Resource Integration Method and System in the Construction of Metaverse Scenarios Based on an Efficient Engine

Abstract

The present invention provides a resource integration method and system in the construction of a metaverse scene based on an efficient engine, which relates to the technical field of the metaverse. It includes classifying and hierarchically storing scene resource data, and predicting the user's target scene area based on a deep learning model, so as to preload relevant resources. Calculate the resource dependency relationship using a resource organizational structure diagram, generate a resource loading sequence and construct a resource index tree, and then adaptively compress resources according to the resource association degree to reduce the resource loading time and bandwidth consumption. In addition, the present invention also provides an alternative resource loading channel to ensure the stability of scene rendering. The present invention can effectively improve the loading speed and rendering efficiency of the metaverse scene and enhance the user experience.

Description

Technical Field

[0001] The present invention relates to the metaverse technology, and in particular, to a resource integration method and system in the construction of metaverse scenes based on an efficient engine. Background Art

[0002] In the construction of metaverse scenes, resource integration is a crucial link, which directly affects the scene loading speed, rendering efficiency, and user experience. To build an immersive and smooth metaverse world, it is necessary to process a large amount of scene resource data, including models, textures, sound effects, animations, etc. These resources are not only huge in quantity but also have complex dependencies on each other. How to effectively organize, manage, and load these resources is a major challenge faced by the current development of metaverse technology.

[0003] However, the existing technologies have the following deficiencies in resource integration:

[0004] First, the resource organization method is relatively rough, lacking refined management of the association relationships between resources. Traditional methods usually use simple folder or database structures to store resources, making it difficult to clearly express the dependencies between resources, resulting in omissions or errors during resource loading and reducing the efficiency of scene rendering.

[0005] Second, the resource loading strategy is not intelligent enough to adaptively adjust according to user behavior. Existing resource loading methods often adopt preloading or on-demand loading strategies, unable to predict the user's next behavior based on the user's historical interaction data, resulting in the user may need to wait for a long time to see the target scene, affecting the user experience. Summary of the Invention

[0006] Embodiments of the present invention provide a method and a system that can solve the problems in the existing technologies.

[0007] In the first aspect of the embodiments of the present invention, a resource integration method in the construction of metaverse scenes based on an efficient engine is provided, including:

[0008] Obtain the scene resource data in the metaverse scene, classify the scene resource data according to the data type, establish an organizational structure diagram of the scene resource data, and record the association relationships between the scene resource data in the organizational structure diagram; classify the scene resource data according to the association degree and usage frequency of the scene resource data, and store the scene resource data in the storage area corresponding to the corresponding level;

[0009] Analyze the historical interaction data of the user in the metaverse scene based on a deep learning model, output the target scene area that the user will access next, and obtain the storage location information of the scene resource data corresponding to the target scene area;

[0010] Calculate the resource dependency relationship of the target scene area based on the organizational structure diagram, and generate a resource loading sequence; establish a resource index tree according to the resource loading sequence, record the hierarchical relationship and association degree between scene resources in the resource index tree, and perform adaptive compression on the scene resources based on the resource index tree, where the compression ratio is inversely proportional to the resource association degree, and the higher the association degree of the resource, the lower the compression ratio is used;

[0011] Monitor the loading status of scene resources in real time. When an abnormal loading of scene resources is detected, start the resource loading backup channel, and obtain the backup data of the abnormal resources through the backup channel; output the integrated scene resource data to the metaverse scene rendering engine for real-time rendering of the metaverse scene.

[0012] Classify the scene resource data according to the data type, and establish an organizational structure diagram of the scene resource data. The association relationships recorded in the organizational structure diagram between scene resource data include:

[0013] Divide the scene resource data into geometric data, material data, environmental data, and interaction data according to the data presentation form. The geometric data includes model mesh data and vertex coordinate data, the material data includes material texture data and normal map data, the environmental data includes light information data and shadow information data, and the interaction data includes collision body data and trigger data;

[0014] Divide the scene resource data into basic element layer data, combined element layer data, and scene element layer data according to the data granularity. The basic element layer data includes single model data and single texture data, the combined element layer data includes model combination data and material combination data, and the scene element layer data includes complete scene unit data;

[0015] Calculate the association strength between scene resource data; construct an organizational structure diagram of scene resource data based on the association strength. In the organizational structure diagram of scene resource data, use the scene resource data as nodes and the association relationships between scene resource data as connection edges, and the weight of the connection edges is determined by the association strength.

[0016] Analyze the historical interaction data of the user in the metaverse scene based on the deep learning model, and output the target scene area that the user will access next and obtain the storage location information of the scene resource data corresponding to the target scene area, including:

[0017] Obtain the historical interaction data of the user in the metaverse scenario, where the historical interaction data includes spatio-temporal dimension feature data and interaction behavior feature data. Among them, the spatio-temporal dimension feature data includes user movement trajectory sequence data and scene residence time distribution data, and the interaction behavior feature data includes operation sequence coding data and interaction intensity matrix data;

[0018] Use a bidirectional long short-term memory network to process the historical interaction data, obtain temporal features through the combination of forward hidden states and backward hidden states, and perform weighted fusion on the temporal features through learnable parameters to obtain fused temporal feature data;

[0019] Construct a user interest model based on the multi-head self-attention mechanism, map the fused temporal feature data into a query matrix, a key matrix, and a value matrix, calculate the attention weight distribution, and obtain user interest feature data through multi-head splicing;

[0020] Perform feature fusion on the fused temporal feature data and the user interest feature data to obtain scene prediction feature data;

[0021] Use a multi-layer perceptron to process the scene prediction feature data and output the predicted probability distribution of the target scene.

[0022] Calculate the resource dependency relationship of the target scene area based on the organizational structure diagram, and generate a resource loading sequence; establish a resource index tree according to the resource loading sequence, record the hierarchical relationship and association degree between scene resources in the resource index tree, and perform adaptive compression on scene resources based on the resource index tree, including:

[0023] Calculate the dependency strength between scene resources based on the scene resource data. The dependency strength is obtained through weighted calculation of functional dependency, temporal dependency, and content dependency. Among them, the functional dependency is calculated based on the resource call data, the temporal dependency is calculated based on the resource loading data, and the content dependency is calculated based on the resource content data;

[0024] Calculate the resource dependency relationship of the target scene area based on the organizational structure diagram. The resource dependency relationship is obtained by multiplying the dependency strengths on the dependency transfer path and applying a distance attenuation function, and the distance attenuation function decreases exponentially with the length of the dependency transfer path;

[0025] Combine the dependency strength and the resource dependency relationship to obtain the loading priority of resource nodes, sort the resource nodes according to the loading priority, and generate a resource loading sequence;

[0026] Construct a resource index tree based on the resource loading sequence, record the hierarchical relationship of resource nodes in the resource index tree, and calculate the association metric value between resource nodes; determine a resource compression strategy according to the association metric value;

[0027] Map the association metric value to a compression ratio through the Sigmoid function; perform adaptive compression processing on the scene resources according to the compression ratio to obtain the compressed scene resource data;

[0028] Load the compressed scene resource data according to the resource loading sequence to achieve optimized loading of scene resources.

[0029] Construct a resource index tree based on the resource loading sequence, record the hierarchical relationship of resource nodes in the resource index tree, and calculate the association metric value between resource nodes; determining a resource compression strategy according to the association metric value includes:

[0030] Construct a resource index tree based on the resource loading sequence, determine the hierarchical value of the resource node by judging the parent-child relationship of the resource node, and the hierarchical value of the resource node is obtained through recursive calculation. For a resource node without a parent node, set the hierarchical value to zero, and for a resource node with a parent node, set the hierarchical value to the hierarchical value of its parent node plus one;

[0031] Calculate the resource node weight according to the hierarchical value of the resource node and the resource size. The resource node weight decays exponentially with the hierarchical value, and the resource node weight is positively correlated with the resource size;

[0032] Calculate the direct association strength between resource nodes. The direct association strength is obtained by multiplying the intersection-over-union ratio of the resource node connection set by the hierarchical difference penalty term, where the hierarchical difference penalty term maps the hierarchical difference to the interval from zero to two through the Sigmoid function;

[0033] Calculate the path association strength between resource nodes based on the direct association strength. The path association strength is obtained by multiplying the product of the direct association strengths on the path by the path attenuation function;

[0034] Weight the direct association strength and the path association strength to obtain the comprehensive association metric value between resource nodes;

[0035] Calculate the balance factor of the resource nodes in the resource index tree. The balance factor is the difference between the height of the left subtree and the height of the right subtree of the resource node. When the absolute value of the balance factor is greater than the preset balance threshold, perform rebalancing processing on the resource index tree;

[0036] Calculate the basic compression rate of the resource node based on the comprehensive correlation metric value, where the basic compression rate maps the difference between the average correlation degree of the resource node and the global average correlation degree to the interval from zero to one through the Sigmoid function;

[0037] Adaptively adjust the basic compression rate according to the resource node weight to obtain the adaptive compression rate, and the adaptive compression rate decreases as the resource node weight increases.

