VR intelligent data caching method and system based on viewport predictability

Abstract

The invention provides a VR intelligent data caching method and system based on viewport predictability. The method comprises the following steps: S1, setting a distributed intelligent storage management caching architecture based on a VR environment; the intelligent storage management cache architecture divides edge nodes into edge servers and data consumers according to functions; s2, predicting future viewpoints of the user according to the viewpoint attenuation mechanism and the parameter information of the user; the viewpoint attenuation mechanism generates an FOV of the user; the parameter information of the user comprises historical head trajectory information of the user; and S3, caching the popular user tiling based on a dynamic tiling caching method. According to the method, efficient data management and storage can be carried out on the edge server, meanwhile, high response support is provided for a data-intensive VR application program, the accuracy and reliability of viewpoint prediction and paving cache algorithms are remarkably improved, and the response time and bandwidth needed when a user requests data are saved.

Description

Technical Field

[0001] The present invention relates to the field of VR data transmission, and in particular to a VR intelligent data caching method and system based on viewport predictability. Background Art

[0002] With the rise of the Metaverse concept, 360° video and virtual reality (VR) as its infrastructure technologies have regained widespread attention in various fields such as entertainment, education, manufacturing, transportation, and medical care.

[0003] VR has shown great commercial and industrial potential due to its ability to provide users with an immersive experience. Users generally watch VR videos through head-mounted displays for a better viewing experience. However, VR transmission streams currently face two major challenges: ultra-high bandwidth challenges and ultra-low latency challenges; and VR head-mounted displays lack portability and mobility.

[0004] Currently, the two mainstream technologies for solving the two major challenges of VR transmission streams are tile-based VR video stream encoding and cloud-based / edge-based architectures used in 360° VR video stream transmission. Tile-based VR video stream encoding refers to dividing the two-dimensional plane obtained by equirectangular projection technology into several tiles of equal size; at this time, the VR video stream only needs to transmit these tiles instead of the entire picture, thereby significantly saving bandwidth. The cloud-based / edge-based architecture used in 360° VR video stream transmission can complete computing-intensive tasks on cloud / edge servers and stream VR videos to users, thereby significantly alleviating the computing and storage pressures of user-side VR devices during the processing process.

[0005] However, in the future metaverse environment, VR will be responsible for bridging the physical world and the virtual world, and data will inevitably grow explosively in the process. Traditional tile-based VR video stream encoding and cloud / edge architecture methods cannot efficiently store and read data, which will greatly increase the cost required for users to request data. At the same time, since most VR devices are resource-limited, and the cloud / edge-based architecture used in 360° VR video streaming transmission is highly dependent on the computing power and caching capabilities of the cloud / edge server, there is a risk of single point failure.

[0006] In the Chinese patent document with publication number CN113920277A, a CloudVR data transmission method and system based on edge computing storage is disclosed, including: using a four-sided pyramid projection method to map and calculate the panoramic image, dividing each panoramic image into multiple video versions with different field of view angles according to the viewpoint projection for storage; obtaining and dynamically predicting the three-dimensional motion data of the user-end device and calculating the user's perspective information; according to the user's perspective information, selecting the video version of the corresponding field of view angle, and performing a pyramid mapping operation on the video version to obtain stereoscopic image data, calculating the viewpoint position movement vector corresponding to the perspective switching, and transmitting the stereoscopic image data to the user-end device. However, the patent document uses technologies such as pyramid mapping operations, while the present invention uses technologies such as viewpoint attenuation mechanisms, so the present invention is essentially different from the technologies and methods used in the patent document. Summary of the invention

[0007] In view of the defects in the prior art, the purpose of the present invention is to provide a VR intelligent data caching method and system based on viewport predictability.

[0008] According to the present invention, a VR intelligent data caching method based on viewport predictability is provided, comprising:

[0009] Step S1: Setting a distributed intelligent storage management cache architecture based on a VR environment;

[0010] The intelligent storage management cache architecture divides edge nodes into edge servers and data consumers according to their functions;

[0011] Step S2: predicting the user's future viewpoint based on the viewpoint attenuation mechanism and the user's parameter information;

[0012] The viewpoint attenuation mechanism generates the user's FOV;

[0013] The user's parameter information includes the user's historical head trajectory information;

[0014] Step S3: caching popular user tiles based on a dynamic tile caching method.

