Regional image processing method and system for display control of virtual exhibition hall

Abstract

The invention discloses a regional image processing method and system for display control of a virtual exhibition hall, and relates to the technical field of image processing, and the method comprises the steps: constructing a double-layer behavior model comprising a cosexual behavior thermodynamic diagram and an individual behavior portrait; establishing a three-dimensional space region under the view angle of the current user based on the model, and dividing the three-dimensional space region into a central region and a neighborhood region; making a thermodynamic response strategy in combination with the coexistence behavior thermodynamic diagram, performing hierarchical decision on the three-dimensional space region, and generating a reference image processing scheme; according to the scheme, the display control of the virtual exhibition hall is executed in the three-dimensional space. Therefore, the technical effects of improving the flexibility and rationality of resource allocation, and improving the display efficiency and the user experience are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of image processing, and particularly to a regional image processing method and system for virtual exhibition hall display control. Background Art

[0002] Existing virtual exhibition hall display technologies usually rely on fixed rendering strategies and display the content of the exhibition hall through preset image processing schemes. These schemes are mainly based on hardware performance optimization or simple user interaction triggering mechanisms, such as switching scenes by mouse clicks or gesture swipes. However, with the expansion of the scale of virtual exhibition halls and the diversification of user behaviors, the existing technologies have gradually exposed limitations in resource allocation and user experience optimization. There are technical problems such as lack of pertinence in resource allocation, resulting in waste of computing resources, and affecting display efficiency and user experience. Summary of the Invention

[0003] The present invention provides a regional image processing method and system for virtual exhibition hall display control to solve the technical problems in the prior art of lack of pertinence in resource allocation, resulting in waste of computing resources, and affecting display efficiency and user experience, and to achieve the technical effects of improving the flexibility and rationality of resource allocation and enhancing display performance and user experience.

[0004] In a first aspect, the present invention provides a regional image processing method for virtual exhibition hall display control, wherein the regional image processing method for virtual exhibition hall display control includes: Construct a two-layer behavior model based on user behavior, wherein the two-layer behavior model includes a common behavior heat map and a personalized behavior portrait.

[0005] Establish a three-dimensional space region from the perspective of the current user according to the two-layer behavior model, wherein the three-dimensional space region is divided into a central space region and a neighborhood space region.

[0006] Construct a heat response strategy, and combine the common behavior heat map to make a hierarchical processing decision for the three-dimensional space region, and output a decision result as a reference regional image processing scheme.

[0007] Based on the reference regional image processing scheme, combine the three-dimensional space region to control the display of the virtual exhibition hall.

[0008] In a feasible implementation manner, constructing a two-layer behavior model based on user behavior, wherein the two-layer behavior model includes a common behavior heat map and a personalized behavior portrait, includes: Obtain the line-of-sight trajectories and operation behavior data of multiple users and analyze to obtain user interaction behavior data, wherein the user interaction behavior data includes line-of-sight trajectory information and operation action information.

[0009] Based on the line-of-sight trajectory information, statistically analyze to obtain the line-of-sight trajectory distribution in the virtual exhibition hall area, and correspondingly generate the common behavior heat map.

[0010] Based on the operation action information, construct the personalized behavior portrait in combination with feature engineering.

[0011] Output the common behavior heat map and the personalized behavior portrait as the double-layer behavior model.

[0012] In a feasible implementation manner, constructing the personalized behavior portrait based on the operation action information in combination with feature engineering includes: According to the unique identification identity of the user, cluster and divide the operation action information.

[0013] Perform feature extraction on the clustering and division results to obtain multiple groups of key behavior feature data.

[0014] Calculate the perspective turning rate, action randomness entropy value, and line-of-sight residence time distribution corresponding to multiple groups of the key behavior feature data respectively, and vectorize the calculation results as the personalized behavior portrait, wherein the personalized behavior portrait is associated with a unique identification identity marked.

[0015] In a feasible implementation manner, according to the double-layer behavior model, establish a three-dimensional space area under the current user's perspective, wherein the three-dimensional space area is divided into a central space area and a neighborhood space area, including: Based on the common behavior heat map and the current user's perspective, crop the area where the heat value is lower than the omitted heat limit to determine the central space area.

[0016] Based on the central space area, perform neighborhood growth in combination with the personalized behavior portrait to determine the corresponding neighborhood space area, wherein the neighborhood growth radius is calculated and determined based on the personalized behavior portrait.

[0017] Combine and smooth the central space area and the neighborhood space area.

[0018] In a feasible implementation manner, construct a heat response strategy, and perform hierarchical processing decision-making on the three-dimensional space area in combination with the common behavior heat map, and output the decision result as the image processing scheme for the reference area, including: According to the processing performance parameters and hardware function information of the target scene, construct the heat response strategy, wherein the heat response strategy includes multiple rendering sub-strategies, and each of the rendering sub-strategies has a differential configuration in terms of image processing accuracy and computing resource scheduling.

[0019] Using the semantic information of the preset area, combined with the common behavior heat map, classify and mark the interest level of the three-dimensional space area.

[0020] According to the mapping relationship in the thermal response strategy, match the image rendering sub-strategy corresponding to the interest level, and output the matching result as the image processing solution for the reference area.

