Immersive user experience method and system based on virtual reality technology

Abstract

The invention discloses an immersive user experience method and system based on a virtual reality technology, and relates to the technical field of virtual reality, the system is composed of a plurality of functional modules, and the system comprises a map construction module which collects user behavior characteristics through multiple sensors and constructs a group cognition map model; the user behavior characteristics comprise interaction preferences, attention distribution and role behavior modes; the map layer generation module is used for calculating a focusing fixation point of a user node to a virtual scene based on the association strength of the user node and the virtual scene in the group cognition map model, performing labeling initialization and optimizing labeling attributes after effective focusing is judged, and generating a three-dimensional labeling map layer through multi-person collaborative editing; the interactive collaboration module introduces a visual view angle to output a rendering shared view angle on the basis of overlapping the focusing fixation point by expanding multi-dimensional feature fusion; and through the group cognition map model, obtaining the attention distribution weight of the user in the watching area, and outputting the position of a warning icon.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of virtual reality, in particular to an immersive user experience method and system based on virtual reality technology. BACKGROUND

[0002] With the improvement of hardware performance and the reduction of cost, virtual reality technology has developed rapidly and has shown its unique role in many fields; VR provides the ability to simulate a real environment, allowing users to practice learning in a safe environment, through VR devices, players can immerse themselves in the game world and have an unprecedented interactive experience. Immersive user experience based on virtual reality technology has changed from a scientific fantasy in the early days to an actual technology widely used in various fields today, not only greatly enriching people's daily life, but also bringing revolutionary changes to various industries.

[0003] Traditional shared perspectives are mostly "broadcasting replication", which only simply transmits picture data without considering attention transmission; the encoding end lacks precise perspective parameter definition, and the compression algorithm is often rough, resulting in loss of details during virtual scene transmission, unreasonable bandwidth occupation, high perspective switching delay, and destruction of immersion; the past warning icons use "preset coordinate broadcasting", and the position calculation is not associated with scene dynamic information; only the midpoint of the user and the target is simply taken, without considering factors such as scene threats and group roles, so it cannot accurately reflect the actual threat direction; in a virtual combat scene, scene threats such as enemy distribution and friendly force position are not considered, and the warning icon is easy to mislead team action. SUMMARY

[0004] (I) Technical problems solved In view of the deficiencies of the prior art, the present application provides an immersive user experience method and system based on virtual reality technology, so as to solve the problem that in the prior art, shared perspectives are mostly "broadcasting replication", which only simply transmits picture data without considering attention transmission.

[0005] (II) Technical solutions To achieve the above purpose, the present application is implemented by the following technical solutions: The immersive user experience system based on virtual reality technology comprises: A graph construction module constructs a group cognitive graph model by collecting user behavior characteristics through multiple sensors; the user behavior characteristics include interaction preferences, attention distribution, and role behavior patterns; A layer generation module calculates the focus gaze point of the user node on the virtual scene based on the association strength of the user node in the group cognitive graph model and the virtual scene, labels and initializes the effective focus after judgment, optimizes the labeling attributes, and generates a three-dimensional labeling layer through multi-person collaborative editing; The interaction coordination module introduces a visual perspective output rendering shared perspective by expanding multi-dimensional feature fusion and introducing a visual perspective output rendering shared perspective on the basis of superimposed focused gaze points; the group cognitive graph model is used to obtain the attention distribution weight of the user in the gaze area and output the warning icon position.

[0006] Further, the process of collecting user behavior characteristics by the multi-sensor is: By multi-sensor cooperation, the interaction data of multiple users in a virtual scene is collected, and the following is performed respectively: obtaining the preference degree in the interaction preference ; constructing an attention distribution vector ; extracting the decision proportion in the role behavior mode .

[0007] Further, the process of constructing the group cognitive graph model is: Based on the user behavior characteristics, a group cognitive graph G is constructed, and the association strength between users and users and the association strength between users and virtual objects are obtained respectively.

[0008] Further, the focused gaze point is calculated: The association strength of the user node and the virtual scene node in the group cognitive graph G; when the user is in the virtual scene node , the group focus state is detected to trigger the generation of the multi-person three-dimensional labeling layer; the association strength of the user node and the virtual scene node in the group cognitive graph is used to calculate the focused gaze point F of the group to the virtual scene node: , Where N is the number of participating users, D is the virtual scene set, M is the total number of objects in the virtual scene; i is the user; j is the index identifier of the object in the virtual scene.