[0038] When it is detected that the scene resource loading is abnormal, start the resource loading backup channel and obtain the backup data of the abnormal resource through the backup channel; Outputting the integrated scene resource data to the metaverse scene rendering engine includes:

[0039] Calculate the priorities of multiple backup channels, where the priorities are obtained by inputting the channel bandwidth, channel reliability, and channel delay into the Sigmoid function, and select the backup channel with the highest priority to load the abnormal resource;

[0040] Obtain the backup data of the abnormal resource from the backup channel, and calculate the consistency check value between the abnormal resource and the backup data, where the consistency check value is calculated based on the hash value difference between the abnormal resource and the backup data;

[0041] When the consistency check value is less than the preset consistency threshold, use the backup data as the replacement data for the abnormal resource. When the consistency check value is not less than the preset consistency threshold, perform data merging processing on the abnormal resource and the backup data to obtain the replacement data;

[0042] Calculate the integrity evaluation value of the scene resource data, where the integrity evaluation value is the ratio of the effective data volume of the scene resource data to the expected data volume; Calculate the rendering adaptability of the scene resource data, where the rendering adaptability is calculated based on the difference between the resource format parameters and the target format parameters;

[0043] Under the constraint conditions that the integrity evaluation value is not lower than the minimum integrity threshold and the rendering adaptability is not lower than the minimum adaptability threshold, output the scene resource data to the metaverse scene rendering engine.

[0044] In the second aspect of the embodiments of the present invention, a resource integration system in the construction of a metaverse scene based on an efficient engine is provided, including:

[0045] The first unit is used to obtain the scene resource data in the metaverse scene, classify the scene resource data according to the data type, establish an organizational structure diagram of the scene resource data, and record the association relationship between the scene resource data in the organizational structure diagram; Classify the scene resource data according to the degree of association and usage frequency of the scene resource data, and store the scene resource data in the storage area corresponding to the level;

[0046] A second unit, configured to analyze historical interaction data of a user in a metaverse scenario based on a deep learning model, output a target scenario area that the user will access next, and obtain location information of a storage location of scenario resource data corresponding to the target scenario area;

[0047] A third unit, configured to calculate a resource dependency relationship of a target scenario area based on the organizational structure diagram, generate a resource loading sequence; establish a resource index tree according to the resource loading sequence, record a hierarchical relationship and an association degree between scenario resources in the resource index tree, and perform adaptive compression on scenario resources based on the resource index tree, wherein a compression ratio is inversely proportional to the resource association degree, and resources with a higher association degree are compressed at a lower compression ratio;

[0048] A fourth unit, configured to monitor a loading state of scenario resources in real time, and when detecting an abnormal loading of scenario resources, start a backup channel for resource loading, and obtain backup data of abnormal resources through the backup channel; output the integrated scenario resource data to a metaverse scenario rendering engine for real-time rendering of the metaverse scenario.

[0049] In a third aspect of the embodiments of the present invention,

[0050] There is provided an electronic device, including:

[0051] A processor;

[0052] A memory for storing instructions executable by the processor;

[0053] Wherein, the processor is configured to call the instructions stored in the memory to execute the method described above.

[0054] In a fourth aspect of the embodiments of the present invention,

[0055] There is provided a computer-readable storage medium, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the method described above is implemented.

[0056] The beneficial effects of the present application are as follows:

[0057] 1. Improve resource loading efficiency: By analyzing historical user interaction data to predict user behavior, preloading resources of the target scenario area, and generating a loading sequence according to the resource dependency relationship, unnecessary resource loading is avoided, thereby shortening the scene loading time and improving the user experience.

[0058] 2. Optimize resource storage and utilization: Hierarchical storage is performed according to the resource association degree and usage frequency, and adaptive compression is performed based on the resource index tree, reducing the storage space occupation and improving the resource utilization rate.

[0059] 3. Enhance system stability: Real-time monitoring of resource loading status and activation of backup channels in case of abnormal loading ensure the stability and smoothness of scene rendering and improve the fault tolerance of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0060] Figure 1 FIG. is a flowchart of a resource integration method in the construction of a metaverse scene based on an efficient engine according to an embodiment of the present invention;

[0061] Figure 2 FIG. is a flowchart of an adaptive compression process for metaverse scene resources;

[0062] Figure 3 FIG. is a schematic diagram of the time distribution of resource loading. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0063] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Apparently, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the protection scope of the present invention.

[0064] The technical solutions of the present invention will be described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments.

[0065] Figure 1 FIG. is a flowchart of a resource integration method in the construction of a metaverse scene based on an efficient engine according to an embodiment of the present invention, as Figure 1 shown, the method includes:

[0066] Obtain the scene resource data in the metaverse scene, classify the scene resource data according to the data type, establish an organizational structure diagram of the scene resource data, and record the association relationship between the scene resource data in the organizational structure diagram; classify the scene resource data according to the association degree and usage frequency of the scene resource data, and store the scene resource data in the storage area corresponding to the corresponding level;

[0067] Analyze the historical interaction data of the user in the metaverse scene based on a deep learning model, output the target scene area that the user will access next, and obtain the storage location information of the scene resource data corresponding to the target scene area;

[0068] Calculate the resource dependency relationship of the target scenario area based on the organizational structure diagram to generate a resource loading sequence; establish a resource index tree according to the resource loading sequence, record the hierarchical relationship and association degree between scenario resources in the resource index tree, and perform adaptive compression on the scenario resources based on the resource index tree, where the compression ratio is inversely proportional to the resource association degree, and the higher the association degree of the resource, the lower the compression ratio;

[0069] Monitor the loading status of scenario resources in real time. When an abnormal loading of scenario resources is detected, start the backup channel for resource loading, and obtain the backup data of the abnormal resources through the backup channel; output the integrated scenario resource data to the metaverse scene rendering engine for real-time rendering of the metaverse scene.

[0070] In an alternative implementation, classify the scenario resource data according to the data type, and establish an organizational structure diagram of the scenario resource data. The association relationships recorded in the organizational structure diagram include:

[0071] Divide the scenario resource data into geometric data, material data, environmental data, and interaction data according to the data presentation form. The geometric data includes model mesh data and vertex coordinate data. The material data includes material texture data and normal map data. The environmental data includes light information data and shadow information data. The interaction data includes collision body data and trigger data;

[0072] Divide the scenario resource data into basic element layer data, combined element layer data, and scenario element layer data according to the data granularity. The basic element layer data includes single model data and single texture data. The combined element layer data includes model combination data and material combination data. The scenario element layer data includes complete scenario unit data;

[0073] Calculate the association strength between scenario resource data; construct an organizational structure diagram of the scenario resource data based on the association strength. In the organizational structure diagram of the scenario resource data, regard the scenario resource data as nodes, and regard the association relationships between the scenario resource data as connection edges. The weight of the connection edges is determined by the association strength.

[0074] Divide the scenario resource data into four categories: geometric data, material data, environmental data, and interaction data. The geometric data includes model mesh data and vertex coordinate data; the material data includes material texture data and normal map data; the environmental data includes light information data and shadow information data; the interaction data includes collision body data and trigger data.

[0075] For example, in a virtual city scene, the geometric data may include the 3D model mesh and vertex coordinates of buildings; the material data may include the texture map and normal map of the building surface; the environmental data may include the daylight illumination information in the city and the shadow information cast by buildings; the interaction data may include the collision volume of buildings and the trigger area for entering buildings.

[0076] The scene resource data is divided into three levels: the basic element layer data, the combined element layer data, and the scene element layer data. The basic element layer data includes single model data and single texture data; the combined element layer data includes model combination data and material combination data; the scene element layer data includes complete scene unit data.

[0077] Taking the virtual city scene as an example, the basic element layer data can be the model of a single building or a single texture of the building surface; the combined element layer data can be the block model combination composed of multiple buildings or the building appearance formed by multiple material combinations; the scene element layer data can be a complete urban area, containing multiple elements such as blocks, roads, and green spaces.

[0078] The association strength can be calculated in the following way:

[0079] For two scene resource data A and B, first determine the association type between them. The association type can be an inclusion relationship, a reference relationship, or an adjacency relationship. The inclusion relationship means that one data contains another data, such as a scene element containing a combined element; the reference relationship means that one data references another data, such as a model referencing a material; the adjacency relationship means that two data are adjacent in the scene, such as two building models being adjacent in space.