[0021] In a feasible implementation manner, based on the image processing solution for the reference area, combined with the three-dimensional space area, perform display control of the virtual exhibition hall, including: Perform object recognition on the central space area and the neighborhood space area respectively to determine the resource call target and the resource pre-storage target.

[0022] Analyze the image processing solution for the reference area, and extract the first call accuracy corresponding to the resource call target and the second call accuracy corresponding to the resource pre-storage target.

[0023] According to the resource call target and the first call accuracy, address and read the corresponding instant digital display resource for image rendering and display control.

[0024] Synchronously, according to the resource pre-storage target and the second call accuracy, read the corresponding predicted digital display resource and store it in the cache medium.

[0025] In a feasible implementation manner, after performing display control of the virtual exhibition hall based on the image processing solution for the reference area, combined with the three-dimensional space area, it further includes: If it is detected that multiple users are located at the same spatial position coordinates, extract the display area images corresponding to each user and perform comparative analysis.

[0026] Based on the comparison result of the display area images, identify the invariant feature information, and according to the invariant feature information, construct a shared image rendering context, where the shared image rendering context is used for fast image reconstruction and image reuse among multiple users.

[0027] In a second aspect, the present invention also provides a regional image processing system for virtual exhibition hall display control, where the regional image processing system for virtual exhibition hall display control includes: A double-layer behavior model construction module for constructing a double-layer behavior model based on user behavior, where the double-layer behavior model includes a common behavior heat map and a personalized behavior portrait.

[0028] A region establishment module for establishing a three-dimensional space region from the perspective of the current user according to the double-layer behavior model, where the three-dimensional space region is divided into a central space region and a neighborhood space region.

[0029] An image processing decision-making module, configured to construct a thermal response strategy, and perform hierarchical processing decision-making on the three-dimensional space region in combination with the common behavior heat map, and output a decision result as an image processing solution for the reference region.

[0030] A display control module, configured to perform display control of the virtual exhibition hall based on the image processing solution for the reference region and in combination with the three-dimensional space region.

[0031] The present invention discloses a method and system for regional image processing for virtual exhibition hall display control, including: constructing a two-layer behavior model based on user behavior, which includes a common behavior heat map and a personalized behavior portrait; on this basis, establishing a three-dimensional space region from the perspective of the current user according to the two-layer behavior model, and dividing the space region into a central space region and a neighborhood space region; subsequently, constructing a thermal response strategy, and performing hierarchical processing decision-making on the three-dimensional space region in combination with the common behavior heat map, and the output result is an image processing solution for the reference region; finally, based on this image processing solution, implementing display control of the virtual exhibition hall in the three-dimensional space region. The method and system for regional image processing for virtual exhibition hall display control disclosed by the present invention solve the technical problems of lack of pertinence in resource allocation, resulting in waste of computing resources, and affecting display efficiency and user experience, and achieve the technical effects of improving the flexibility and rationality of resource allocation, and improving display efficiency and user experience. Description of the Drawings

[0032] Figure 1 It is a schematic flowchart of a method for regional image processing for virtual exhibition hall display control according to the present invention.

[0033] Figure 2 It is a schematic structural diagram of a system for regional image processing for virtual exhibition hall display control according to the present invention.

[0034] Description of the reference numerals: the two-layer behavior model construction module 11, the region establishment module 12, the image processing decision-making module 13, the display control module 14. Detailed Embodiments

[0035] The following will describe the above technical solutions in detail in combination with the accompanying drawings of the specification and specific embodiments to better understand the above technical solutions. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without making creative efforts belong to the scope of protection of the present invention. In addition, it should be noted that for the sake of description, only parts related to the present invention are shown in the drawings rather than all.

[0036] Embodiment 1, as follows Figure 1 This is a schematic flowchart of a regional image processing method for virtual exhibition hall display control according to the present invention. Among them, the regional image processing method for virtual exhibition hall display control includes: Construct a two - layer behavior model based on user behavior, where the two - layer behavior model includes a common behavior heat map and a personalized behavior portrait.

[0037] Specifically, the two - layer behavior model contains two levels of behavior analysis models (common behavior heat map and personalized behavior portrait), so as to capture the common behaviors of the user group and the personalized behaviors of individual users simultaneously. Among them, the common behavior heat map is generated by statistical analysis of the behavior data of multiple users, and is used to reflect the concentration trend and hot areas of user behavior in the virtual exhibition hall. Preferably, it usually represents the frequency or intensity of behavior occurrence with the depth of color. The personalized behavior portrait is the modeling result of the behavior characteristics of a single user. By analyzing the operation action information of the user, characteristic data reflecting the unique behavior pattern of the user, such as the viewing angle turning rate, action randomness entropy value, and line - of - sight residence time distribution, can be extracted, and they are vectorized and represented as the personalized behavior portrait.

[0038] In the above steps, the common behavior heat map can help the system identify the areas generally concerned by the user group, so that these areas are preferentially considered during resource allocation, avoiding blind resource allocation and waste. At the same time, the personalized behavior portrait enables the system to perform personalized display control according to the specific behavior characteristics of each user, further improving the user experience. The above - mentioned behavior modeling method combining commonness and individuality can improve the display efficiency and then optimize the overall performance of the virtual exhibition hall.