[0015] Preferably, the edge server includes a Master module and a data container; the Master module has communication and management functions; and the data container completes the storage tasks assigned by the Master module.

[0016] Preferably, step S2 extracts features based on historical viewport information of VR videos watched by the user, generates an initial label matrix and records the user's historical viewing information.

[0017] Preferably, the viewpoint attenuation mechanism selects a viewpoint attenuation parameter, and attenuates the tiled labels covered by the user's viewport on the initial label matrix circle by circle with the user as the center to generate an attenuated label matrix; and predicts the user's future viewpoint based on deep learning to generate the user's FOV.

[0018] Preferably, the dynamic tiling cache method stores the predicted user's FOV in the cloud or in a local data storage container according to its popularity; the decision-making process of the Master is regarded as a Markov chain, and the Master chooses whether to cache in the local data storage container according to the strategy π, and the optimal decision-making process of finding the Master is converted into an optimization problem:

[0019] π * (s) = argmax a∈A (Q(s,a)),s∈S;

[0020] Among them, Q(s,a) represents the expected reward obtained by performing action a in state s.

[0021] According to the present invention, a VR intelligent data caching system based on viewport predictability is provided, comprising:

[0022] Intelligent storage management cache architecture module: Set up a distributed intelligent storage management cache architecture and divide edge nodes into edge servers and data consumers according to their functions;

[0023] Viewpoint prediction module: predicts the user's future viewpoint based on the viewpoint attenuation mechanism and the user's parameter information;

[0024] The viewpoint attenuation mechanism generates the user's FOV;

[0025] The user's parameter information includes the user's historical head trajectory information;

[0026] Tile caching function module: cache popular user tiles based on dynamic tile caching method;

[0027] Communication management function module: manages the request flow and data flow between the cloud and data consumers, as well as the data interaction with the adjacent edge server;

[0028] Multi-user data writing and mapping function module: manages data storage containers and storage space, and supports multiple users to write to a file.

[0029] Preferably, the edge server includes a Master module and a data container; the Master module has communication and management functions; and the data container completes the storage tasks assigned by the Master module.

[0030] Preferably, the viewpoint prediction module extracts features based on historical viewport information of VR videos watched by the user, generates an initial label matrix and records the user's historical viewing information.

[0031] Preferably, the viewpoint attenuation mechanism selects a viewpoint attenuation parameter, and attenuates the tiled labels covered by the user's viewport on the initial label matrix circle by circle with the user as the center to generate an attenuated label matrix; and predicts the user's future viewpoint based on deep learning to generate the user's FOV.

[0032] Preferably, the dynamic tiling cache method stores the predicted user's FOV in the cloud or in a local data storage container according to its popularity; the decision-making process of the Master is regarded as a Markov chain, and the Master chooses whether to cache in the local data storage container according to the strategy π, and the optimal decision-making process of finding the Master is converted into an optimization problem:

[0033] π * (s) = argmax a∈A (Q(s,a)),s∈S;

[0034] Among them, Q(s,a) represents the expected reward obtained by performing action a in state s.

[0035] Compared with the prior art, the present invention has the following beneficial effects:

[0036] 1. The method provided by the present invention designs a viewpoint attenuation mechanism to transmit the user's video viewing history information. The optimized cache structure can significantly improve the accuracy and reliability of the viewpoint prediction and tiling cache algorithm for data sets with different head movement rates.

[0037] 2. The method provided by the present invention proposes a distributed intelligent storage management cache architecture suitable for VR environment. The architecture applies the idea of ​​distributed storage in cloud computing to edge computing to solve the ultra-high bandwidth and ultra-low latency problems encountered in current VR transmission streams. By using distributed storage, efficient data management and storage can be performed on edge servers, while providing high-response support for data-intensive VR applications.

[0038] 3. The method provided by the present invention is based on deep learning and reinforcement learning, which samples and marks the user's historical viewing information, and further predicts the user's future viewpoint; at the same time, according to the popularity of the user's VR viewing, the popular content is cached in a cache data container in a distributed intelligent storage management cache architecture, which significantly saves the response time and bandwidth required when the user requests data.

[0039] 4. The method and system provided by the present invention can help solve the ultra-high bandwidth challenges and ultra-low latency challenges currently faced by VR transmission, and can be effective even on data sets with different user head movement rates; the distributed intelligent storage management system provided in the VR environment provides a way to realize VR transmission under limited resource conditions on the edge side; the present invention establishes a more powerful and predictable intelligent tiled caching method in the edge environment with limited resources, which is of great significance to the infrastructure construction of the metaverse in terms of VR transmission streams.