[0080] Then, assign a basic weight value according to the association type. The basic weight for the inclusion relationship can be set to 0.8, the basic weight for the reference relationship can be set to 0.6, and the basic weight for the adjacency relationship can be set to 0.4.

[0081] If both data A and B are frequently used data (such as the main building model or commonly used material), the frequency factor can be set to 1.2; if they are medium-frequency used data, the frequency factor can be set to 1.0; if they are low-frequency used data, the frequency factor can be set to 0.8.

[0082] If the spatial distance between data A and B in the scene is very close, the distance factor can be set to 1.2; if the distance is moderate, the distance factor can be set to 1.0; if the distance is far, the distance factor can be set to 0.8.

[0083] Multiply the base weight by the frequency factor and the distance factor to obtain the final association strength value. For example, if there is a reference relationship (base weight 0.6) between building model A and material B, and both are data used frequently (frequency factor 1.2) with a moderate spatial distance (distance factor 1.0), then the association strength between them is 0.6×1.2×1.0 = 0.72.

[0084] Based on the calculated association strength, construct an organizational structure diagram of the scene resource data. In this organizational structure diagram, the scene resource data is used as nodes, and the association relationships between the scene resource data are used as connecting edges, and the weights of the connecting edges are determined by the association strength.

[0085] When specifically implemented, a graph data structure can be used to represent the organizational structure diagram. Each node contains the following attributes: node ID, data type (geometric data, material data, environmental data, or interaction data), data granularity (basic element layer, combined element layer, or scene element layer), data content description, storage location, etc. Each edge contains the following attributes: starting node ID, target node ID, association type, association strength value, etc.

[0086] Taking the virtual city scene as an example, assume there is a building model node A (geometric data, basic element layer) and a material texture node B (material data, basic element layer), and there is a reference relationship between them with an association strength of 0.72. In the organizational structure diagram, there will be a directed edge with a weight of 0.72 between node A and node B, indicating that A references B.

[0087] To optimize the organizational structure diagram, an association strength threshold can be set, and only the connecting edges with an association strength greater than the threshold are retained. For example, if the threshold is set to 0.5, then only the association relationships with an association strength greater than 0.5 will be displayed in the organizational structure diagram. This can reduce the complexity of the graph and highlight the important association relationships.

[0088] In addition, according to different application scenarios, different views of the organizational structure diagram can be displayed. For example, only the association relationships between geometric data can be displayed, or only the association relationships from the basic element layer to the combined element layer can be displayed to meet different analysis requirements.

[0089] The organizational structure diagram of the scene resource data constructed by the above method can clearly display the association relationships between the scene resource data, which is helpful for the management, optimization, and reuse of the scene resources, and improves the efficiency and quality of scene construction.

[0090] In an alternative implementation, grading the scene resource data according to the association degree and usage frequency of the scene resource data and storing the scene resource data in the corresponding storage area levels includes:

[0091] Calculate the priority index of the scenario resource data, where the priority index is obtained by weighting the usage frequency factor, the calculation complexity factor, and the correlation factor. The usage frequency factor is determined based on the access times, the calculation complexity factor is determined based on the resource processing overhead, and the correlation factor is determined based on the average correlation strength;

[0092] Divide the scenario resource data into different levels of storage areas according to the priority index. Store the scenario resource data with a priority index greater than the first classification threshold in the high-priority storage area, store the scenario resource data with a priority index greater than the second classification threshold and less than the first classification threshold in the medium-priority storage area, and store the scenario resource data with a priority index less than the second classification threshold in the low-priority storage area.

[0093] The system obtains the relevant information of the scenario resource data, including the usage frequency, the calculation complexity, and the correlation with other resources. Based on this information, the system calculates the priority index of each scenario resource data and allocates the data to different levels of storage areas according to the index.

[0094] The priority index is a comprehensive score calculated by weighting the usage frequency factor, the calculation complexity factor, and the correlation factor. The specific calculation method is as follows:

[0095] The calculation of the priority index adopts the weighted average method, that is, multiply the usage frequency factor, the calculation complexity factor, and the correlation factor by the corresponding weight coefficients respectively and then add them up. In practical applications, the weights of each factor can be adjusted according to the specific scenario. For example, in a typical configuration, the weight of the usage frequency factor is 0.5, the weight of the calculation complexity factor is 0.3, and the weight of the correlation factor is 0.2.

[0096] The usage frequency factor is determined based on the access times of the scenario resource data. The system records the number of times each resource is accessed within a period of time (such as 30 days) and normalizes it to a value between 0 and 1.

[0097] When specifically implemented, the system first counts the access times of all resources and finds the maximum access times max_Count and the minimum access times min_Count . Then, for the resource with the access times of count , its usage frequency factor is calculated as: ( (count - min_Count) / (max_Count - min_Count) .

[0098] For example, assume that within the past 30 days, Resource A was accessed 500 times, Resource B was accessed 100 times, and Resource C was accessed 50 times. The maximum number of accesses is 500, and the minimum number of accesses is 50. Then, the usage frequency factor of Resource A is (500 - 50) / (500 - 50) = 1, the usage frequency factor of Resource B is (100 - 50) / (500 - 50) = 0.11, and the usage frequency factor of Resource C is (50 - 50) / (500 - 50) = 0.

[0099] The computation complexity factor is determined based on the computational overhead required to process the resources. The system records metrics such as the average CPU time and memory usage required to process each resource and synthesizes them into a complexity score.

[0100] In a specific implementation, the system can record the CPU occupancy time (in milliseconds) and the peak memory usage (in MB) when processing the resources. Assume the CPU time weight is 0.6 and the memory usage weight is 0.4. Then, the complexity score is: 0.6 × CPU time + 0.4 × memory usage. Then, the system finds the maximum complexity score among all resources max_Complexity and the minimum complexity score min_Complexity , for a resource with a complexity score of complexity , its computation complexity factor is: ( complexity - min_Complexity ) / ( max_Complexity - min_Complexity ).

[0101] For example, when processing Resource A, the average CPU time is 200 milliseconds and the memory usage is 150 MB. Its complexity score is 0.6 × 200 + 0.4 × 150 = 180. The complexity score of Resource B is 120, and the complexity score of Resource C is 90. The maximum complexity score is 180, and the minimum complexity score is 90. Then, the computation complexity factor of Resource A is (180 - 90) / (180 - 90) = 1, the computation complexity factor of Resource B is (120 - 90) / (180 - 90) = 0.33, and the computation complexity factor of Resource C is (90 - 90) / (180 - 90) = 0.

[0102] The correlation factor is determined based on the average correlation strength between the scenario resource data and other resources. The system analyzes the reference relationships between resources and calculates the average correlation strength of each resource.

[0103] In specific implementation, the system constructs an association graph among resources, recording the number of times each resource is referenced by other resources and the number of times it references other resources. For each pair of associated resources, the system calculates the association strength (ranging from 0 to 1) based on metrics such as reference frequency and co-access frequency. Then, for each resource, the average of its association strengths with all relevant resources is calculated as the average association strength of that resource.

[0104] For example, if the association strength between resource A and resource D is 0.8, with resource E is 0.6, and with resource F is 0.4, then the average association strength of resource A is (0.8 + 0.6 + 0.4) / 3 = 0.6. Assuming the maximum average association strength among all resources is 0.8 and the minimum is 0.2, then the association degree factor of resource A is (0.6 - 0.2) / (0.8 - 0.2) = 0.67.

[0105] Based on the calculated priority index, the system divides the scenario resource data into different levels of storage areas:

[0106] 1. High-priority storage area: Stores scenario resource data with a priority index greater than the first classification threshold. These data usually have a high access frequency, high computational complexity, or high association with other resources and require fast access. The high-priority storage area usually uses high-speed solid-state drives or memory caches to provide the best read and write performance.

[0107] 2. Medium-priority storage area: Stores scenario resource data with a priority index greater than the second classification threshold and less than the first classification threshold. The access requirements for these data are moderate and can be stored on ordinary solid-state drives or high-performance mechanical hard drives.

[0108] 3. Low-priority storage area: Stores scenario resource data with a priority index less than the second classification threshold. These data have a low access frequency, low computational complexity, or low association and can be stored on large-capacity mechanical hard drives or cloud storage.

[0109] In practical applications, the first classification threshold and the second classification threshold can be adjusted according to system resources and performance requirements. For example, the first classification threshold can be set to 0.7 and the second classification threshold to 0.3.