[0039] In some embodiments, constructing a two - layer behavior model based on user behavior, where the two - layer behavior model includes a common behavior heat map and a personalized behavior portrait, includes: Obtain the line - of - sight trajectories and operation behavior data of multiple users and analyze to obtain user interaction behavior data, where the user interaction behavior data includes line - of - sight trajectory information and operation action information; based on the line - of - sight trajectory information, statistically analyze to obtain the line - of - sight trajectory distribution of the virtual exhibition hall area, and correspondingly generate the common behavior heat map; based on the operation action information, construct the personalized behavior portrait by combining feature engineering; output the common behavior heat map and the personalized behavior portrait as the two - layer behavior model.

[0040] Specifically, user interaction behavior data refers to various data generated when users interact with the system in the virtual exhibition hall, including historical user sight trajectory information and operation action information. Among them, sight trajectory information refers to the path and stop position of the user's sight movement in the virtual exhibition hall (which can be represented as a time-series data sequence of the sight direction vector), reflecting the user's attention to different areas; operation action information includes various operation behaviors of users in the virtual exhibition hall, such as clicking, sliding, zooming, etc. These operation behaviors can reflect the user's interests and interaction habits.

[0041] Specifically, first, eye tracking devices or human-computer interaction devices record the user's line of sight, and at the same time record the user's operation actions, such as clicks and slides, through system logs. For example, in a virtual art exhibition hall, it can record the user's stay time in front of a painting, the path of the line of sight, and whether the user clicks to enlarge the painting. The above data together constitute the user interaction behavior data, which provides basic data support for subsequent model construction.

[0042] Specifically, the collected sight trajectory information is then statistically analyzed, that is, by calculating the dwell frequency and cumulative dwell time of all sights in different areas, a common behavior heat map is generated. For example, in a virtual exhibition hall, if most users' sights are focused on certain paintings, these areas will appear as high-heat areas on the heat map. These areas can be considered to be the focus of general attention of the user group, and may require higher attention or image processing capabilities. The above-mentioned common behavior heat map can intuitively reflect the common behavioral characteristics of the user group, providing an important basis for subsequent resource allocation and display control.

[0043] Specifically, by extracting and converting the features of the operation action information, key data that effectively describes the user's behavior characteristics can be extracted, such as the viewing angle turning rate, the action randomness entropy value, and the distribution of the sight dwelling time, etc.; then, these features are vectorized to form a personality behavior portrait, in which each feature indicator corresponds to a vector dimension, that is, the personality behavior portrait is represented as a multidimensional vector. For example, for a certain user, his personality behavior portrait may show that he prefers to quickly browse multiple exhibits and has a higher frequency of attention to certain specific types of exhibits. The personality behavior portrait obtained through the above process can accurately reflect the personalized characteristics of individual users' habits, preferences, etc., which helps to provide each user with a personalized display experience.

[0044] Furthermore, the generated common behavior heat map and individual behavior portrait are integrated into a two-layer behavior model as a decision basis for subsequent display control and resource allocation. The construction of the above two-layer behavior model not only helps to improve the flexibility and rationality of resource allocation, but also helps to improve display efficiency and user experience.

[0045] In some implementations, constructing the personality behavior portrait based on the operation action information in combination with feature engineering includes: According to the user's unique identification identity, the operation action information is clustered; feature extraction is performed on the clustering result to obtain multiple groups of key behavior feature data; the perspective turning rate, action randomness entropy value and gaze dwell time distribution corresponding to the multiple groups of key behavior feature data are calculated respectively, and the calculation result is vectorized into the personality behavior portrait, wherein the personality behavior portrait is associated with a unique identification identity.

[0046] Specifically, first, the operation action information generated by a specific user in the system is extracted, and the user's operation action information is grouped according to the user identity based on the unique identification identity bound to it, to ensure that the behavior data of each user is processed independently; then, feature extraction is performed on each type of operation sample after clustering, and multiple groups of key behavior feature data are constructed, such as the viewing angle and frequency sequence, the curvature of the gesture trajectory, the start and end time difference of the operation, the number of turning points in the eye movement path, etc. In other words, the key behavior feature data can be considered as the basic representation of the user's personality.

[0047] Furthermore, based on the above key behavioral feature data, the perspective turning rate (average and peak values ​​of angular velocity change) corresponding to each set of features is calculated; the action randomness entropy value of the behavior sequence is calculated (such as Shannon entropy and sample entropy, which measure the regularity of operations); and the distribution of gaze dwell time (average gaze time and dwell frequency on different areas or interface elements) is statistically analyzed. The above multiple indicators can be considered as a more comprehensive representation of user personality based on key behavioral feature data.

[0048] Furthermore, the extracted key behavioral feature data are normalized and quantified to form a behavioral feature vector with a fixed dimension as a personality behavior portrait to facilitate subsequent model construction and analysis, wherein the personality behavior portrait is associated with a unique identification tag to facilitate subsequent display control and resource allocation.

[0049] Through the above process, a personalized behavior portrait is constructed based on the operation action information, which realizes the refined and personalized analysis of user behavior and provides data support for subsequent personalized display. Among them, the key behavior feature data is quantified to form a personalized behavior portrait, which can effectively reflect the user's operation habits and preferences, thereby realizing the precise allocation of resources and the optimization of display strategies.