[0009] Further, the process of initializing labeling after determining effective focus is: When F exceeds the threshold value , it is determined that the current state is effective focus, and the labeling initialization is performed; When F is less than or equal to the threshold value , it is determined that the group attention is in a scattered state, and the current adjustment mechanism is maintained; After determining effective focus, a three-dimensional bounding box is constructed for the three-dimensional space, key semantic elements are extracted based on semantic analysis , and initial labeling positions are allocated in the three-dimensional space , satisfying the three-dimensional space bounding box constraint and the group viewing requirement.

[0010] Further, the process of optimizing the labeling attribute is: Optimizing the labeling attribute through group cognitive atlas , the formula is: , Among them, is the base color; is the adjustment coefficient; is the user i's attention time in the region; K is the total number of regions divided in the virtual scene; k' is the semantic element index; is the semantic region in the virtual space; is the user i's attention time in the virtual scene region.

[0011] Further, the process of generating a three-dimensional labeling layer through multi-person collaborative editing is: When multi-person collaborative editing, record the user operation sequence , calculate the operation priority Determine the operation execution order according to the priority, and solve the editing conflict.

[0012] Further, the process of introducing visual perspective output rendering shared perspective based on superimposed focus gaze points is: By defining the user i's visual perspective as a camera parameter set , using a deep learning image compression algorithm to compress the perspective image into a low-bandwidth data stream, the receiving end decodes to obtain the shared perspective image , superimpose the original user gaze point mark , the gaze point position is determined by eye movement data , and rendered into a shared perspective with gaze marks.

[0013] Further, the process of obtaining the attention distribution weight of the user in the gaze region and outputting the warning icon position is: Through the acquisition device, obtain the attention related features , use a support vector machine SVM classifier to obtain a classification result y, after judging the detection event, according to the user space position and the enemy position , calculate the warning icon position , the formula is: , Among them, is the position fusion coefficient.

[0014] The immersive user experience method based on virtual reality technology comprises the following steps: Step one: collect user behavior characteristics through multiple sensors to build a group cognitive graph model; user behavior characteristics include interaction preferences, attention distribution, and role behavior patterns; Step two: based on the association strength between user nodes and virtual scenes in the group cognitive graph model, calculate the focus gaze point of the user node on the virtual scene, and after judging as effective focus, initialize and optimize the labeling attributes, generate a three-dimensional labeling layer through multi-person collaborative editing; Step three: through the expansion of multi-dimensional feature fusion, introduce visual perspective output rendering shared perspective based on the superposition of focus gaze points; through the group cognitive graph model, obtain the attention distribution weight of the user in the gaze area and output the warning icon position.

[0015] (Three) beneficial effects The present application provides an immersive user experience method and system based on virtual reality technology, which has the following beneficial effects: (1) In the present application, the focus gaze point is calculated by the association strength between the user node and the virtual scene node in the group cognitive graph, which comprehensively considers the complex association relationship between multiple users and the virtual scene, and reflects the attention focus of the group in the virtual scene from the group cognitive level. After judging as effective focus, initialization is performed, and the generation of the label is based on the attention distribution of the group; through the group cognitive graph, the attention degree of the user to different parts of the virtual scene and other information are obtained, so that the label can accurately point to the key area of the group's attention, rather than random or uniform mode of labeling, which improves the pertinence and effectiveness of the label; (2) The present application initializes the label after judging as effective focus, and the generation of the label is based on the attention distribution of the group. Through the group cognitive graph, the attention degree of the user to different parts of the virtual scene and other information are obtained, so that the label can accurately point to the key area of the group's attention, rather than random or uniform mode of labeling, which improves the pertinence and effectiveness of the label. Through multi-person collaborative editing, a three-dimensional labeling layer is generated, breaking the traditional single-person labeling or simple multi-person sequential labeling mode. In the virtual scene, multiple users can participate in labeling at the same time, improving the efficiency and diversity of labeling. This deep integration enables the system to understand the collaborative intention of the user from the group cognitive level, not only realizing the generation of the labeling layer, but also providing more intelligent support for group collaboration. BRIEF DESCRIPTION OF DRAWINGS

[0016] Figure 1 The system flowchart of the present application is shown in the figure; Figure 2 The overall method schematic diagram of the present application is shown in the figure. DETAILED DESCRIPTION