[0110] The following is a specific implementation case demonstrating the process of hierarchical storage of scenario resource data:

[0111] Suppose there are 10 scenario resource data (R1 to R10), and the system collects their usage frequency, computational complexity, and association information and calculates the corresponding factor values:

[0112] R1: Usage frequency factor = 0.9, computational complexity factor = 0.8, association degree factor = 0.7;

[0113] R2: Usage frequency factor = 0.8, computational complexity factor = 0.7, correlation factor = 0.6;

[0114] R3: Usage frequency factor = 0.7, computational complexity factor = 0.6, correlation factor = 0.5;

[0115] R4: Usage frequency factor = 0.6, computational complexity factor = 0.5, correlation factor = 0.4;

[0116] R5: Usage frequency factor = 0.5, computational complexity factor = 0.4, correlation factor = 0.3;

[0117] R6: Usage frequency factor = 0.4, computational complexity factor = 0.3, correlation factor = 0.2;

[0118] R7: Usage frequency factor = 0.3, computational complexity factor = 0.2, correlation factor = 0.1;

[0119] R8: Usage frequency factor = 0.2, computational complexity factor = 0.1, correlation factor = 0.0;

[0120] R9: Usage frequency factor = 0.1, computational complexity factor = 0.0, correlation factor = 0.0;

[0121] R10: Usage frequency factor = 0.0, computational complexity factor = 0.0, correlation factor = 0.0;

[0122] Use the weight configuration: Usage frequency factor weight = 0.5, computational complexity factor weight = 0.3, correlation factor weight = 0.2, and calculate the priority index of each resource:

[0123] Priority index of R1: 0.5×0.9 + 0.3×0.8 + 0.2×0.7 = 0.83;

[0124] Priority index of R2: 0.5×0.8 + 0.3×0.7 + 0.2×0.6 = 0.73;

[0125] Priority index of R3: 0.5×0.7 + 0.3×0.6 + 0.2×0.5 = 0.63

[0126] Priority index of R4: 0.5×0.6 + 0.3×0.5 + 0.2×0.4 = 0.53;

[0127] Priority index of R5: 0.5×0.5 + 0.3×0.4 + 0.2×0.3 = 0.43;

[0128] R6 priority index: 0.5×0.4 + 0.3×0.3 + 0.2×0.2 = 0.33;

[0129] R7 priority index: 0.5×0.3 + 0.3×0.2 + 0.2×0.1 = 0.23;

[0130] R8 priority index: 0.5×0.2 + 0.3×0.1 + 0.2×0.0 = 0.13;

[0131] R9 priority index: 0.5×0.1 + 0.3×0.0 + 0.2×0.0 = 0.05;

[0132] R10 priority index: 0.5×0.0 + 0.3×0.0 + 0.2×0.0 = 0.00;

[0133] Set the first classification threshold to 0.7 and the second classification threshold to 0.3, then the resource allocation is as follows:

[0134] High-priority storage area: R1 (0.83), R2 (0.73);

[0135] Medium-priority storage area: R3 (0.63), R4 (0.53), R5 (0.43), R6 (0.33);

[0136] Low-priority storage area: R7 (0.23), R8 (0.13), R9 (0.05), R10 (0.00);

[0137] Through this hierarchical storage method, the system can store resources in appropriate storage media according to the importance and access requirements of the resources, thereby optimizing the system performance and resource utilization.

[0138] In an alternative embodiment, analyzing the historical interaction data of the user in the metaverse scenario based on a deep learning model, outputting the target scenario area that the user will access next, and obtaining the storage location information of the scenario resource data corresponding to the target scenario area includes:

[0139] Obtain the historical interaction data of the user in the metaverse scenario, where the historical interaction data includes spatio-temporal dimension feature data and interaction behavior feature data, and the spatio-temporal dimension feature data includes user movement trajectory sequence data and scene residence time distribution data, and the interaction behavior feature data includes operation sequence coding data and interaction intensity matrix data;

[0140] The historical interaction data is processed using a bidirectional long short-term memory network, temporal features are obtained through the combination of forward hidden states and backward hidden states, and the temporal features are weighted and fused using learnable parameters to obtain fused temporal feature data;

[0141] A user interest model is constructed based on the multi-head self-attention mechanism. The fused temporal feature data is mapped into a query matrix, a key matrix, and a value matrix, the attention weight distribution is calculated, and the user interest feature data is obtained through multi-head concatenation;

[0142] The fused temporal feature data and the user interest feature data are feature-fused to obtain scene prediction feature data;

[0143] A multi-layer perceptron is used to process the scene prediction feature data, and the predicted probability distribution of the target scene is output.

[0144] The historical interaction data mainly includes two categories: spatio-temporal dimension feature data and interaction behavior feature data. The spatio-temporal dimension feature data includes user movement trajectory sequence data and scene residence time distribution data; the interaction behavior feature data includes operation sequence coding data and interaction intensity matrix data.

[0145] The user movement trajectory sequence data records the position change information of the user in the metaverse scene and can be represented as a series of three-dimensional coordinate points (x, y, z) and corresponding timestamps t. For example, the movement trajectory of a user in the "Science and Technology Museum" scene can be recorded as: [(125.3, 67.8, 0.0, 1623401256), (126.5, 70.2, 0.0, 1623401262),...], indicating the spatial positions of the user at different time points.

[0146] The scene residence time distribution data records the residence duration of the user in each scene area and can be represented as a mapping relationship between area identifiers and residence times. For example: {"Science and Technology Exhibition Area A": 325, "Interactive Experience Area B": 178, "Rest Area C": 42}, with the unit of seconds.

[0147] The operation sequence coding data records the user's interaction operations, and different types of operations are encoded as numerical values. For example, a click operation is encoded as 1, a drag operation is encoded as 2, and a zoom operation is encoded as 3, forming an operation sequence such as [1, 1, 3, 2, 1,...].

[0148] The interaction intensity matrix data records the interaction frequency and intensity of the user with each interaction object in the scene and is constructed in matrix form. For example, the interaction intensity of the user with "Science and Technology Exhibit D" is 0.85, and the interaction intensity with the "Virtual Tour Guide" is 0.32, etc.

[0149] The historical interaction data is processed using a Bidirectional Long Short-Term Memory Network (BiLSTM) to extract temporal features. The BiLSTM network consists of two directions, a forward LSTM and a backward LSTM, which can consider historical and future context information simultaneously.

[0150] In a specific implementation, the user's historical interaction data is organized into a sequence in chronological order and input into the BiLSTM network. The forward LSTM processes the sequence starting from the beginning position, generating a forward hidden state sequence; the backward LSTM processes the sequence starting from the end position, generating a backward hidden state sequence. For each time step t in the sequence, the corresponding forward hidden state and backward hidden state are concatenated to obtain the BiLSTM output at that time step.

[0151] For example, assume that the hidden state of the forward LSTM at time step t is [0.32, 0.45, 0.67, 0.21], and the hidden state of the backward LSTM at the same time step is [0.56, 0.23, 0.78, 0.44]. Then the BiLSTM output at this time step is [0.32, 0.45, 0.67, 0.21, 0.56, 0.23, 0.78, 0.44].

[0152] To perform weighted fusion of features at different time steps, learnable attention parameters are introduced. The attention scores for each time step are calculated, normalized as weights, and multiplied and summed with the BiLSTM output at the corresponding time step to obtain the fused temporal feature data.

[0153] For example, assume that the sequence length is 5, and the attention weights for each time step are [0.15, 0.25, 0.30, 0.20, 0.10]. Then the fused temporal feature is the weighted sum of the BiLSTM outputs at each time step.

[0154] The fused temporal feature data is mapped to a query matrix (Q), a key matrix (K), and a value matrix (V) through a linear transformation.

[0155] In the multi-head self-attention mechanism, first, the fused temporal feature data is copied h times (h is the number of attention heads, for example, h = 8), and each copy is passed through a different linear transformation to obtain different query, key, and value matrices. For each attention head, the dot product of the query matrix and the key matrix is calculated, scaled, and Softmax-normalized to obtain the attention weights. The attention weights are multiplied by the value matrix to obtain the output of that attention head. Finally, the outputs of all attention heads are concatenated and passed through a linear transformation to obtain the final output of the multi-head self-attention, which is the user interest feature data.

[0156] For example, assume that the dimension of the fused temporal features is 64, 8 attention heads are set, and the output dimension of each head is 8. Then the final output dimension of the multi-head self-attention is 64 (8×8).

[0157] Fuse the fused temporal feature data and the user interest feature data to obtain the scene prediction feature data. Feature fusion can be achieved through concatenation, addition, or gating mechanisms. In this embodiment, fusion is performed by concatenating followed by a fully connected layer.

[0158] Specifically, concatenate the fused temporal feature data and the user interest feature data in the feature dimension to obtain the concatenated feature. Then, perform a non-linear transformation on the concatenated feature through a fully connected layer to obtain the scene prediction feature data.

[0159] For example, assume that the dimension of the fused temporal features is 64 and the dimension of the user interest features is 64. Then the dimension of the concatenated feature is 128. Map the 128-dimensional feature to 96 dimensions through a fully connected layer to obtain the scene prediction feature data.