[0040] Other beneficial effects of the present invention will be explained in the specific implementation manner through the introduction of specific technical features and technical solutions. Through the introduction of these technical features and technical solutions, those skilled in the art should be able to understand the beneficial technical effects brought about by the technical features and technical solutions. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0042] Figure 1 Schematic diagram of a distributed intelligent storage management cache architecture suitable for a VR environment in an embodiment of the present invention.

[0043] Figure 2 Schematic diagram of a viewpoint prediction and tile caching algorithm based on deep learning and reinforcement learning in an embodiment of the present invention. DETAILED DESCRIPTION

[0044] The present invention is described in detail below in conjunction with specific embodiments. The following embodiments will help those skilled in the art to further understand the present invention, but are not intended to limit the present invention in any form. It should be noted that, for those of ordinary skill in the art, several changes and improvements can also be made without departing from the concept of the present invention. These all belong to the protection scope of the present invention.

[0045] The present invention discloses a VR intelligent data caching method based on viewport predictability, comprising:

[0046] Reference Figure 1 As shown, first, a distributed intelligent storage management cache architecture suitable for VR environment is designed to provide a system architecture design for a predictable intelligent tiled cache method.

[0047] Among them, the intelligent storage management cache architecture divides edge nodes into edge servers and data consumers according to their functions, and the VR head-mounted display is regarded as a data consumer because it receives user requests and forwards them to the edge server.

[0048] The edge server also includes a Master module with communication and management functions and a data container that completes the storage tasks assigned by the Master module. The Master module has three main functions: an intelligent viewpoint prediction algorithm based on artificial intelligence and dynamic tile caching, communication management, and multi-user data mapping and storage. The main function of the Master is to predict future viewports and dynamically store content according to customer requirements. By tracking past changes in user requests, the Master predicts future requests and dynamically saves the most popular content to reduce the bandwidth and latency required for 360° video streaming. In addition, the Master is also responsible for managing request flows and data exchanges with adjacent edge servers. The data mapping table makes it easy to locate data, and it allows the stored data to be modified and deleted, and a file can have multiple writers.

[0049] Reference Figure 2 As shown in the figure, the user's future viewpoint is predicted based on the viewpoint attenuation mechanism and the user's historical head trajectory information. The output is a probability matrix of the future viewpoint falling on a 2D plane tile. The two tile centers with the highest probability are selected as the predicted future viewpoints, and the FOV is generated and delivered to the user in advance. The accuracy of viewpoint prediction and the efficiency of pre-transmitted FOV can be balanced by adjusting the radius of the generated FOV.

[0050] Among them, the viewpoint attenuation mechanism first selects a viewpoint attenuation parameter, and attenuates the tiled labels covered by the user's viewport on the initial label matrix in a circle with the user as the center, generates an attenuated label matrix, and predicts the user's future viewpoint based on deep learning to generate the user's FOV. Features are extracted based on the historical viewport information of the VR video watched by the user, and an initial label matrix is ​​generated to record the user's historical viewing information. The 360° video is projected into a 360°×180° 2D plane through equirectangular projection, and the 2D plane is divided into 162 equal-sized 20°×20° tiles. The present invention designs a 162×20 two-dimensional initial label matrix. For example, the method first samples the 360° video at a frequency of 10Hz, and uses 2s as a prediction window. The method predicts the user's future viewpoint based on the change of the user's viewpoint in this prediction window. The first dimension of the two-dimensional initial label matrix in this method represents the 162 tiles into which the 2D plane is divided, and the second dimension represents the samples in the prediction window. The initial label matrix is ​​a Boolean matrix, in which the tiles covered by the user FOV are marked as 1 and the rest are marked as 0.

[0051] For example, the viewpoint attenuation mechanism processes the initial label matrix to obtain an optimized label matrix and pre-sets a viewpoint attenuation parameter δ. In the optimized label matrix, the tile where the user's viewpoint is located will still be marked as 1, and the rest of the tiles covered by the FOV will be marked by multiplying δ for every 20° distance between the center and the viewpoint, that is, if the distance is k*20°, it will be marked as δ k The optimized label matrix is ​​the input of the intelligent viewpoint prediction method.

