CN117331622B - AssetBundle-based webpage version museum scene loading method - Google Patents
AssetBundle-based webpage version museum scene loading method Download PDFInfo
- Publication number
- CN117331622B CN117331622B CN202311337077.3A CN202311337077A CN117331622B CN 117331622 B CN117331622 B CN 117331622B CN 202311337077 A CN202311337077 A CN 202311337077A CN 117331622 B CN117331622 B CN 117331622B
- Authority
- CN
- China
- Prior art keywords
- scene
- user
- knowledge
- loading
- assetbundle
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000011068 loading method Methods 0.000 title claims abstract description 149
- 230000007704 transition Effects 0.000 claims abstract description 26
- 238000012856 packing Methods 0.000 claims abstract description 6
- 230000033001 locomotion Effects 0.000 claims description 56
- 238000000034 method Methods 0.000 claims description 38
- 230000008569 process Effects 0.000 claims description 20
- 238000002372 labelling Methods 0.000 claims description 15
- 230000006870 function Effects 0.000 claims description 13
- 238000004364 calculation method Methods 0.000 claims description 11
- 238000001514 detection method Methods 0.000 claims description 9
- 238000005516 engineering process Methods 0.000 claims description 9
- 238000012544 monitoring process Methods 0.000 claims description 8
- 230000000007 visual effect Effects 0.000 claims description 8
- 239000000463 material Substances 0.000 claims description 7
- 238000005457 optimization Methods 0.000 claims description 6
- 230000006837 decompression Effects 0.000 claims description 4
- 238000013461 design Methods 0.000 claims description 4
- 239000000203 mixture Substances 0.000 claims description 4
- 238000004806 packaging method and process Methods 0.000 claims description 4
- 230000002776 aggregation Effects 0.000 claims description 3
- 238000004220 aggregation Methods 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 3
- 238000004064 recycling Methods 0.000 claims description 3
- 238000007726 management method Methods 0.000 description 4
- 230000004044 response Effects 0.000 description 3
- 230000006399 behavior Effects 0.000 description 2
- 230000015556 catabolic process Effects 0.000 description 2
- 230000001427 coherent effect Effects 0.000 description 2
- 238000006731 degradation reaction Methods 0.000 description 2
- 238000000605 extraction Methods 0.000 description 2
- 238000000227 grinding Methods 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000009877 rendering Methods 0.000 description 2
- 230000011218 segmentation Effects 0.000 description 2
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012536 packaging technology Methods 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000004043 responsiveness Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44568—Immediately runnable code
- G06F9/44578—Preparing or optimising for loading
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F8/00—Arrangements for software engineering
- G06F8/70—Software maintenance or management
- G06F8/71—Version control; Configuration management
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F9/00—Arrangements for program control, e.g. control units
- G06F9/06—Arrangements for program control, e.g. control units using stored programs, i.e. using an internal store of processing equipment to receive or retain programs
- G06F9/44—Arrangements for executing specific programs
- G06F9/445—Program loading or initiating
- G06F9/44521—Dynamic linking or loading; Link editing at or after load time, e.g. Java class loading
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
Landscapes
- Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Health & Medical Sciences (AREA)
- Computer Security & Cryptography (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Information Transfer Between Computers (AREA)
Abstract
The invention provides a AssetBundle-based webpage version museum scene loading method, which comprises the following steps of: packing museum scene resources, generating AssetBundle files and storing the AssetBundle files into a database; distinguishing scene loading types according to user selection, performing space transition type scene loading when the user moves across the scene, and performing space switching type scene loading when the user moves across the exhibition area; when a user loads a space transition type scene, carrying out scene model instantiation according to AssetBundle files, predicting a moving route of the user, and preloading a scene passed by the user according to the moving route of the user; when the user loads the space switching type scene, the scene model is instantiated according to AssetBundle files, similar scenes in the system knowledge nodes or knowledge field are searched based on the scene model, and the similar scenes are preloaded.