[0160] Use a multi-layer perceptron (MLP) to process the scene prediction feature data and output the predicted probability distribution of the target scene. The multi-layer perceptron consists of multiple fully connected layers, and a non-linear activation function is added between each layer.

[0161] In a specific implementation, the multi-layer perceptron includes two hidden layers and an output layer. The first hidden layer maps the 96-dimensional scene prediction feature to 64 dimensions, the second hidden layer maps the 64-dimensional feature to 32 dimensions, and the output layer maps the 32-dimensional feature to a vector of the dimension of the number of scene regions. Apply the Softmax function to the output vector to obtain the predicted probability distribution of each scene region.

[0162] For example, assume that the metaverse scene contains 10 regions. Then the dimension of the output layer is 10. After being processed by the Softmax function, the obtained predicted probability distribution may be [0.05, 0.12, 0.03, 0.42, 0.08, 0.07, 0.06, 0.09, 0.04, 0.04], indicating that the probability of the user accessing the 4th scene region next is the highest, which is 0.42.

[0163] According to the predicted probability distribution, select the scene region with the highest probability as the target scene region for the user to access next. Then, query the pre-established scene resource mapping table to obtain the storage location information of the scene resource data corresponding to the target scene region.

[0164] The scene resource mapping table stores the correspondence between scene areas and resource storage locations. For example: {"Technology Exhibition Area A": "cdn: / / resources / tech_zone_a / ", "Interactive Experience Area B": "cdn: / / resources / interactive_zone_b / "}.

[0165] Through the above method, it is possible to predict the target scene area that the user will visit next based on the user's historical interaction data, and obtain the corresponding scene resource data in advance, thereby optimizing the resource loading process and enhancing the user's interaction experience in the metaverse scene.

[0166] In an alternative embodiment, calculate the resource dependency relationship of the target scene area based on the organizational structure diagram, and generate a resource loading sequence; establish a resource index tree according to the resource loading sequence, and record the hierarchical relationship and association degree between scene resources in the resource index tree. Adaptive compression of scene resources based on the resource index tree includes:

[0167] Calculate the dependency strength between scene resources based on the scene resource data. The dependency strength is obtained through weighted calculation of functional dependency, temporal dependency, and content dependency. The functional dependency is calculated based on the resource call data, the temporal dependency is calculated based on the resource loading data, and the content dependency is calculated based on the resource content data;

[0168] Calculate the resource dependency relationship of the target scene area based on the organizational structure diagram. The resource dependency relationship is obtained by multiplying the dependency strengths on the dependency transfer path and applying a distance attenuation function. The distance attenuation function decreases exponentially with the length of the dependency transfer path;

[0169] Combine the dependency strength with the resource dependency relationship to obtain the loading priority of resource nodes, sort the resource nodes according to the loading priority, and generate a resource loading sequence;

[0170] Construct a resource index tree based on the resource loading sequence, record the hierarchical relationship of resource nodes in the resource index tree, and calculate the association metric value between resource nodes; determine the resource compression strategy according to the association metric value;

[0171] Map the association metric value to a compression ratio through the Sigmoid function; perform adaptive compression processing on the scene resources according to the compression ratio to obtain the compressed scene resource data;

[0172] Load the compressed scene resource data according to the resource loading sequence to achieve optimized loading of scene resources.

[0173] Figure 2 This is a schematic diagram of the adaptive compression process for metaverse scene resources. Exemplarily, an organizational structure diagram is a data structure that describes the organizational relationships between scene resources and includes the connection relationships between nodes. In practical applications, taking a game scene as an example, the organizational structure diagram can be represented as a directed graph, where nodes represent scene resources (such as models, textures, sound effects, etc.) and edges represent the dependency relationships between resources.

[0174] Calculating the dependency strength between scene resources is a key step, and the dependency strength is obtained through weighted calculation of functional dependency, temporal dependency, and content dependency. Specifically, the functional dependency is calculated based on resource call data. For example, the proportion of the number of times resource A is called by resource B to the total number of calls of B; the temporal dependency is calculated based on resource loading data, such as the reciprocal of the loading time interval between resource A and resource B; the content dependency is calculated based on resource content data, such as the content similarity between resource A and resource B.

[0175] Taking a certain game scene as an example, assume that the texture resource T1 is called 10 times by the model resource M1, and the total number of calls of M1 is 20 times. Then the functional dependency of T1 on M1 is 0.5; the loading time interval between T1 and M1 is 0.2 seconds, so the temporal dependency is 5; the content similarity between T1 and M1 is 0.3, so the content dependency is 0.3. If the weights of the three are 0.5, 0.3, and 0.2 respectively, then the dependency strength of T1 on M1 is 0.5×0.5 + 0.3×5 + 0.2×0.3 = 1.91.

[0176] The resource dependency relationship is obtained by multiplying the dependency strengths on the dependency transfer path and applying a distance attenuation function. The distance attenuation function decreases exponentially with the length of the dependency transfer path and can be expressed as the power of the attenuation factor to the path length. For example, if the attenuation factor is 0.8, the path length from resource A to resource C is 2 (A depends on B, and B depends on C), the dependency strength of A on B is 1.5, and the dependency strength of B on C is 2.0, then the resource dependency relationship of A on C is 1.5×2.0×0.8² = 1.92.

[0177] Combining the dependency strength and the resource dependency relationship to obtain the loading priority of resource nodes. The combination method can use weighted summation, such as priority = 0.6×dependency strength + 0.4×resource dependency relationship. Sort the resource nodes according to the loading priority to generate a resource loading sequence. For example, if the scene contains resources A, B, and C, and their loading priorities are 8.5, 6.2, and 9.1 respectively, then the resource loading sequence is C→A→B.

[0178] Construct a resource index tree based on the resource loading sequence, and record the hierarchical relationships of resource nodes in the resource index tree. The resource index tree is a tree-like data structure, where the root node is the resource with the highest priority, and the child nodes are resources that have a direct dependency relationship with the parent node. In the resource index tree, calculate the association metric values between resource nodes. The association metric values can be calculated based on factors such as the shortest path distance between nodes and the number of co-dependent resources.

[0179] Taking a virtual reality scene as an example, the root node of the resource index tree is the main scene model M0, and its child nodes include sub-models M1 and M2. The child nodes of M1 include textures T1 and T2, and the child nodes of M2 include texture T3. The association metric value between M1 and T1 is 0.9 (direct dependency), and the association metric value between M1 and T3 is 0.3 (indirect dependency).

[0180] Determine the resource compression strategy based on the association metric values. Map the association metric values to compression ratios through the Sigmoid function. The Sigmoid function can convert the association metric values into values between 0 and 1, representing the proportion of the original resources to be retained. For example, when the association metric value is 0.9, the compression ratio after mapping through the Sigmoid function is 0.8, indicating that 80% of the original resources are retained; when the association metric value is 0.3, the compression ratio is 0.4, indicating that 40% of the original resources are retained.

[0181] Perform adaptive compression processing on the scene resources according to the compression ratios to obtain the compressed scene resource data. For texture resources, compression can be achieved by reducing the resolution, reducing the color depth, etc.; for model resources, compression can be achieved by reducing the number of polygons, simplifying the bone structure, etc.; for audio resources, compression can be achieved by reducing the sampling rate, reducing the number of channels, etc.

[0182] Taking a mobile game as an example, the original texture T1 has a resolution of 1024×1024, a 32-bit color depth, and a size of 4MB; when the compression ratio is 0.8, the resolution can be reduced to 912×912, maintaining a 32-bit color depth, and the compressed size is approximately 3.2MB. The original model M1 contains 10,000 polygons, and when the compression ratio is 0.6, it can be reduced to 6,000 polygons.

[0183] Finally, load the compressed scene resource data according to the resource loading sequence to achieve optimized loading of the scene resources. In practical applications, the compression ratio can be dynamically adjusted according to the device performance and network conditions to further optimize the loading process. For example, the compression ratio can be increased on low-end devices and decreased on high-end devices; the compression ratio can be decreased when the network condition is good and increased when the network condition is poor.

[0184] Table 1 is a comparison table of the scene resource loading times. Figure 3It is a schematic diagram of the resource loading time distribution. The bar chart clearly shows the significant differences in the loading time distribution of different technical solutions. The bar of this technical solution significantly shifts to the left (i.e., shorter time), indicating that its loading performance is significantly better than the other three methods. Especially in the time period of 2 - 4 seconds, the sample proportion of this technical solution is the highest, meaning that most resource loadings can be completed in an extremely short time.