[0052] Finally, the previously obtained FOV is used as input to count the popularity characteristics of each tile, and popular tiles are cached based on the dynamic tile caching method, so that users can directly obtain the cached video content from the cache when requesting it again in subsequent requests without having to obtain it from the cloud, thereby reducing the delay required for VR transmission streaming.

[0053] The dynamic tiled cache method stores the predicted user's FOV in the cloud or local data storage container according to its popularity. The Master's decision process is regarded as a Markov chain. The Master chooses whether to cache in the local data storage container according to the strategy π, thus converting the best decision process of finding the Master into an optimization problem:

[0054] π * (s) = argmax a∈A (Q(s,a)),s∈S;

[0055] Among them, Q(s,a) represents the expected reward obtained by performing action a in state s.

[0056] The above is a basic embodiment of the present invention. The technical solution of the present invention is further described below through a preferred embodiment.

[0057] Reference Figure 1 As shown, first, a distributed intelligent storage management cache architecture suitable for VR environment is designed to provide a system architecture design for a predictable intelligent tiled cache method.

[0058] Among them, the intelligent storage management cache architecture divides edge nodes into edge servers and data consumers according to their functions, and the VR head-mounted display is regarded as a data consumer because it receives user requests and forwards them to the edge server.

[0059] The edge server further includes a Master module with communication and management functions and data containers that complete the storage tasks assigned by the Master module. The Master module mainly has three functions: an intelligent view point prediction algorithm and dynamic tiling caching based on artificial intelligence, communication management, and multi-user data mapping and storage. The most important function of the Master is to predict the future viewport and dynamically store content according to the requirements of customers. By tracking the changes in past user requests, the Master predicts future requests and dynamically saves the most popular content to reduce the bandwidth and latency required for 360° video streaming. In addition, the Master is also responsible for managing the request flow and data exchange with adjacent edge servers. The data mapping table can conveniently locate data. At the same time, it allows the modification and deletion of stored data, and a file can have multiple writers.

[0060] Referring to Figure 2 As shown, according to the view point attenuation mechanism and the historical head trajectory information of the user, the future view point of the user is predicted. The output is a probability matrix of the future view point falling on the 2D plane tiling. Select the two tiling centers with the highest probability as the predicted future view point, and generate the FOV in advance for the user. The accuracy of view point prediction and the efficiency of pre-transmitted FOV can be balanced by adjusting the radius of the generated FOV.

[0061] Among them, the view point attenuation mechanism first selects a view point attenuation parameter, and attenuates the tiling labels covered by the user's viewport circle by circle on the initial label matrix centered on the user to generate an attenuated label matrix, and predicts the future view point of the user based on deep learning to generate the FOV of the user.

[0062] Specifically, for the initial label matrix V, oversampling or undersampling operations are performed on the data set, and the data set format is converted to 10Hz. The dimension of the initial label matrix V is 162×20, where 162 represents 162 tiles in the 2D plane, and 20 represents 2s of time; 20 FOV information of the 2D plane is sampled in the prediction window. In the time prediction window, when the user's FOV covers all or part of the kth tile in the tth time frame, the element v k,t takes a value of 1, otherwise it is 0.

[0063] For the optimized label matrix H obtained after being processed by the view point attenuation mechanism, H is also a 162×20 matrix, where 162 represents 162 tiles in the 2D plane, and 20 represents 20 FOV information of the 2D plane sampled in the time prediction window. The view point attenuation mechanism pre-sets a view point attenuation parameter δ. In the matrix H, when the distance D between the center point of a tile and the center point of the tile where the user's view point is located satisfies (m - 1)·l < D ≤ m·l, the tile represented by this element in the matrix H is marked as δ m, where l is the side length of each tile. Using the matrix H as the input of the LSTM algorithm, a probability matrix can be obtained, and each element in the matrix represents the possibility of the predicted future viewpoint in different tiles.

[0064] Finally, popular user tiles are cached based on a dynamic tile caching method to reduce the latency required for VR transmission streaming.

[0065] Specifically, each Master independently designs its own optimized cache and content replacement strategy. For one Mastern, n∈N, N represents the total number of Masters, and the state at a certain moment is s, s satisfies s∈S, S represents all possible states. Each state is obtained by extracting specific features from the user's historical requests. The state space S n It can be obtained by counting the number of requests for each tile from the beginning to the current time. When the user receives a new request w tk When tk Representing the kth tile of the tth frame, the Master may encounter the following two possible situations:

[0066] a) The tiles requested by the user are cached;

[0067] b) The tile requested by the user is not cached.