[0050] According to the two-layer behavior model, a three-dimensional space region under the current user's perspective is established, wherein the three-dimensional space region is divided into a central space region and a neighborhood space region.

[0051] Specifically, the current user perspective refers to the observation or operation position and direction of the user at a specific interaction time node, which is determined by perspective coordinates (such as camera position + orientation vector), eye tracking, or interface focus information; the three-dimensional space region is a spatial region structure constructed around the current user perspective, used to map the attention intensity or the possibility of behavioral response of the user to different spatial positions in the environmental scene, and guide the range that subsequent image processing needs to focus on.

[0052] Specifically, the three-dimensional space region includes a central space region and a neighborhood space region. Among them, the central space region corresponds to the high-attention space region near the user's line of sight or operation focus; the neighborhood space region is the low-attention region distributed around the central space region, representing the environmental space range that may be noticed or indirectly affected; through the dynamic modeling of the neighborhood space, it helps to prepare in advance for the user's possible behavior trends such as turning and attention switching, thereby enhancing the response forward-looking and personalized adaptability of the interaction system.

[0053] In some embodiments, according to the double-layer behavior model, a three-dimensional space region under the current user perspective is established, where the three-dimensional space region is divided into a central space region and a neighborhood space region, including: Based on the common behavior heat map and the current user perspective, crop the regions where the heat value is lower than the omitted heat limit to determine the central space region; based on the central space region, combine the personalized behavior portrait for neighborhood growth to determine the corresponding neighborhood space region, where the neighborhood growth radius is calculated based on the personalized behavior portrait; combine and smooth the central space region and the neighborhood space region.

[0054] Specifically, the heat value refers to the value representing the occurrence frequency or intensity of the behavior in the heat map; the omitted heat limit refers to a set threshold used to determine which regions have lower heat values and can be ignored.

[0055] Specifically, first, through the common behavior heat map and the current user perspective, determine the core region that the user is currently concerned about. For example, in a virtual exhibition hall, the system combines the common behavior heat map according to the current perspective position and direction of the user, crops the regions where the heat value is lower than the set threshold (i.e., the omitted heat limit) (such as pure black or fixed background regions), and retains the high-heat regions that the user is concerned about as the central space region. This cropping can ensure that resources are concentrated in the regions that the user truly cares about and avoid unnecessary calculations for the unconcerned regions.

[0056] Specifically, then, with the central space region as the core, combined with the personalized behavior portrait of the current user, analyze the attention transfer tendency and behavior extension range, dynamically calculate the neighborhood growth radius, and perform neighborhood expansion operations in the three-dimensional space; Exemplarily, the greater the perspective turning rate, the greater the action randomness entropy value, and the more dispersed the line-of-sight residence time distribution, the greater the user's behavior randomness, and the corresponding growth radius is larger. For example, if the user prefers to browse quickly, the neighborhood growth radius is appropriately expanded to include more potentially interesting content.

[0057] Furthermore, merge the central space region and the neighborhood space region, and perform smoothing processing through interpolation algorithms or boundary smoothing techniques to ensure a natural transition at the boundary between the central region and the neighborhood region, avoiding an abrupt sense of boundary; The above processing not only improves the visual effect but also makes the user's browsing experience smoother.

[0058] The above process establishes a three-dimensional space region based on a two-layer behavior model, which can significantly improve the flexibility and rationality of resource allocation. First, cut off the low heat value regions to ensure that resources are concentrated in the areas that the user truly pays attention to, avoiding resource waste; The neighborhood growth process dynamically adjusts the neighborhood range according to the personalized behavior portrait, making the displayed content more in line with the user's personalized needs; Combining and smoothing processing ensures a natural transition at the region boundary, enhancing the fluency and aesthetics of the display effect. Such a region division method based on user behavior not only helps improve the display efficiency but also significantly enhances the user experience.

[0059] Construct a thermal response strategy, and combine the common behavior heat map to make a hierarchical processing decision for the three-dimensional space region, and output the decision result as the image processing scheme for the reference region.

[0060] Specifically, the thermal response strategy is a dynamic resource allocation strategy based on the user behavior heat map, which can allocate different processing resources and rendering qualities according to the importance of different space regions to achieve a balance between processing efficiency and user experience. Among them, according to the interest level or importance of the region, the three-dimensional space region can be prioritized, so as to allocate corresponding image processing schemes for each level.

[0061] Specifically, the image processing scheme for the reference region is the final image processing scheme output according to the hierarchical processing decision, which is used to guide the display control of the virtual exhibition hall, including multiple hierarchical region units and corresponding display control methods, control parameters, etc.

[0062] In some embodiments, constructing a thermal response strategy, and combining the common behavior heat map to make a hierarchical processing decision for the three-dimensional space region, and outputting the decision result as the image processing scheme for the reference region includes: Construct the thermal response strategy according to the processing performance parameters and hardware function information of the target scenario, where the thermal response strategy includes multiple rendering sub-strategies, and each rendering sub-strategy has differential configurations in terms of image processing accuracy and computing resource scheduling; use the preset regional semantic information to combine with the common behavior heat map to divide and mark the interest levels of the three-dimensional space region; according to the mapping relationship in the thermal response strategy, match the image rendering sub-strategy corresponding to the interest level, and output the matching result as the image processing solution for the reference region.