[0185] Table 1 is a comparison table of the scene resource loading times:

[0186]

[0187] Through the above method, the adaptive compression and optimized loading of scene resources can be achieved, improving the scene loading speed, reducing resource occupancy, and enhancing the user experience. In a virtual reality application test, after adopting this method, the scene loading time is reduced from the original 8.5 seconds to 3.2 seconds, and the memory occupancy is reduced from 450MB to 280MB, while maintaining a good visual effect and interaction experience.

[0188] In an optional implementation, a resource index tree is constructed based on the resource loading sequence, the hierarchical relationship of resource nodes is recorded in the resource index tree, and the association metric value between resource nodes is calculated; determining the resource compression strategy according to the association metric value includes:

[0189] Construct a resource index tree based on the resource loading sequence, determine the hierarchical value of the resource node by judging the parent - node relationship of the resource node, and the hierarchical value of the resource node is obtained through recursive calculation. For a resource node without a parent node, the hierarchical value is set to zero, and for a resource node with a parent node, the hierarchical value is set to the hierarchical value of its parent node plus one;

[0190] Calculate the resource node weight according to the hierarchical value of the resource node and the resource size. The resource node weight decays exponentially with the hierarchical value, and the resource node weight is positively correlated with the resource size;

[0191] Calculate the direct association strength between resource nodes. The direct association strength is obtained by multiplying the intersection - union ratio of the resource node connection set by the hierarchical difference penalty term, where the hierarchical difference penalty term maps the hierarchical difference to the interval from zero to two through the Sigmoid function;

[0192] Calculate the path association strength between resource nodes based on the direct association strength. The path association strength is obtained by multiplying the product of the direct association strengths on the path by the path decay function;

[0193] Weight the direct association strength and the path association strength to obtain the comprehensive association metric value between resource nodes;

[0194] Calculate the balance factor of resource nodes in the computing resource index tree. The balance factor is the difference between the height of the left subtree and the height of the right subtree of the resource node. When the absolute value of the balance factor is greater than the preset balance threshold, perform rebalancing processing on the resource index tree;

[0195] Calculate the basic compression rate of the resource node based on the comprehensive association metric value. The basic compression rate maps the difference between the average association degree of the resource node and the global average association degree to the interval from zero to one through the Sigmoid function;

[0196] Adaptively adjust the basic compression rate according to the resource node weight to obtain the adaptive compression rate, and the adaptive compression rate decreases as the resource node weight increases.

[0197] When the system loads resources, record the loading order of each resource and its parent-child relationship. For example, in a game scene, if model B is loaded immediately after scene A is loaded, there may be a parent-child relationship. Determine the level value by judging the parent node relationship between resource nodes, and adopt a recursive calculation method: for the root node (the resource node without a parent node), the level value is set to 0; for the resource node with a parent node, the level value is the level value of its parent node plus 1.

[0198] Taking the loading of game resources as an example, the main scene resource is the root node with a level value of 0; the character model resource in the main scene has a level value of 1; the weapon resource in the character model has a level value of 2. This level division reflects the inclusion relationship between resources and helps with subsequent association metric calculations.

[0199] The node weight calculation uses an exponential decay model, that is, the weight exponentially decays as the level value increases and is positively correlated with the resource size. In specific implementation, the weight calculation formula can be set as: node weight = resource size × decay factor ^ level value. For example, if the decay factor is set to 0.8, the weight of a 10MB resource node with a level value of 2 is 10 × 0.8^2 = 6.4, while the weight of an equally sized resource with a level value of 0 is 10.

[0200] Subsequently, calculate the direct association strength between resource nodes. The direct association strength is obtained by multiplying the intersection-over-union ratio of the resource node connection sets by the level difference penalty term. The intersection-over-union ratio of the connection sets reflects the proportion of the nodes jointly connected by two resource nodes in the total connected nodes. The level difference penalty term maps the level difference to the interval from 0 to 2 through the Sigmoid function. The smaller the level difference, the closer the penalty term is to 2; the larger the level difference, the closer the penalty term is to 0.

[0201] For example, if Resource A and Resource B have 3 common connection nodes, and their respective total connection nodes are 5 and 7, then the intersection-to-union ratio is 3 / 9 ≈ 0.33; if the level value of A is 1 and the level value of B is 3, and the level difference is 2, the corresponding penalty term may be 0.5, then the direct association strength is 0.33 × 0.5 = 0.165.

[0202] Based on the direct association strength, calculate the path association strength between resource nodes. The path association strength takes into account the multiple possible connection paths between resource nodes and is obtained by multiplying the direct association strengths on the path and the path attenuation function. The path attenuation function is related to the path length, and the longer the path, the more severe the attenuation.

[0203] Taking three resource nodes A, B, and C as an example, if the direct association strength from A to B is 0.8, the direct association strength from B to C is 0.6, the path length is 2, and the path attenuation factor is set to 0.9, then the path association strength from A to C is 0.8 × 0.6 × 0.9^2 ≈ 0.39.

[0204] Weight the direct association strength and the path association strength to obtain the comprehensive association metric value. The weight of the direct association strength can be set to 0.7, and the weight of the path association strength can be set to 0.3. Then, the comprehensive association metric value = 0.7 × direct association strength + 0.3 × path association strength.

[0205] To ensure the balance of the resource index tree, calculate the balance factor of each resource node, which is the difference between the height of the left subtree and the height of the right subtree. When the absolute value of the balance factor is greater than the preset balance threshold (usually 1 or 2), perform rebalancing on the resource index tree, which can be achieved by using operations such as left rotation, right rotation, or a combination of left and right rotations.

[0206] For example, if the height of the left subtree of node A is 4, the height of the right subtree is 1, and the balance factor is 3, exceeding the preset threshold of 2, then a right rotation operation is required to rebalance the tree. After rebalancing, the query efficiency is improved, which is beneficial to the subsequent calculation of the association metric value.

[0207] Calculate the basic compression rate of the resource node based on the comprehensive association metric value. First, calculate the average association degree of each resource node (the average of the association metric values with other nodes), and then calculate the global average association degree (the average of all node average association degrees). Map the difference between the resource node average association degree and the global average association degree to the interval from 0 to 1 through the Sigmoid function to obtain the basic compression rate.

[0208] For example, if the average association degree of resource node A is 0.6, the global average association degree is 0.4, and the difference is 0.2, the basic compression rate obtained through Sigmoid mapping may be 0.7, indicating that this resource can be compressed to 70% of its original size.

[0209] The adaptive compression ratio decreases as the weight of the resource node increases, ensuring that more details are retained for important resources (with high weights). The adjustment formula can be set as: Adaptive compression ratio = Base compression ratio × (1 - Weight adjustment factor × Normalized node weight).

[0210] Taking the weight adjustment factor as 0.5, the normalized weight of resource node A as 0.8, and the base compression ratio as 0.7 as an example, then the adaptive compression ratio = 0.7 × (1 - 0.5 × 0.8) = 0.7 × 0.6 = 0.42, indicating that the resource is finally compressed to 42% of its original size.

[0211] Table 2 is a comparison table of the comprehensive performance test results of different methods. As shown in Table 2, the technical solution of the present invention is in a leading position in each performance index, and the average improvement rate remains at a high level of 63%.

[0212] Table 2 is a comparison table of the comprehensive performance test results of different methods:

[0213]

[0214] The resource compression strategy constructed by the above method can adaptively adjust the compression ratio according to the association relationship and importance degree among resources, effectively reduce the overall resource volume while ensuring the quality of important resources, and improve the system performance.

[0215] In an alternative embodiment, when it is detected that the scene resource loading is abnormal, a backup channel for resource loading is started, and backup data of the abnormal resource is obtained through the backup channel; outputting the integrated scene resource data to the metaverse scene rendering engine includes:

[0216] Calculating the priorities of multiple backup channels, where the priorities are obtained by inputting the channel bandwidth, channel reliability, and channel delay into the Sigmoid function, and selecting the backup channel with the highest priority to load the abnormal resource;

[0217] Obtaining the backup data of the abnormal resource from the backup channel, and calculating the consistency check value between the abnormal resource and the backup data, where the consistency check value is calculated based on the hash value difference between the abnormal resource and the backup data;

[0218] When the consistency check value is less than the preset consistency threshold, using the backup data as the alternative data for the abnormal resource, and when the consistency check value is not less than the preset consistency threshold, performing data merging processing on the abnormal resource and the backup data to obtain the alternative data;

[0219] Calculate the integrity evaluation value of the scene resource data, where the integrity evaluation value is the ratio of the effective data volume of the scene resource data to the expected data volume; calculate the rendering adaptability of the scene resource data, where the rendering adaptability is calculated based on the difference between the resource format parameters and the target format parameters.

[0220] Under the constraint conditions that the integrity evaluation value is not lower than the minimum integrity threshold and the rendering adaptability is not lower than the minimum adaptability threshold, output the scene resource data to the metaverse scene rendering engine.