[0068] For the first case a), the Master will keep the current cache unchanged;

[0069] For the second case b), the Master now has a possible operation set A of size C+1. tk . Where C is the size of the cache space. tk The collection contains elements {A tk0 ,A tk1 ,...,A tki ,...,A tkC}, where A tk0 Indicates that the content cached on the Master remains unchanged, while A tki All tiles of the i-th video cached on the Master will be replaced by the requested tile. Therefore, at time t, the action space A t That is A t =∏ k∈K A tk Considering the decision-making process of the Master as a Markov chain, the Master chooses whether to cache the requested tile in the local data storage container according to the strategy π. Then, finding the best decision process of the Master is transformed into an optimization problem:

[0070] π * (s) = argmaxa∈A (Q(s,a)),s∈S.

[0071] Among them, Q(s,a) is the expected reward obtained by performing action a in state s. The optimal cache strategy of the Master can be obtained by iterating the optimization problem using the reinforcement learning DQN method until convergence.

[0072] For the evaluation matrix, three indicators, namely viewpoint prediction accuracy, cache hit rate and cache utilization, are used to evaluate the performance of the intelligent caching method.

[0073] The viewpoint prediction accuracy is the ratio of the number of FOVs generated by the intelligent viewpoint prediction algorithm where the user's real viewpoint falls to the sum of the total number of the user's real viewpoints; the cache hit rate is the ratio of the cache tiles hit by the dynamic tile cache algorithm to the total number of tiles; the cache utilization is the ratio of the cache tiles hit by the dynamic tile cache algorithm to the cache capacity.

[0074] This embodiment, based on deep learning and reinforcement learning, proposes a distributed intelligent storage management cache architecture suitable for VR environments, and a system architecture design provided by a predictable intelligent tiled cache method. By utilizing an intelligent viewpoint prediction algorithm and a dynamic tiled cache algorithm, the user's future viewpoint can be predicted while generating the user's field of view, and popular video content can be cached according to the popularity of the user's VR video content request, greatly reducing the bandwidth and delay burden of the VR video transmission stream. In summary, the present invention has established a more powerful and predictable intelligent tiled cache method in an edge environment with limited resources, which is of great significance to the infrastructure construction of the metaverse in terms of VR transmission streams.

[0075] The present invention also provides a VR intelligent data caching system based on viewport predictability, which can be implemented by executing the process steps of the VR intelligent data caching method based on viewport predictability, that is, those skilled in the art can understand the VR intelligent data caching method based on viewport predictability as a preferred implementation of the VR intelligent data caching system based on viewport predictability.

[0076] Specifically, a VR intelligent data caching system based on viewport predictability includes:

[0077] Intelligent storage management cache architecture module: Set up a distributed intelligent storage management cache architecture and divide edge nodes into edge servers and data consumers according to their functions;

[0078] Viewpoint prediction module: predicts the user's future viewpoint based on the viewpoint attenuation mechanism and the user's parameter information;

[0079] The viewpoint attenuation mechanism generates the user's FOV;

[0080] The user's parameter information includes the user's historical head trajectory information;

[0081] Tile caching function module: cache popular user tiles based on dynamic tile caching method;

[0082] Communication management function module: manages the request flow and data flow between the cloud and data consumers, as well as the data interaction with the adjacent edge server;

[0083] Multi-user data writing and mapping function module: manages data storage containers and storage space, and supports multiple users to write to a file.

[0084] The edge server includes a Master module and a data container; the Master module has communication and management functions; and the data container completes the storage task assigned by the Master module.

[0085] The viewpoint prediction module extracts features based on the historical viewport information of the VR video watched by the user, and generates an initial label matrix to record the user's historical viewing information.

[0086] The viewpoint attenuation mechanism selects a viewpoint attenuation parameter, and attenuates the tiled labels covered by the user's viewport on the initial label matrix in a circle with the user as the center to generate an attenuated label matrix; the user's future viewpoint is predicted based on deep learning to generate the user's FOV.

[0087] The dynamic tiled caching method stores the predicted user's FOV in the cloud or in a local data storage container according to its popularity. The decision-making process of the Master is regarded as a Markov chain. The Master chooses whether to cache in the local data storage container according to the strategy π, and the optimal decision-making process of finding the Master is transformed into an optimization problem:

[0088] π * (s) = argmax a∈A (Q(s,a)),s∈S;

[0089] Among them, Q(s,a) represents the expected reward obtained by performing action a in state s.