[0063] Specifically, the rendering sub-strategy is a subset strategy in the thermal response strategy, with different parameter configurations such as image rendering accuracy, computing complexity, resource occupancy, etc. Among them, high-quality rendering has rich details but high resource consumption, such as full-scale rendering with high resolution, high frame rate, high texture, and high shadow; medium-quality rendering balances performance and accuracy, such as rendering based on or low-quality rendering (fast response but detail compression).

[0064] Specifically, the regional semantic information is used to describe the semantic attributes of each sub-region in the three-dimensional space, such as "operation area", "observation area", "edge area", etc., providing a semantic basis for the interest level division.

[0065] Exemplarily, the high-quality rendering sub-strategy uses the original high-resolution image or an equivalent high-precision image source; retains complete details such as texture, shadow, reflection, etc., and is suitable for the central high-interest area with long user gaze time and frequent interactions; is applicable to scenarios with sufficient terminal computing power, such as: using a high-throughput graphics card renderer, virtual reality headset, etc.; the medium-quality rendering sub-strategy needs to comprehensively consider processing efficiency and visual fidelity, such as using new neural network processing methods such as AI super-resolution algorithm, AI frame interpolation model, etc. (based on CUDA or tensor cores) to obtain an approximate high-quality effect with medium image resource input; among them, this strategy can be used for some medium-interest areas, especially suitable for optimization processing in the AI-accelerated hardware environment; the low-quality rendering sub-strategy is mainly applicable to the edge or low-attention areas, and traditional image enhancement methods such as frame interpolation method based on optical flow estimation, edge anti-aliasing algorithm, etc. can be used for fast compression rendering, which can effectively reduce the processing load, improve the system response speed, and at the same time support the operation of devices with lower hardware specifications and resource-sensitive devices.

[0066] Specifically, first, collect the image processing performance parameters of the target usage scenario (such as the upper limit of frame rate, GPU computing power, cache bandwidth, etc.) and the terminal hardware functions (such as graphics card model, parallel processing ability); establish a responsiveness model based on the above parameters, and construct a thermal response strategy that includes multiple rendering sub-strategies, each sub-strategy having different image processing levels (such as: resolution level, detail level, texture mapping frequency, etc.); among them, the thermal response strategy has the ability of strategy-resource mapping, and can match the region of interest and the resource allocation mode as needed; then, for the three-dimensional space region constructed in the previous stage, perform weight calculation (such as weighted sum) on the thermal values in the regional semantic information (such as identifying "user's permanent perspective", "frequently operated object") and the common behavior heat map, and output the interest level division result, which is used to guide the differential image rendering process.

[0067] Furthermore, based on the level-strategy mapping table in the thermal response strategy, automatically match the appropriate rendering sub-strategies for different interest levels, and combine the matching results of all sub-strategies to generate a benchmark regional image processing scheme, which is used to indicate the application of different image processing means at different spatial positions to achieve the optimal allocation of resources and the maximization of experience. Minimize the overall image calculation overhead of the system without sacrificing the clarity of key visual areas.

[0068] Based on the benchmark regional image processing scheme, combined with the three-dimensional space region, perform the display control of the virtual exhibition hall.

[0069] In some embodiments, based on the benchmark regional image processing scheme, combined with the three-dimensional space region, perform the display control of the virtual exhibition hall, including: Perform object recognition on the central space region and the neighborhood space region respectively to determine the resource call target and the resource pre-storage target; analyze the benchmark regional image processing scheme, and extract the first call accuracy corresponding to the resource call target and the second call accuracy corresponding to the resource pre-storage target; according to the resource call target and the first call accuracy, address and read the corresponding instant digital display resources for image rendering and display control; synchronously, according to the resource pre-storage target and the second call accuracy, read the corresponding predicted digital display resources and store them in the cache medium.

[0070] Specifically, object recognition is first performed on the central space region and the neighborhood space region respectively. Based on the regional perspective and interaction frequency analysis, the immediate interaction objects to be rendered currently (defined as resource call targets) are automatically recognized, as well as the candidate objects that will enter the user's attention range (defined as resource pre-storage targets). In other words, the resource call target is the region or object that needs to be immediately rendered and displayed, including exhibits, exhibition stands, lighting, walls, etc. in the virtual exhibition hall. The resource pre-storage target refers to the region or object that may need to be rendered in the future and is pre-loaded to improve the display efficiency. For example, in a virtual exhibition hall, the main exhibits in the central space region are recognized as resource call targets, while the secondary exhibits or backgrounds in the neighborhood space region are resource pre-storage targets. Through the above recognition process, it is possible to ensure that high-priority content is processed first, and at the same time, it helps to pre-load the content that may be needed.

[0071] Specifically, analyze the rendering sub-strategies matched by each region in the benchmark region image processing scheme, and extract the first call accuracy corresponding to the resource call target and the second call accuracy corresponding to the resource pre-storage target from it. Among them, the above call accuracy is matched with parameters such as image resolution, texture level, model detail level, anti-aliasing intensity, etc. in the rendering sub-strategy.