Description
Technical Field
The invention relates to the technical field of multimedia exhibition and display, in particular to a AssetBundle-based webpage version museum scene loading method.
Background
In the traditional webpage 3D scene display, the scene model data volume is huge, and the overall loading and rendering can lead to the problems of user experience degradation, insufficient browser memory and overlong loading time.
Chinese patent CN116012523a, a virtual three-dimensional exhibition hall-based scene loading method, system and storage medium, discloses a virtual three-dimensional exhibition hall-based scene loading method, which loads scene data by determining the scene data to be loaded. However, when the method faces a scene with excessive resource data quantity, segmentation and segmentation downloading cannot be performed, and the travelling direction of the user during actual browsing cannot be predicted, so that the user cannot feel the feeling of being personally on the scene.
Disclosure of Invention
In view of this, the invention provides a method for loading a museum scene of a webpage version based on AssetBundle, which is developed based on AssetBundle technology, packages a grid model, materials and maps of the museum scene to reduce the memory pressure and increase the loading speed, predicts the scene of the loading user in the moving direction through the advancing direction, optimizes the user experience, and solves the problems that the user needs to wait for too long loading time and the browsing experience is not real enough when browsing the museum scene of the webpage version.
The technical scheme of the invention is realized as follows: the invention provides a AssetBundle-based webpage version museum scene loading method, which comprises the following steps of:
s1, packing museum scene resources, generating AssetBundle files and storing the AssetBundle files into a database;
S2, distinguishing scene loading types according to user selection, performing space transition type scene loading when the user moves across the scene, and performing space switching type scene loading when the user moves across the exhibition area;
S3, when the user loads the space transition type scene, carrying out scene model instantiation according to AssetBundle files, predicting the moving route of the user, and preloading the scene passed by the user according to the moving route of the user;
s4, when the user loads the space switching type scene, the scene model is instantiated according to AssetBundle files, similar scenes in the system knowledge nodes or knowledge field are searched based on the scene model, and the similar scenes are preloaded;
The preloading operation is performed by using idle bandwidth resources only when the bandwidth is in an idle state.
Preferably, step S1 comprises the steps of:
collecting and packaging a grid model, materials and a map of a museum scene;
Creating AssetBundle a file using a Unity engine, the AssetBundle file comprising museum scene resources;
and storing the generated AssetBundle files in a database.
Preferably, step S2 comprises the steps of:
Distinguishing the user movement type according to the operation of the user movement scene, loading resources according to the user movement type, and loading a space transition scene by the user when the user switches between adjacent exhibition halls; when a user switches between different exhibition areas, the user loads a space switching type scene;
the step of loading the space transition type scene is as follows:
detecting a movement intention of a user;
loading AssetBundle resource packages of the next scene according to the mobile intention of the user;
and preloading the exhibition hall scene to be browsed by the user according to AssetBundle resource packages of the next scene.
Preferably, the process of detecting the movement intention of the user is as follows:
Forming a motion vector array through the triggering sequence of the multi-layer triggers, and performing unified calculation to obtain a final movement intention vector;
judging the direction and length of the movement intention vector, and determining the specific movement intention of the user;
when the movement intention vector is in the positive direction and the length reaches the requirement, the movement intention of the user is judged to be forward, when the movement intention vector is in the opposite direction, the movement intention of the user is judged to be backward, and when the movement intention vector is not in the sufficient length, the movement intention of the user is judged to be backward.
Preferably, the process of loading AssetBundle resource packages of the next scene according to the mobile intention of the user is as follows:
And loading a scene AssetBundle resource package of the next exhibition hall according to the movement intention determined by the user, and loading, decompressing and instantiating the selected AssetBundle resource package to finish loading the scene to be loaded.