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0018] Example 1: Please refer to Figure 1 This embodiment provides an immersive user experience method based on virtual reality technology. The early warning system includes: The graph construction module collects user behavior characteristics through multiple sensors to build a group cognitive graph model; user behavior characteristics include interaction preferences, attention distribution, and role behavior patterns; The specific process of collecting user behavior characteristics is as follows: By coordinating multiple sensors, interaction data of multiple users in a virtual scene can be collected; Millimeter-wave radar array: Simultaneous positioning of multiple user spatial coordinates to achieve millimeter-level positioning and capture limb movement trajectories; Eye tracker: Collects the coordinates of the user's gaze point; Electromyography and electroencephalography (EEG) sensors: record EEG features related to force exertion and attention to assist in extracting character behavior patterns; Motion capture equipment: tracks sequences of actions such as gestures and body postures, and counts the frequency of operations on virtual objects; In multi-user virtual interaction scenarios, it is necessary to record each user's unique interaction preferences. Attention distribution and role behavior patterns .

[0019] Interaction Preference Quantification: By defining user interactions with virtual objects Operating frequency User i on virtual objects Preference The calculation formula is as follows: , Where M is the total number of objects in the virtual scene, and j is the index of an object in the virtual scene, used to traverse from the 1st to the Mth object in calculating the probability. In this process, the relevant data of all objects in the virtual scene are comprehensively considered by summing j from 1 to M. By statistically analyzing the frequency of user operations on different virtual objects and normalizing the data, a preference score is obtained, reflecting the user's interest in virtual content.

[0020] Attention distribution modeling: Utilizing eye-tracking and EEG signal analysis techniques, this method obtains the proportion of user attention in different areas of a virtual scene. It also calculates the virtual region the user is focusing on. Time percentage Construct attention distribution vector K represents the total number of regions in the virtual scene. This vector visually represents the user's focus of attention in the virtual space, providing a basis for subsequent scene interaction optimization.

[0021] Role behavior pattern extraction: For collaborative scenarios, analyze user action sequences and decision-making participation levels. Taking the dominant role as an example, calculate the number of times it initiates decisions. Number of total decisions ratio ,Right now The formula for determining the decision-making percentage is: , By using behavioral clustering algorithms, we can discover the role characteristics of different users in collaborative tasks, such as dominant and auxiliary roles, and provide behavioral pattern references for group interaction and collaboration.

[0022] By outputting user behavior features including user interaction preference vectors, attention distribution vectors, and role behavior pattern features, the system serves as the basic input for constructing a group cognitive graph, accurately depicting the individual behavioral characteristics of users in the virtual environment and supporting subsequent group interaction analysis.

[0023] Constructing a group cognitive graph model: Based on user behavior characteristics, a group cognitive graph G=(V,E) is constructed to realize the visualization and quantitative analysis of the relationship between users and virtual scene elements.

[0024] Node definition: Node V contains user nodes. and virtual objects User node Storage interaction preferences Attention distribution Role Behavior Patterns Virtual objects Record its own attributes, such as type, physical characteristics and relationship with users, to form a "digital profile" of virtual scene elements; user nodes are sub-nodes of multiple user groups.

[0025] Edge weight calculation (association strength): User-user association strength: different user nodes and The strength of the correlation between Based on the calculation of interaction collaboration degree, the formula is: , wherein, is the number of times of jointly operating virtual elements, is the total number of operations of user i, is a weight coefficient, a is used to measure the importance of "interaction synergy" in the calculation of user association strength, reflecting the influence weight of actual interaction behavior (jointly operating virtual elements) between users on the association relationship, and β is used to measure the importance of "attention focus consistency", reflecting the influence weight of the behavior of focusing on the same area of the virtual scene between users on the association relationship, satisfying are positive numbers, and ; is the area proportion of the common focus area of different user attentions. This formula comprehensively considers the synergy of user interaction behavior and the consistency of attention focus, quantifies the closeness of the association between users; the user node and the user refer to the same.

[0026] User node-virtual object association strength: the association strength between the user node and the virtual object , the formula is: , wherein, is a weight coefficient, measuring the weight proportion of user subjective preference (long-term interest reflected by operation frequency) and real-time attention focus (short-term attention reflected by area gaze time) in the calculation of association strength, and ; is the gaze time of user i on the area where the element of j' is located, is the element set in the area. By integrating preference and attention distribution, the attention degree of the user to the virtual element is accurately reflected.

[0027] Through the construction of nodes and edges, the group cognitive graph clearly presents the complex association between multiple users and virtual scene elements, and outputs a graph structure containing node attributes and edge weight relationships, providing topological relationship support for group interaction analysis and scene intelligent response, and realizing the deep association mapping of group behavior and scene elements in the virtual environment.