[0072] Specifically, the immediate digital display resources refer to the digital resources corresponding to the resource call targets that need to be immediately used currently and need to meet the first call accuracy, such as high-precision models, videos, etc.; the predicted digital display resources refer to the digital resources that may be needed according to the user's behavior and are usually pre-loaded into the cache and meet the second call accuracy.

[0073] Furthermore, call the graphics rendering pipeline (such as based on graphics interfaces such as OpenGL / DirectX / Vulkan), perform real-time image rendering according to the obtained immediate digital display resources, and execute dynamic display control to ensure that the core content that the user is currently concerned about is displayed in the highest quality. At the same time, according to the resource pre-storage target and the second call accuracy, read the resources in the neighborhood space region (i.e., the predicted digital display resources) from the storage device and pre-load them into the local cache medium (such as GPU video memory, system cache, VRAM buffer pool, etc.) to reduce the latency when accessing these resources in the future and improve the display efficiency.

[0074] Through the above process, a dynamic display resource scheduling mechanism driven by the image processing strategy in the three-dimensional virtual exhibition hall can be realized, thereby improving the overall resource utilization efficiency, image rendering response speed, and scalability of the display system while ensuring the user's high-immersion experience.

[0075] In some implementation manners, based on the benchmark region image processing scheme, combined with the three-dimensional space region, the display control of the virtual exhibition hall is performed. After that, it further includes: When multiple users are detected to be at the same spatial position coordinates, extract the display area images corresponding to each user and perform comparative analysis; based on the comparison results of the display area images, identify invariant feature information, and construct a shared image rendering context according to the invariant feature information, where the shared image rendering context is used for fast image reconstruction and image reuse among multiple users.

[0076] Specifically, the display area image refers to the image content seen by the user from the current perspective in the virtual exhibition hall; the invariant feature information refers to the visual features that remain unchanged in the display area images of multiple users. The shared image rendering context is a shared rendering environment used to store and manage image resources and rendering parameters that can be reused among multiple users.

[0077] Exemplarily, the invariant feature information includes motion direction vectors, material texture features, and lighting parameter features, etc. Among them, the motion direction vector refers to the motion direction and speed of the object in the image, which is used for the analysis of dynamic scenes; the material texture feature refers to the surface material and texture information of the object in the image, such as color, roughness, etc.; the lighting parameter feature refers to the lighting conditions in the image, such as the light source direction, intensity, etc.

[0078] Optionally, if multiple users are detected to be at the same or approximate spatial position coordinates through head-mounted device pose data or behavior prediction trajectories, trigger the shared rendering logic, extract the display area image data from each user's current perspective, including image frame buffer information, rendering parameter sets, etc.; then, compare the display area images, identify the similar segments or stable presentation content (such as fixed exhibits and their materials, textures, shadows, etc.) therein, perform image reconstruction difference analysis, and the obtained invariant feature information indicates the same aspects in the displays of different users and can be reused.

[0079] Optionally, construct a shared image rendering context based on the above invariant feature information for image reuse by multiple users in the shared area, where the shared image rendering context may include a unified scene graph structure, a pre-rendered frame buffer, a common material buffer, and image reconstruction parameter configuration, etc. For example, store the invariant material texture features and lighting parameter features in the shared rendering context so that they can be shared by multiple users, thereby reducing duplicate calculations and resource waste.

[0080] The above process can significantly improve the flexibility and rationality of resource allocation by identifying invariant feature information and constructing a shared image rendering context. Among them, the comparative analysis and the identification of invariant feature information ensure that the sharing possibilities among multiple users are fully utilized, reducing duplicate calculations; the construction of the shared image rendering context enables resources to be efficiently reused among multiple users, avoiding resource waste; the fast image reconstruction and reuse significantly improve the display efficiency, especially in multi-user scenarios, enhancing the overall performance of the system and the user experience, and realizing efficient and personalized display control in the virtual exhibition hall.

[0081] In summary, a regional image processing method for virtual exhibition hall display control provided by the present invention has the following technical effects: By constructing a two-layer behavior model based on user behavior, the model includes a common behavior heat map and a personalized behavior portrait; on this basis, a three-dimensional space area from the perspective of the current user is established according to the two-layer behavior model, and the space area is divided into a central space area and a neighborhood space area; subsequently, a heat response strategy is constructed, and the three-dimensional space area is hierarchically processed and decided in combination with the common behavior heat map, and the output result is a reference area image processing scheme; finally, based on this image processing scheme, the display control of the virtual exhibition hall is implemented in the three-dimensional space area, so as to achieve the technical effects of improving the flexibility and rationality of resource allocation, and improving the display efficiency and user experience.

[0082] Embodiment 2, as Figure 2 is a schematic structural diagram of a regional image processing system for virtual exhibition hall display control according to the present invention. For example, Figure 1 In the flow schematic diagram of a regional image processing method for virtual exhibition hall display control according to the present invention can be implemented by such as Figure 2 shown structure.

[0083] Based on the same concept as a regional image processing method for virtual exhibition hall display control in the above embodiment, the present invention also provides a regional image processing system for virtual exhibition hall display control, including: A two-layer behavior model construction module 11, configured to construct a two-layer behavior model based on user behavior, wherein the two-layer behavior model includes a common behavior heat map and a personalized behavior portrait.