[0090] The above describes the specific embodiments of the present invention. It should be understood that the present invention is not limited to the above specific embodiments, and those skilled in the art can make various changes or modifications within the scope of the claims, which does not affect the essence of the present invention. In the absence of conflict, the embodiments of the present application and the features in the embodiments can be combined with each other arbitrarily.

Claims

1. A VR intelligent data caching method based on viewport predictability, characterized in that: include: Step S1: Setting a distributed intelligent storage management cache architecture based on a VR environment; The intelligent storage management cache architecture divides edge nodes into edge servers and data consumers according to their functions; Step S2: predicting the user's future viewpoint based on the viewpoint attenuation mechanism and the user's parameter information; The viewpoint attenuation mechanism generates the user's FOV; The user's parameter information includes the user's historical head trajectory information; Step S3: caching popular user tiles based on a dynamic tile caching method.

2. The method for VR intelligent data caching based on viewport predictability according to claim 1, characterized in that: The edge server includes a Master module and a data container; the Master module has communication and management functions; and the data container completes the storage task assigned by the Master module.

3. The method for VR intelligent data caching based on viewport predictability according to claim 1, characterized in that: The step S2 extracts features based on the historical viewport information of the VR videos watched by the user, generates an initial label matrix and records the historical viewing information of the user.

4. The method for VR intelligent data caching based on viewport predictability according to claim 1, characterized in that: The viewpoint attenuation mechanism selects a viewpoint attenuation parameter, and attenuates the tiled labels covered by the user's viewport on the initial label matrix in a circle with the user as the center to generate an attenuated label matrix; the user's future viewpoint is predicted based on deep learning to generate the user's FOV.

5. The method for VR intelligent data caching based on viewport predictability according to claim 1, characterized in that: The dynamic tiled caching method stores the predicted user's FOV in the cloud or in a local data storage container according to its popularity. The decision-making process of the Master is regarded as a Markov chain. The Master chooses whether to cache in the local data storage container according to the strategy π, and the optimal decision-making process of finding the Master is transformed into an optimization problem: π * (s)=argmax a∈A (Q(s,a)),s∈S; Among them, Q(s,a) represents the expected reward obtained by performing action a in state s.

6. A VR intelligent data caching system based on viewport predictability, characterized in that: include: Intelligent storage management cache architecture module: Set up a distributed intelligent storage management cache architecture and divide edge nodes into edge servers and data consumers according to their functions; Viewpoint prediction module: predicts the user's future viewpoint based on the viewpoint attenuation mechanism and the user's parameter information; The viewpoint attenuation mechanism generates the user's FOV; The user's parameter information includes the user's historical head trajectory information; Tile caching function module: cache popular user tiles based on dynamic tile caching method; Communication management function module: manages the request flow and data flow between the cloud and data consumers, as well as the data interaction with the adjacent edge server; Multi-user data writing and mapping function module: manages data storage containers and storage space, and supports multiple users to write to a file.

7. The VR intelligent data caching system based on viewport predictability according to claim 6, characterized in that: The edge server includes a Master module and a data container; the Master module has communication and management functions; and the data container completes the storage task assigned by the Master module.

8. The VR intelligent data caching system based on viewport predictability according to claim 6, characterized in that: The viewpoint prediction module extracts features based on the historical viewport information of the VR video watched by the user, generates an initial label matrix and records the user's historical viewing information.

9. The VR intelligent data caching system based on viewport predictability according to claim 6, characterized in that: The viewpoint attenuation mechanism selects a viewpoint attenuation parameter, and attenuates the tiled labels covered by the user's viewport on the initial label matrix in a circle with the user as the center to generate an attenuated label matrix; the user's future viewpoint is predicted based on deep learning to generate the user's FOV.

10. The VR intelligent data caching system based on viewport predictability according to claim 6, characterized in that: The dynamic tiled caching method stores the predicted user's FOV in the cloud or in a local data storage container according to its popularity. The decision-making process of the Master is regarded as a Markov chain. The Master chooses whether to cache in the local data storage container according to the strategy π, and the optimal decision-making process of finding the Master is transformed into an optimization problem: π * (s)=argmax a∈A (Q(s,a)),s∈S; Among them, Q(s,a) represents the expected reward obtained by performing action a in state s.

Images