[0028] The layer generation module calculates the focus gaze point based on the association strength between the user nodes and the virtual scene nodes in the group cognitive graph model, initializes the labeling after judging as effective focus, and generates a three-dimensional labeling layer through multi-person collaborative editing; The focus gaze point is calculated based on the association strength between the user nodes and the virtual scene nodes in the group cognitive graph G: When the user is in the virtual scene node ​At that time, by accurately detecting the group's focus state, a multi-person 3D annotation layer is generated; the correlation strength between user nodes and virtual scene nodes in the group's cognitive graph is used. Let i be the user, j be the index of an object in the virtual scene, and calculate the focal point F of the group's gaze on the virtual scene node: , Where N is the number of participating users, D is the set of virtual scenes, and M is the total number of objects in the virtual scenes.

[0029] The judgment process is as follows: When F exceeds the threshold (as set) When the threshold is 0.6 (meaning 60% of the group's attention is focused on the same document), the system will determine the current state as effective focus. The core of this determination mechanism is to quantify the concentration of group attention. When the threshold is reached, it means that most users' focus is highly unified. At this time, subsequent processes will be triggered, such as starting the deep analysis module to explore user interests, or pushing relevant extended resources based on the focused content, so as to optimize the interaction logic and content distribution strategy of the virtual reality immersive user experience system. When F is less than or equal to the threshold When the system determines that the group's attention is scattered, it will activate a dynamic content adjustment mechanism. On the one hand, it will use diverse content recommendation algorithms to push differentiated resources to stimulate user interest and prevent user fatigue due to monotonous content. On the other hand, it will use cluster analysis technology to group users with similar behavioral patterns and push customized content based on the characteristics of different groups. At the same time, it will collect user feedback data to optimize the subsequent content distribution model and gradually guide the group's attention to a state of effective focus.

[0030] Based on the correlation of the group's cognitive graph, the algorithm quantifies the focus point of the group's attention, accurately identifies the triggering conditions for collaborative editing, and outputs the focus detection result (whether it is triggered), providing a start signal for the generation of 3D annotation layers.

[0031] After determining that the focus is valid, perform annotation initialization: After focus is triggered, a 3D bounding box is constructed for the 3D space, using the following formula: , Extracting key semantic elements based on semantic analysis Assign initial annotation positions in three-dimensional space It satisfies the bounding box constraints in three-dimensional space and the requirements for group viewing.

[0032] The formula for optimizing labeled attributes (color, size) using a group cognitive graph is: , wherein, is the base color, is the adjustment coefficient, used to control the magnitude of the adjustment of the annotation color according to the user gaze time, and the value determines the sensitivity of the color change with the attention level; is the semantic region in the virtual space; is the gaze time of user i on the virtual scene region k'; is the gaze time of user i on the semantic element k; is the gaze time of user i on the semantic element k; represents the sum of the gaze time of all users on all semantic elements, and the numerator is normalized by the denominator, so as to dynamically adjust the annotation color based on the group gaze data. The annotation color of the semantic element with high attention level is brighter, which is consistent with the group attention characteristics. The output of this link is an initial annotation set containing three-dimensional position and attributes, which constructs a document annotation basic framework in the virtual space and provides a visual interactive object for collaborative editing.

[0033] When multiple people collaborate to edit, operation conflicts need to be solved to ensure the orderliness of editing.

[0034] Operation priority calculation: record the user operation sequence (moving, modifying, etc.), calculate the operation priority , the formula is: , wherein, is the user role weight (based on role behavior pattern , the dominant weight is high), is the operation time sequence weight, the earlier the operation, the higher the weight, . The role and time factors are fused to quantify the operation priority.

[0035] Conflict resolution and execution: determine the operation execution order according to the priority to solve the conflict. If a user simultaneously modifies the same annotation, the operation with high priority is executed first to ensure the consistency of collaborative editing. The output is the processed annotation editing result, which updates the annotation state in the virtual space in real time, realizes efficient collaborative editing of documents by the group, and makes multiple users complete annotation collaboration smoothly in the virtual environment, improves the efficiency of information interaction and knowledge co-creation, dynamically associates with the group cognitive graph, and optimizes the editing process according to the group behavior characteristics.