[0084] An area establishment module 12, configured to establish a three-dimensional space area from the perspective of the current user according to the two-layer behavior model, wherein the three-dimensional space area is divided into a central space area and a neighborhood space area.

[0085] An image processing decision module 13, configured to construct a heat response strategy, and perform hierarchical processing and decision on the three-dimensional space area in combination with the common behavior heat map, and output a decision result as a reference area image processing scheme.

[0086] A display control module 14 is configured to perform display control of a virtual exhibition hall based on the image processing scheme of the reference area in combination with the three-dimensional space area.

[0087] In some embodiments, the dual-layer behavior model construction module 11 includes: A user interaction behavior data acquisition unit is configured to acquire the line-of-sight trajectories and operation behavior data of multiple users and analyze to obtain user interaction behavior data, where the user interaction behavior data includes line-of-sight trajectory information and operation action information.

[0088] A common behavior heat map generation unit is configured to statistically analyze the line-of-sight trajectory distribution of the virtual exhibition hall area based on the line-of-sight trajectory information and correspondingly generate the common behavior heat map.

[0089] A personalized behavior portrait construction unit is configured to construct the personalized behavior portrait based on the operation action information in combination with feature engineering.

[0090] A dual-layer behavior model output unit is configured to output the common behavior heat map and the personalized behavior portrait as the dual-layer behavior model.

[0091] In some implementation manners, the personalized behavior portrait construction unit in the dual-layer behavior model construction module 11 includes: An operation action information clustering unit is configured to perform clustering division on the operation action information according to the unique identification identity of the user.

[0092] A key behavior feature data extraction unit is configured to perform feature extraction on the clustering division result to obtain multiple groups of key behavior feature data.

[0093] A personalized behavior portrait vectorization unit is configured to calculate the perspective turning rate, action randomness entropy value, and line-of-sight residence time distribution corresponding to multiple groups of the key behavior feature data respectively, and vectorize the calculation results as the personalized behavior portrait, where the personalized behavior portrait is associated with a unique identification identity.

[0094] In some embodiments, the area establishment module 12 includes: A central space area determination unit is configured to cut off the areas with heat values lower than the omitted heat limit based on the common behavior heat map and the current user's perspective to determine the central space area.

[0095] A neighborhood space area determination unit is configured to perform neighborhood growth based on the central space area in combination with the personalized behavior portrait to determine the corresponding neighborhood space area, where the neighborhood growth radius is calculated and determined based on the personalized behavior portrait.

[0096] A three-dimensional spatial region combination smoothing unit for combining and smoothing the central spatial region and the neighborhood spatial region.

[0097] In some embodiments, the image processing decision module 13 includes: A thermal response strategy construction unit for constructing the thermal response strategy according to the processing performance parameters and hardware function information of the target scene, wherein the thermal response strategy includes a plurality of rendering sub-strategies, and each of the rendering sub-strategies has a differential configuration in terms of image processing accuracy and computing resource scheduling.

[0098] A three-dimensional spatial region interest level division unit for using preset region semantic information and combining with the common behavior heat map to divide and mark the interest level of the three-dimensional spatial region.

[0099] A reference region image processing scheme generation unit for matching the image rendering sub-strategy corresponding to the corresponding interest level according to the mapping relationship in the thermal response strategy, and outputting the matching result as the reference region image processing scheme.

[0100] In some embodiments, the display control module 14 includes: An object recognition and resource target determination unit for respectively performing object recognition on the central spatial region and the neighborhood spatial region to determine the resource call target and the resource pre-storage target.

[0101] A reference region image processing scheme analysis unit for analyzing the reference region image processing scheme and extracting the first call accuracy corresponding to the resource call target and the second call accuracy corresponding to the resource pre-storage target.

[0102] An instant digital display resource call unit for addressing and reading the corresponding instant digital display resource according to the resource call target and the first call accuracy for image rendering and display control.

[0103] A predicted digital display resource pre-storage unit for synchronously reading the corresponding predicted digital display resource according to the resource pre-storage target and the second call accuracy and storing it in the cache medium.

[0104] In some embodiments, the system further includes: A multi-user image comparison and analysis unit for extracting the display area images corresponding to each user and performing comparison and analysis when it is detected that multiple users are located at the same spatial position coordinates.

[0105] A sharing and image reuse unit, configured to identify invariant feature information based on a comparison result of the display area image, and construct a shared image rendering context according to the invariant feature information, where the shared image rendering context is used for fast image reconstruction and image reuse among multiple users.

[0106] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the foregoing Embodiment 1 are equally applicable to the area image processing system for virtual exhibition hall display control described in Embodiment 2. For the sake of brevity of the specification, no further elaboration will be made here.

[0107] It should be understood that the disclosed embodiments of the present invention and the above descriptions enable those skilled in the art to implement the present invention using the present invention. At the same time, the present invention is not limited to the above-mentioned part of the embodiments. It should be understood that those of ordinary skill in the art can still modify the technical solutions recorded in the foregoing embodiments, or perform equivalent replacements on some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included in the protection scope of the present invention.