Preferably, the preloading procedure of the exhibition hall scene to be browsed by the user according to AssetBundle resource packages of the next scene is as follows:
When the scene loading used by the user is completed, the current use requirement of the user is met, the program downloading queue is empty, the bandwidth is in an idle state, the idle bandwidth resource is utilized to predict and calculate the moving direction of the next step of the user, and the exhibition hall which the user goes to next step is preloaded;
predicting the next moving direction of a calculated user by using a pedestrian track tracking technology, calculating a exhibition hall scene covered by a user route, loading a scene model of the exhibition hall which is predicted to be covered according to a prediction result obtained by calculation, and loading adjacent scenes of the current scene by referring to a spatial locality principle in a computer composition principle when the bandwidth is still idle;
When the scene model of the next exhibition hall meets the inequality equation set or the Rect function calculation, judging that the scene model of the exhibition hall is covered by the user route, wherein the inequality equation set is as follows:
Wherein { D Path point } is the set of path points for the next movement of the user, r Exhibition hall is the distance from the current scene of the user to the scene model of the next exhibition hall, and l Exhibition hall is the distance from the center point of the scene model of the next exhibition hall to the edge of the scene model;
the Rect function judges whether a scene model of the exhibition hall is covered by a user route according to the position and the size of the exhibition hall;
In the process of loading the space transition type scene, the downloading and memory optimization of the resources are as follows:
and accelerating the downloading of AssetBundle resource packages to terminal equipment through the CDN, caching the resource packages in a IndexDB database, monitoring the memory of the database by using a front-end performance monitoring function, and recycling garbage according to the sequence of the queues when the occupied memory exceeds a preset threshold.
Preferably, step S4 comprises the steps of:
When a user switches the exhibition areas, the scene resource file of the exhibition area is directly downloaded from the CDN according to the exhibition area selected by the user, decompression and instantiation operations are carried out on terminal equipment, and the scene resource of the selected exhibition area is loaded into the memory;
according to the preset knowledge point node number sequence, a preloading rule of space switching type scene loading is set;
Loading the exhibition halls in the same knowledge point or the same knowledge field as a preloading target according to the preset sequence of knowledge point node trees in the exhibition hall, and carrying out knowledge node association labeling for AssetBundle resource packages when the AssetBundle resource packages of the scene model are uploaded, and inquiring the scene model needing to inquire unified knowledge points according to the knowledge points; and for a scene model in the field of related knowledge to be queried, tracing back to an upper-level root knowledge node through a knowledge point node tree, traversing downwards to obtain all sub-knowledge nodes of the root knowledge node, and performing query preloading according to the sub-knowledge nodes.
Preferably, the construction process of the knowledge point node tree comprises the following steps:
Establishing a first-level root knowledge node according to the correspondence of the exhibition halls, establishing a second-level root knowledge node according to the correspondence of different exhibition halls in the exhibition halls, establishing a sub-knowledge node according to the correspondence of different exhibition halls in each exhibition halls, and forming a knowledge point node tree by the first-level knowledge node, the second-level knowledge node and the sub-node according to the design and the planning of the professional level of the museum;
Extracting image features of scene models of the sub-knowledge nodes, classifying the scene models of the sub-knowledge nodes based on the image features, and labeling the sub-knowledge nodes;
similar label aggregation is carried out on the sub-knowledge nodes according to the labels, and sub-knowledge nodes with the same or similar labels are associated;
When a scene model is inquired, the first-level root knowledge nodes and the corresponding second-level root knowledge nodes are mutually associated, the second-level root knowledge nodes and the corresponding sub-knowledge nodes are mutually associated, and the first-level root knowledge nodes, the second-level root knowledge nodes and the sub-knowledge nodes of the same kind are mutually associated;
when a user uses an exhibition hall scene model, preloading sub-knowledge nodes, secondary root knowledge nodes and scene models corresponding to the primary root knowledge nodes which are associated with the exhibition hall scene model, and traversing the sub-knowledge nodes to obtain a next selected scene model according to the primary root knowledge nodes to the secondary root knowledge nodes.