[0221] Scene resources include but are not limited to 3D models, texture maps, audio files, animation data, etc. When the system detects that a certain resource fails to load, times out, or is damaged, it is determined that the resource loading is abnormal. For example, when the system attempts to load the building model with the ID "model_building_01", if the loading fails to complete within the preset timeout of 5 seconds, or the data verification fails after loading, an exception handling process is triggered.

[0222] After detecting resource loading anomalies, the system activates the alternative channel selection mechanism. The system maintains multiple alternative channels, including but not limited to local cache channels, CDN backup channels, P2P distribution channels, etc. For each alternative channel, the system calculates its priority, and the priority calculation method is as follows:

[0223] For each alternative channel, the system obtains its current bandwidth value (e.g., 10MB / s), reliability value (e.g., 0.95, representing a 95% success rate), and channel delay value (e.g., 200ms). After normalizing these three parameters, they are converted into values between 0 and 1 through the Sigmoid function. Specifically, for the bandwidth value, it is divided by the preset maximum bandwidth value (such as 50MB / s) to obtain the normalized value; for the reliability value, it is directly used; for the delay value, it is divided by the preset maximum acceptable delay (such as 1000ms) and then inverted to obtain the normalized value. Then, the weighted sum of these three normalized values is input into the Sigmoid function to obtain the final priority value. The weights can be set as 0.4 for bandwidth, 0.4 for reliability, and 0.2 for delay.

[0224] For example, if an alternative channel has a bandwidth of 20MB / s, a reliability of 0.9, and a delay of 300ms, the normalized values are 0.4, 0.9, and 0.7 respectively, and the weighted sum is 0.4×0.4 + 0.4×0.9 + 0.2×0.7 = 0.66. After inputting into the Sigmoid function, the priority value is approximately 0.659. The system calculates the priority values of all alternative channels and selects the channel with the highest priority to load the abnormal resource.

[0225] After obtaining the backup data of the abnormal resource from the selected alternative channel, the system calculates the consistency check value between the abnormal resource and the backup data. The consistency check value is calculated based on the hash value difference between the abnormal resource and the backup data. Specifically, the system calculates the SHA-256 hash values of the abnormal resource and the backup data respectively, then calculates the Hamming distance (the number of different bits) between the two hash values, and then divides this distance by the total number of bits of the hash value (256) to obtain the consistency check value.

[0226] For example, if the hash value of the abnormal resource is "a1b2c3...", and the hash value of the backup data is "a1b2d4...", and the Hamming distance between them is 64 bits, then the consistency check value is 64 / 256 = 0.25. The system compares this value with a preset consistency threshold (such as 0.2).

[0227] When the consistency check value is less than the preset consistency threshold, it indicates that the backup data is basically consistent with the abnormal resource, and the system directly uses the backup data as the replacement data for the abnormal resource. For example, if the consistency check value is 0.15, which is less than the threshold 0.2, then the backup data is directly used.

[0228] When the consistency check value is not less than the preset consistency threshold, it indicates that there are significant differences between the backup data and the abnormal resource, and the system needs to perform data merging processing on the abnormal resource and the backup data to obtain the replacement data. Different strategies are adopted for data merging processing according to the resource type: for 3D models, a mesh repair algorithm can be used; for texture maps, an image repair algorithm can be used; for audio files, an audio interpolation algorithm can be used.

[0229] For example, for a damaged 3D model, the system can extract the intact mesh part from the abnormal resource, extract the corresponding missing part from the backup data, and then merge them into a complete model through boundary matching and smoothing processing. For texture maps, the system can use an image repair algorithm based on the Poisson equation to fuse the valid pixels in the abnormal resource with the corresponding area in the backup data.

[0230] After completing the processing of the abnormal resource, the system calculates the integrity evaluation value of the scene resource data. The integrity evaluation value is the ratio of the effective data volume of the scene resource data to the expected data volume. The system counts the total number of all resources in the scene (such as 100) and the number of successfully loaded resources (such as 95), and calculates the ratio to obtain the integrity evaluation value (such as 0.95).

[0231] Meanwhile, the system calculates the rendering adaptability of the scene resource data. The rendering adaptability is calculated based on the difference between the resource format parameters and the target format parameters. The system compares the format parameters of each resource (such as resolution, number of polygons, texture depth, etc.) with the format parameters required by the target rendering engine to calculate the degree of difference. For example, if the resolution of a texture resource is 1024×1024 and the target requirement is 2048×2048, the adaptability of this parameter is 0.5; if the number of polygons of a model is 10,000 and the target requirement is not more than 12,000, the adaptability of this parameter is 0.833. The system calculates the weighted average of the adaptabilities of all parameters to obtain the overall rendering adaptability.

[0232] Finally, the system checks whether the integrity evaluation value and the rendering adaptability meet the output conditions. If the integrity evaluation value is not lower than the minimum integrity threshold (such as 0.9) and the rendering adaptability is not lower than the minimum adaptability threshold (such as 0.8), the system outputs the processed scene resource data to the metaverse scene rendering engine; otherwise, the system issues a warning to the user indicating that the scene may not be rendered properly and provides a degraded rendering option.

[0233] The resource integration system in the construction of the metaverse scene based on the efficient engine includes:

[0234] The first unit is used to obtain the scene resource data in the metaverse scene, classify the scene resource data according to the data type, establish an organizational structure diagram of the scene resource data, and record the association relationship between the scene resource data in the organizational structure diagram; classify the scene resource data according to the association degree and usage frequency of the scene resource data, and store the scene resource data in the storage area corresponding to the corresponding level;

[0235] The second unit is used to analyze the historical interaction data of the user in the metaverse scene based on the deep learning model, output the target scene area that the user will visit next, and obtain the storage location information of the scene resource data corresponding to the target scene area;

[0236] The third unit is used to calculate the resource dependency relationship of the target scene area based on the organizational structure diagram, generate a resource loading sequence; establish a resource index tree according to the resource loading sequence, record the hierarchical relationship and association degree between the scene resources in the resource index tree, and perform adaptive compression on the scene resources based on the resource index tree, where the compression ratio is inversely proportional to the resource association degree, and the higher the association degree of the resource, the lower the compression ratio is used;

[0237] The fourth unit is used to monitor the loading status of scene resources in real time. When it detects an abnormal loading of scene resources, it starts a backup channel for resource loading and obtains backup data of the abnormal resources through the backup channel; it outputs the integrated scene resource data to the metaverse scene rendering engine for real-time rendering of the metaverse scene.

[0238] In the third aspect of the embodiments of the present invention,

[0239] A kind of electronic device is provided, including:

[0240] A processor;

[0241] A memory for storing executable instructions of the processor;

[0242] Wherein, the processor is configured to call the instructions stored in the memory to execute the method described above.

[0243] In the fourth aspect of the embodiments of the present invention,

[0244] A computer-readable storage medium is provided, on which computer program instructions are stored, and when the computer program instructions are executed by a processor, the method described above is implemented.

[0245] The present invention can be a method, a device, a system and / or a computer program product. The computer program product may include a computer-readable storage medium, on which computer-readable program instructions for executing various aspects of the present invention are uploaded.

[0246] Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it is still possible to modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements on some or all of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A resource integration method in the construction of a metaverse scene based on an efficient engine, characterized in that, Including: Obtain the scene resource data in the metaverse scene, classify the scene resource data according to the data type, establish an organizational structure diagram of the scene resource data, and record the association relationship between the scene resource data in the organizational structure diagram; classify the scene resource data according to the association degree and usage frequency of the scene resource data, and store the scene resource data in the storage area corresponding to the corresponding level; Analyze the historical interaction data of the user in the metaverse scene based on the deep learning model, output the target scene area that the user will access next, and obtain the storage location information of the scene resource data corresponding to the target scene area; Calculate the resource dependence relationship of the target scene area based on the organizational structure diagram, and generate a resource loading sequence; Establish a resource index tree according to the resource loading sequence, record the hierarchical relationship and association degree between scene resources in the resource index tree, and perform adaptive compression on the scene resources based on the resource index tree, where the compression ratio is inversely proportional to the resource association degree, and the higher the association degree of the resource, the lower the compression ratio is used; Monitor the loading status of the scene resources in real time. When it is detected that the scene resources are loaded abnormally, start the resource loading backup channel, and obtain the backup data of the abnormal resources through the backup channel; output the integrated scene resource data to the metaverse scene rendering engine for real-time rendering of the metaverse scene; Calculate the dependence strength between scene resources based on the scene resource data, and the dependence strength is obtained by weighted calculation of the functional dependence degree, the timing dependence degree, and the content dependence degree, where the functional dependence degree is calculated based on the resource call data, the timing dependence degree is calculated based on the resource loading data, and the content dependence degree is calculated based on the resource content data; Calculate the resource dependence relationship of the target scene area based on the organizational structure diagram, and the resource dependence relationship is obtained by multiplying the dependence strengths on the dependence transfer path and applying a distance attenuation function, and the distance attenuation function decreases exponentially with the length of the dependence transfer path; Combine the dependence strength and the resource dependence relationship to obtain the loading priority of the resource nodes, sort the resource nodes according to the loading priority, and generate a resource loading sequence; Construct a resource index tree based on the resource loading sequence, record the hierarchical relationship of the resource nodes in the resource index tree, and calculate the association metric value between the resource nodes; determine the resource compression strategy according to the association metric value; Map the association metric value to the compression ratio through the Sigmoid function; perform adaptive compression processing on the scene resources according to the compression ratio to obtain the compressed scene resource data; Load the compressed scene resource data according to the resource loading sequence to realize the optimized loading of the scene resources.