Claims

1. An area image processing method for virtual exhibition hall display control, characterized in that, The method includes: Constructing a two - layer behavior model based on user behavior, where the two - layer behavior model includes a common behavior heat map and a personalized behavior portrait; Establishing a three - dimensional space area from the perspective of the current user according to the two - layer behavior model, where the three - dimensional space area is divided into a central space area and a neighborhood space area; Constructing a heat response strategy, and combining the common behavior heat map to perform hierarchical processing decision - making on the three - dimensional space area, and outputting the decision result as an image processing scheme for the reference area; Based on the image processing scheme for the reference area, combining the three - dimensional space area to perform display control of the virtual exhibition hall.

2. The regional image processing method for virtual exhibition hall display control according to claim 1, characterized in that Constructing a two - layer behavior model based on user behavior, where the two - layer behavior model includes a common behavior heat map and a personalized behavior portrait, including: Obtaining the line - of - sight trajectories and operation behavior data of multiple users and analyzing to obtain user interaction behavior data, where the user interaction behavior data includes line - of - sight trajectory information and operation action information; Based on the line - of - sight trajectory information, statistically analyzing to obtain the line - of - sight trajectory distribution of the virtual exhibition hall area, and correspondingly generating the common behavior heat map; Based on the operation action information, constructing the personalized behavior portrait by combining feature engineering; Outputting the common behavior heat map and the personalized behavior portrait as the two - layer behavior model.

3. The regional image processing method for virtual exhibition hall display control according to claim 2, wherein, Based on the operation action information, constructing the personalized behavior portrait by combining feature engineering, including: According to the unique identification identity of the user, clustering and dividing the operation action information; Performing feature extraction on the clustering and division results to obtain multiple groups of key behavior feature data; Calculating the perspective turning rate, action randomness entropy value, and line - of - sight residence time distribution corresponding to multiple groups of the key behavior feature data respectively, and vectorizing the calculation results as the personalized behavior portrait, where the personalized behavior portrait is associated with and marked with a unique identification identity.

4. The regional image processing method for virtual exhibition hall display control according to claim 3, characterized in that, According to the two - layer behavior model, establishing a three - dimensional space area from the perspective of the current user, where the three - dimensional space area is divided into a central space area and a neighborhood space area, including: Based on the common behavior heat map and the current user's perspective, cropping the area where the heat value is lower than the omitted heat limit to determine the central space area; Based on the central space area, performing neighborhood growth by combining the personalized behavior portrait to determine the corresponding neighborhood space area, where the neighborhood growth radius is calculated and determined based on the personalized behavior portrait; Combining and smoothing the central space area and the neighborhood space area.

5. A regional image processing method for virtual exhibition hall display control according to claim 4, characterized in that, Constructing a heat response strategy, and combining the common behavior heat map to perform hierarchical processing decision - making on the three - dimensional space area, and outputting the decision result as an image processing scheme for the reference area, including: Constructing the heat response strategy according to the processing performance parameters and hardware function information of the target scene, where the heat response strategy includes multiple rendering sub - strategies, and each rendering sub - strategy has a differential configuration in terms of image processing accuracy and computing resource scheduling; Using the preset area semantic information, combining the common behavior heat map, to perform interest level division and marking on the three - dimensional space area; Match the image rendering sub-strategy corresponding to the interest level according to the mapping relationship in the thermal response strategy, and output the matching result as the image processing solution for the reference area.

6. The regional image processing method for virtual exhibition hall display control according to claim 5, characterized in that, Based on the image processing solution for the reference area, combine with the three-dimensional space area to perform display control of the virtual exhibition hall, including: Perform object recognition on the central space area and the neighborhood space area respectively to determine the resource call target and the resource pre-storage target; Analyze the image processing solution for the reference area, and extract the first call accuracy corresponding to the resource call target and the second call accuracy corresponding to the resource pre-storage target; According to the resource call target and the first call accuracy, address and read the corresponding instant digital display resource for image rendering and display control; Synchronously, according to the resource pre-storage target and the second call accuracy, read the corresponding predicted digital display resource and store it in the cache medium.

7. A regional image processing method for virtual exhibition hall display control according to claim 1, characterized in that, Based on the image processing solution for the reference area, combine with the three-dimensional space area to perform display control of the virtual exhibition hall. After that, it further includes: If it is detected that multiple users are located at the same spatial position coordinates, extract the display area images corresponding to each user and perform comparative analysis; Based on the comparison result of the display area images, identify the invariant feature information, and construct a shared image rendering context according to the invariant feature information, where the shared image rendering context is used for fast image reconstruction and image reuse among multiple users.

8. An area image processing system for virtual exhibition hall display control, characterized in that, A regional image processing method for virtual exhibition hall display control for implementing any one of claims 1-7, including: A double-layer behavior model construction module for constructing a double-layer behavior model based on user behavior, where the double-layer behavior model includes a common behavior heat map and a personalized behavior portrait; A region establishment module for establishing a three-dimensional space area from the perspective of the current user according to the double-layer behavior model, where the three-dimensional space area is divided into a central space area and a neighborhood space area; An image processing decision module for constructing a thermal response strategy and performing hierarchical processing decision on the three-dimensional space area in combination with the common behavior heat map, and outputting the decision result as the image processing solution for the reference area; A display control module for performing display control of the virtual exhibition hall in combination with the three-dimensional space area based on the image processing solution for the reference area.

Images