the three-dimensional object generation unit is used to generate three-dimensional object models corresponding to each key element provided by the scene understanding module; the scene layout adjustment unit is used to complete the spatial position adjustment and fusion between the three-dimensional object models.

2. The scene generation system based on artificial intelligence according to claim 1, characterized in that: The key elements extracted by the key element extraction unit also include description clarity of each key element. The description clarity of each key element is obtained by obtaining the attribute integrity and semantic clarity of each key element.

3. The scene generation system based on artificial intelligence according to claim 1, characterized in that: The layout understanding unit includes an element layout subunit, an element connection word library and a conflict element adjustment subunit. The element layout subunit is used to generate a preliminary layout of each key element according to each key element and its layout information; the element connection word library is a preset database containing a quantitatively expressed degree of connection between different key elements; the conflict element adjustment subunit is used to identify and resolve layout conflicts arising from the preliminary layout of each key element by combining the description clarity and connection degree of each key element, and generate adjusted key elements and their final layout.

4. The scene generation system based on artificial intelligence according to claim 1, characterized in that: The three-dimensional object generation unit uses a generative adversarial network to dynamically generate three-dimensional object models corresponding to each key element provided by the conflict element adjustment subunit; in the process of generating the three-dimensional object model corresponding to each key element, it includes adjusting the degree of detail of the three-dimensional object model according to the description clarity of each key element, thereby ensuring that the generated three-dimensional object model conforms to the user input description.

5. The scene generation system based on artificial intelligence according to claim 1, characterized in that: The scene layout adjustment unit adjusts and merges the spatial positions of the generated three-dimensional object models according to the final layout provided by the conflict element adjustment sub-unit; this includes adjusting the relative positions and angles between the three-dimensional object models in combination with geometric constraints and simulated physical rules, thereby completing the construction of the three-dimensional scene.

6. The scene generation system based on artificial intelligence according to claim 1, characterized in that: The rendering optimization module renders the three-dimensional scene constructed by the scene layout adjustment unit through ray tracing and physically based rendering technology based on the rendering information provided by the scene understanding module, thereby generating a final three-dimensional scene.

7. The scene generation system based on artificial intelligence according to claim 1, characterized in that: The scene adjustment module includes a scene feedback update unit and a rendering feedback update unit; The scene feedback update unit is used to receive user feedback information on the generated scene and perform a modification operation on the generated scene; the rendering update unit is used to receive user rendering feedback information and perform a rendering operation on the modified scene to generate a modified three-dimensional scene.

Images