[0036] Interactive collaboration module: render the shared view by introducing the visual view on the basis of superimposing the focused gaze point; obtain the attention distribution weight of the user in the gaze area through the group cognitive graph model, and output the shared view with enhanced markers; Render the shared view by introducing the visual view on the basis of superimposing the focused gaze point: Cross-user sharing of visual perspective enhances the accuracy of group collaboration information transmission.

[0037] Perspective encoding transmission: define the visual perspective of user i as a set of camera parameters , the formula is: , Among them, is the projection matrix, is the rotation matrix, using deep learning image compression algorithm, the formula is: , Among them, Encoder is the encoding network, which compresses the perspective image into a low-bandwidth data stream for efficient transmission.

[0038] Shared perspective rendering: the receiver decodes the shared perspective image , superimposed with the original user's gaze point marker . The gaze point position is determined by eye movement data , rendered as a specific marker such as a highlighted circle, the formula is: , Among them, is the shared perspective image with gaze marker; is the output visual focus marker; let the receiver clearly identify the original user's gaze focus, improve the accuracy of visual information transmission in group collaboration. The output is the shared perspective content with gaze marker, realizing visual perspective sharing and focus transmission between users in virtual space, optimizing information presentation of perspective sharing based on user attention features of group cognitive map, and strengthening the sense of immersion in group collaboration.

[0039] Attention distribution weight: In the virtual team combat scene, realize the cross-user sharing of user attention signal, and improve the efficiency of group collaborative combat.

[0040] Attention event detection: obtain attention-related features through acquisition equipment , using support vector machine SVM classifier, the formula is: , Among them, y is the classification result, which judges whether the attention concentration event of "finding hidden enemy" is detected, and accurately captures the key collaboration signal.

[0041] Warning icon generation and broadcast: after detecting the event, according to the user space position and the enemy position , calculate the warning icon position , the formula is: , Among them, The position fusion coefficient is, for example, , the user and the enemy position influence are balanced. A warning icon is generated and broadcast to the team interface, realizing the sharing of attention signals across users. The output is a warning icon interaction content shared by the team, which converts individual attention into group cooperation signals, optimizes the transmission of warning signals based on user roles and position information in the group cognitive map, enhances the group synergy of the virtual combat scene, and enables users to efficiently cooperate in an immersive experience.

[0042] Embodiment 2: Please refer to Figure 2 Based on Embodiment 1, this embodiment also provides an immersive user experience method based on virtual reality technology, comprising the following specific steps: Step 1: Collect user behavior characteristics through multiple sensors to build a group cognitive map model; user behavior characteristics include interaction preferences, attention distribution, and role behavior patterns; Step 2: Based on the association strength between user nodes in the group cognitive map model and the virtual scene, calculate the focused gaze point of the user node to the virtual scene, and after judging as an effective focus, initialize and optimize the labeling attributes, and generate a three-dimensional labeling layer through multi-person collaborative editing; Step 3: Through the expansion of multi-dimensional feature fusion, introduce visual perspective output to render a shared perspective based on the superimposed focused gaze point; through the group cognitive map model, obtain the attention distribution weight of the user in the gaze area to output the warning icon position.

[0043] The above embodiments can be realized in whole or in part by software, hardware, firmware or any combination thereof. When implemented by software, the above embodiments can be realized in whole or in part in the form of a computer program product. Those skilled in the art can realize that the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein can be realized by electronic hardware or a combination of computer software and electronic hardware. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical solutions.

[0044] The units described as separate components can or can not be physically separated, and the components shown as units can or can not be physical units, which can be located in one place or distributed on multiple network units. Part or all of the units can be selected to achieve the purpose of the embodiment according to actual needs.

[0045] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto, and any skilled person in the art can easily think of changes or replacements within the technical scope disclosed in the present application, which should be covered within the protection scope of the present application.

Claims

1. An immersive user experience system based on virtual reality technology, characterized in that: The system comprises: A graph construction module, which collects user behavior characteristics through a plurality of sensors and constructs a group cognitive graph model; the user behavior characteristics include interaction preferences, attention distribution, and role behavior patterns; A layer generation module, which calculates a focused gaze point of a user node on a virtual scene based on an association strength of the user node on the virtual scene in the group cognitive graph model, labels and initializes the focused gaze point after determining that the focused gaze point is valid, optimizes label attributes, and generates a three-dimensional label layer through multi-person collaborative editing; An interaction collaboration module, which, through extended multi-dimensional feature fusion, introduces a visual perspective output to render a shared perspective based on the focused gaze point, and, through the group cognitive graph model, acquires an attention distribution weight of a user in a gaze area to output a warning icon position.