2. The method according to claim 1, wherein Classify the scene resource data according to the data type, and establish an organizational structure diagram of the scene resource data. The association relationship between the scene resource data recorded in the organizational structure diagram includes: The scene resource data is divided into geometric data, material data, environmental data, and interaction data according to the data representation form. The geometric data includes model mesh data and vertex coordinate data. The material data includes material texture data and normal map data. The environmental data includes lighting information data and shadow information data. The interaction data includes collider data and trigger data; The scene resource data is divided into basic element layer data, combined element layer data, and scene element layer data according to the data granularity. The basic element layer data includes single model data and single texture data. The combined element layer data includes model combination data and material combination data. The scene element layer data includes complete scene unit data; Calculate the association strength between the scene resource data; construct an organizational structure diagram of the scene resource data based on the association strength. In the organizational structure diagram of the scene resource data, the scene resource data is used as nodes, and the association relationship between the scene resource data is used as connection edges. The weight of the connection edges is determined by the association strength.

3. The method according to claim 2, wherein Classify the scene resource data according to the association degree and usage frequency of the scene resource data, and store the scene resource data in the corresponding storage area, including: Calculate the priority index of the scene resource data. The priority index is obtained by weighting the usage frequency factor, calculation complexity factor, and association degree factor. The usage frequency factor is determined according to the access times. The calculation complexity factor is determined according to the resource processing overhead. The association degree factor is determined according to the average association strength; Divide the scene resource data into different levels of storage areas according to the priority index. Store the scene resource data with a priority index greater than the first classification threshold in the high-priority storage area. Store the scene resource data with a priority index greater than the second classification threshold and less than the first classification threshold in the medium-priority storage area. Store the scene resource data with a priority index less than the second classification threshold in the low-priority storage area.

4. The method according to claim 1, characterized in that Analyze the historical interaction data of the user in the metaverse scene based on the deep learning model, output the target scene area that the user will access next, and obtain the storage location information of the scene resource data corresponding to the target scene area, including: Obtain the historical interaction data of the user in the metaverse scene. The historical interaction data includes spatio-temporal dimension feature data and interaction behavior feature data. The spatio-temporal dimension feature data includes user movement trajectory sequence data and scene residence time distribution data. The interaction behavior feature data includes operation sequence coding data and interaction intensity matrix data; Process the historical interaction data using a bidirectional long short-term memory network, obtain temporal features through the combination of forward hidden states and backward hidden states, and perform weighted fusion on the temporal features through learnable parameters to obtain fused temporal feature data; Construct a user interest model based on the multi-head self-attention mechanism, map the fused temporal feature data into a query matrix, a key matrix, and a value matrix, calculate the attention weight distribution, and obtain user interest feature data through multi-head splicing; Fuse the fused time-series feature data with the user interest feature data to obtain scene prediction feature data; Process the scene prediction feature data using a multi-layer perceptron to output the predicted probability distribution of the target scene.

5. The method according to claim 1, wherein Construct a resource index tree based on the resource loading sequence, record the hierarchical relationship of resource nodes in the resource index tree, and calculate the association metric value between resource nodes; Determine the resource compression strategy according to the association metric value, including: Construct a resource index tree based on the resource loading sequence, determine the hierarchical value of the resource node by judging the parent-child relationship of the resource node, and the hierarchical value of the resource node is obtained by recursive calculation. For a resource node without a parent node, set the hierarchical value to zero. For a resource node with a parent node, set the hierarchical value to the hierarchical value of its parent node plus one; Calculate the resource node weight according to the hierarchical value and resource size of the resource node. The resource node weight decays exponentially with the hierarchical value, and the resource node weight is positively correlated with the resource size; Calculate the direct association strength between resource nodes. The direct association strength is obtained by multiplying the intersection-over-union of the resource node connection set by the hierarchical difference penalty term, where the hierarchical difference penalty term maps the hierarchical difference to the interval from zero to two through the Sigmoid function; Calculate the path association strength between resource nodes based on the direct association strength. The path association strength is obtained by multiplying the product of the direct association strengths on the path by the path decay function; Weight the direct association strength and the path association strength to obtain the comprehensive association metric value between resource nodes; Calculate the balance factor of the resource nodes in the resource index tree. The balance factor is the difference between the height of the left subtree and the height of the right subtree of the resource node. When the absolute value of the balance factor is greater than the preset balance threshold, rebalance the resource index tree; Calculate the basic compression rate of the resource node based on the comprehensive association metric value. The basic compression rate maps the difference between the average association degree of the resource node and the global average association degree to the interval from zero to one through the Sigmoid function; Adaptively adjust the basic compression rate according to the resource node weight to obtain the adaptive compression rate. The adaptive compression rate decreases as the resource node weight increases.

6. The method according to claim 1, wherein When a scene resource loading anomaly is detected, activate the resource loading backup channel and obtain the backup data of the abnormal resource through the backup channel; Output the integrated scene resource data to the metaverse scene rendering engine, including: Calculate the priorities of multiple backup channels. The priorities are obtained by inputting the channel bandwidth, channel reliability, and channel delay into the Sigmoid function, and select the backup channel with the highest priority to load the abnormal resource; Obtain the backup data of the abnormal resource from the backup channel and calculate the consistency check value between the abnormal resource and the backup data. The consistency check value is calculated based on the hash value difference between the abnormal resource and the backup data; When the consistency check value is less than the preset consistency threshold, the backup data is used as the replacement data for the abnormal resource. When the consistency check value is not less than the preset consistency threshold, data merging processing is performed on the abnormal resource and the backup data to obtain the replacement data; Calculate the integrity evaluation value of the scene resource data, where the integrity evaluation value is the ratio of the effective data volume to the expected data volume of the scene resource data; calculate the rendering adaptability of the scene resource data, where the rendering adaptability is calculated based on the difference between the resource format parameters and the target format parameters; Under the constraint conditions that the integrity evaluation value is not lower than the minimum integrity threshold and the rendering adaptability is not lower than the minimum adaptability threshold, the scene resource data is output to the metaverse scene rendering engine.

7. A resource integration system in the construction of a metaverse scenario based on an efficient engine, which is used to implement the method described in any one of claims 1-6, characterized in that, Comprising: The first unit is used to obtain the scene resource data in the metaverse scene, classify the scene resource data according to the data type, establish an organizational structure diagram of the scene resource data, and record the association relationship between the scene resource data in the organizational structure diagram; classify the scene resource data according to the association degree and usage frequency of the scene resource data, and store the scene resource data in the storage area corresponding to the corresponding level; The second unit is used to analyze the historical interaction data of the user in the metaverse scene based on the deep learning model, output the target scene area that the user will access next, and obtain the storage location information of the scene resource data corresponding to the target scene area; The third unit is used to calculate the resource dependency relationship of the target scene area based on the organizational structure diagram and generate a resource loading sequence; Establish a resource index tree according to the resource loading sequence, record the hierarchical relationship and association degree between the scene resources in the resource index tree, and perform adaptive compression on the scene resources based on the resource index tree, where the compression ratio is inversely proportional to the resource association degree, and the higher the association degree of the resource, the lower the compression ratio is used; The fourth unit is used to monitor the loading status of the scene resources in real time. When it is detected that the scene resources are loaded abnormally, start the resource loading backup channel, and obtain the backup data of the abnormal resources through the backup channel; output the integrated scene resource data to the metaverse scene rendering engine for real-time rendering of the metaverse scene.

8. An electronic device, characterized in that, Comprising: A processor; A memory for storing instructions executable by the processor; Wherein, the processor is configured to call the instructions stored in the memory to execute the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having computer program instructions stored thereon, characterized in that, When the computer program instructions are executed by the processor, the method according to any one of claims 1 to 6 is implemented.

Images