Preferably, the image feature extraction is performed on the scene model of the sub-knowledge node, the classification is performed on the scene model of the sub-knowledge node based on the image feature, and the labeling process of the sub-knowledge node comprises the following steps:
Constructing a multidimensional prediction model, fusing the features of different dimensions of the scene model through the root semantic information of the high-dimensional features and the sub-semantic information of the low-dimensional features, and carrying out feature detection on the fused feature map;
Classifying the scene model of the sub knowledge node based on the detected characteristic data, and labeling according to the classification result;
The multi-dimensional prediction model performs target detection independently in two dimensions, the root semantic information of the high-dimensional features is an original knowledge point node tree of the scene model, and the sub-semantic information of the low-dimensional features comprises extracted visual image features of the scene model.
Preferably, the process of associating child knowledge nodes with the same or similar labels according to the labels is as follows:
extracting a semantic vector of a current tag, presetting a semantic similarity threshold, and associating the tag with the semantic similarity of the current tag within the threshold with the current tag;
When a scene model is inquired, preloading the scene model corresponding to the sub-knowledge node corresponding to the label associated with the labels according to the inquired labels of the scene model.
Compared with the prior art, the method for loading the scene of the museum of the webpage version based on AssetBundle has the following beneficial effects:
(1) By packing the museum scene resources into AssetBundle files and storing the AssetBundle files into a database, the high-efficiency loading and management of the resources are realized, the network transmission time and loading time are reduced, and the scene loading efficiency is improved;
(2) According to the region and the movement type selected by the user, the space transition type scene loading or the space switching type scene loading is intelligently carried out, smooth scene transition is provided in the navigation process of the user, loading time and the clamping condition are reduced, and the browsing experience of the user is improved;
(3) By utilizing prediction calculation and the moving route of the user, the exhibition hall or the scene to be browsed by the user is loaded in advance, so that the loading delay is reduced, and the operation response speed and the scene switching fluency of the user are improved;
(4) Dynamic loading and unloading of resources are carried out according to the equipment performance and the memory state, so that the memory use efficiency is ensured, the memory pressure of a browser is reduced, and the resource utilization rate is improved;
(5) Through association and preloading according to the knowledge point node tree of the exhibition hall, the efficiency of the user for acquiring the related exhibition hall is optimized, and richer scene browsing experience is provided.
Drawings
In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flow chart of a method for loading a scene of a museum based on AssetBundle web pages.
Detailed Description
The following description of the embodiments of the present invention will clearly and fully describe the technical aspects of the embodiments of the present invention, and it is apparent that the described embodiments are only some embodiments of the present invention, not all embodiments. All other embodiments, which can be made by those skilled in the art based on the embodiments of the present invention without making any inventive effort, are intended to fall within the scope of the present invention.
The method for loading the webpage version museum scene based on AssetBundle is provided, as shown in fig. 1, and comprises the following steps:
s1, packing museum scene resources, generating AssetBundle files and storing the AssetBundle files into a database;
S2, distinguishing scene loading types according to user selection, performing space transition type scene loading when the user moves across the scene, and performing space switching type scene loading when the user moves across the exhibition area;
S3, when the user loads the space transition type scene, carrying out scene model instantiation according to AssetBundle files, predicting the moving route of the user, and preloading the scene passed by the user according to the moving route of the user;
s4, when the user loads the space switching type scene, the scene model is instantiated according to AssetBundle files, similar scenes in the system knowledge nodes or knowledge field are searched based on the scene model, and the similar scenes are preloaded;
The preloading operation is performed by using idle bandwidth resources only when the bandwidth is in an idle state.