2. The virtual reality technology based immersive user experience system, as claimed in claim 1, wherein, The process of collecting user behavior characteristics through the plurality of sensors is as follows: Through the cooperation of multiple sensors, the interaction data of multiple users in the virtual scene is collected, and the following is performed respectively: obtaining the preference degree in the interaction preference ; constructing an attention distribution vector ; extracting the decision proportion in the role behavior mode .

3. The virtual reality technology based immersive user experience system, as claimed in claim 2, wherein, The process of constructing the group cognitive graph model is as follows: Based on the user behavior characteristics, a group cognitive graph G is constructed, and the association strength between users and users and the association strength between users and virtual objects are obtained respectively .​ 4. The virtual reality technology based immersive user experience system, as claimed in claim 1, wherein: The process of calculating the focused gaze point is as follows: User nodes in the group cognitive graph G With virtual scene nodes The strength of the association; users in virtual scene nodes At that time, by detecting the group's focus state, a multi-person 3D annotation layer is generated; the correlation strength between user nodes and virtual scene nodes in the group's cognitive graph is used. The formula for calculating the focal point F of the group's gaze on the virtual scene node is: ， Wherein, N is the number of participating users, D is a virtual scene set, M is a total number of objects in the virtual scene; i is a user; j is an index identifier of an object in the virtual scene.

5. The virtual reality technology based immersive user experience system as claimed in claim 4, wherein: The process of labeling and initializing the focused gaze point after determining that the focused gaze point is valid is as follows: When F exceeds the threshold value the current state is determined to be valid focus, and annotation initialization is performed. When F is less than or equal to a threshold value If the group attention is determined to be in a state of distraction, the current adjustment mechanism is maintained. After determining the effective focusing, a three-dimensional bounding box is constructed for the three-dimensional space, and key semantic elements are extracted based on semantic analysis An initial marking position is assigned in the three-dimensional space The three-dimensional space bounding box constraint and the group viewing requirement are met.

6. The virtual reality technology based immersive user experience system of claim 5, wherein: The process of optimizing the label attributes is as follows: Optimizing annotation attributes through group cognitive mapping The formula is: ， wherein, is a base color; is an adjustment factor; is a gaze time of user i on region. K is the total number of regions divided in the virtual scene; k' is the semantic element index; is a semantic region in the virtual space; is a gaze time of user i on the virtual scene region.

7. The virtual reality technology based immersive user experience system as claimed in claim 6, wherein: The process of generating the three-dimensional label layer through multi-person collaborative editing is as follows: Recording user operation sequences when multiple people are collaboratively editing Computing operation priorities Determining operation execution order according to priorities, resolving editing conflicts.

8. The virtual reality technology based immersive user experience system, as claimed in claim 1, wherein: The process of introducing the visual perspective output to render the shared perspective based on the focused gaze point is as follows: By defining the visual perspective of user i as a set of camera parameters , the perspective image is compressed into a low-bandwidth data stream using a deep learning image compression algorithm, and the receiving party decodes to obtain a shared perspective image , superimposing the original user's gaze point marker , the gaze point position is determined by eye movement data , and rendered into a shared perspective with gaze markers.

9. The virtual reality technology based immersive user experience system as claimed in claim 8, wherein: The process of acquiring the attention distribution weight of the user in the gaze area to output the warning icon position is as follows: Obtaining attention-related features by a collection device , a support vector machine (SVM) classifier is used to obtain a classification result y, after judging the detection event, a warning icon position is calculated according to a user space position and an enemy position , and a formula is as follows: ​ ， wherein, is a position fusion coefficient.

10. A method of immersive user experience based on virtual reality technology, characterized in that: The method comprises the following steps: Step one: collecting user behavior characteristics through a plurality of sensors and constructing a group cognitive graph model; the user behavior characteristics include interaction preferences, attention distribution, and role behavior patterns; Step two: calculating a focused gaze point of a user node on a virtual scene based on an association strength of the user node on the virtual scene in the group cognitive graph model, labeling and initializing the focused gaze point after determining that the focused gaze point is valid, optimizing label attributes, and generating a three-dimensional label layer through multi-person collaborative editing; Step three: introducing a visual perspective output to render a shared perspective based on the focused gaze point through extended multi-dimensional feature fusion, and acquiring an attention distribution weight of a user in a gaze area through the group cognitive graph model to output a warning icon position.

Images