It should be noted that: in the traditional webpage 3D scene display, the scene model data volume is huge, and the overall loading and rendering can lead to the problems of user experience degradation, insufficient browser memory and overlong loading time;
in order to optimize the user experience, research and development are performed based on AssetBundle technology, and the grid model, materials and maps of the museum scene are packaged so as to reduce the memory pressure and increase the loading speed. On the basis, a self-grinding intelligent dynamic loading and preloading technology is introduced, so that a large number of 3D graphical files can be efficiently segmented and intelligently downloaded in a slicing way;
In the aspect of space transition type scene loading, the moving direction of a user is accurately judged through prediction of the moving direction, assetBundle packages are loaded in advance and an instantiation scene is decompressed according to the scene in the moving direction of the user, and the next scene can be accurately loaded when the user moves forwards, backwards or forwards and backwards through establishing association of the moving direction vector and the movement intention, so that the user experience is improved;
in the aspect of space switching type scene loading, switching an exhibition area as an access point, carrying out intelligent scene loading according to the switching behavior of a user, and predicting a venue which the user possibly goes to next and loading related scenes based on the preset knowledge point and the associated information of the knowledge field so as to reduce the overall scene loading time;
when the program downloading queue is empty and the bandwidth is in an idle state, the idle bandwidth can be utilized, and preloading is carried out for a venue where a user possibly goes next, and all preloading in the invention is carried out only when the bandwidth is in the idle state;
The method comprises the steps of utilizing AssetBundle dynamic resource packaging technology to carry out block packaging on a scene model, and improving loading speed and memory utilization rate through self-grinding intelligent preloading algorithm and browser memory dynamic load optimization; the scheme aims to provide smoother webpage version 3D museum scene display experience, optimize resource loading and memory management, and reduce loading time and memory occupation. The method can adapt to the characteristics of huge 3D graphical files, improves the browsing experience of users, and provides better virtual visiting experience for the users.
Step S1 comprises the steps of:
collecting and packaging a grid model, materials and a map of a museum scene;
Creating AssetBundle a file using a Unity engine, the AssetBundle file comprising museum scene resources;
and storing the generated AssetBundle files in a database.
It should be noted that:
after the scene resources are packed into AssetBundle files, the scene resources are flexibly managed and loaded through the loading and unloading functions of AssetBundle, so that the efficiency of resource loading can be improved, and the loading time is reduced;
The occupation of storage space can be reduced by packing the resources into AssetBundle files, compared with directly storing a large number of grid models, materials and mapping files, the AssetBundle files can be used for compressing and optimizing the resources, the volume of the files is reduced, and the storage space is saved;
After the museum scene resources are packaged into AssetBundle files, the data structure and the logic codes of the resources can be separated from the logic codes of the game operation, so that maintainability and reusability of the codes are improved, different developers can concentrate on different tasks, and development efficiency is improved;
the grid model, the materials and the map of the museum scene are packaged into AssetBundle files and stored in the database, so that the resource management and loading efficiency can be improved, the storage space can be saved, the data structure and the logic code can be separated, and the cross-platform compatibility can be improved. Providing basic support for subsequent scene loading and running.
Step S2 comprises the steps of:
Distinguishing the user movement type according to the operation of the user movement scene, loading resources according to the user movement type, and loading a space transition scene by the user when the user switches between adjacent exhibition halls; when the user switches between different exhibition areas, the user performs space switching type scene loading.
It should be noted that:
according to different movement types of the user switching between different exhibition halls and between different exhibition areas, resource loading can be carried out according to specific loading requirements and optimization strategies, and by distinguishing the movement types of the user, space transition type scene loading or space switching type scene loading can be carried out according to the needs, so that unnecessary resource loading and memory occupation are reduced, and the speed and efficiency of resource loading are optimized;
the space transition type scene loading is used for loading the scene in the moving direction of the user according to the traveling direction prediction for the user who browses the museum in a normal exhibition hall; the space switching type scene loading is that for a user browsing by using exhibition area switching, the whole time of scene loading is reduced for the intelligent loading of the scene;
According to the moving type of the user, different types of scene loading is carried out, so that scene switching is smoother and more natural, loading delay and clamping phenomenon are reduced, browsing experience of the user is improved, and when the user switches among different exhibition halls of the same exhibition hall, seamless switching and continuous browsing experience can be realized through space transition type scene loading; when a user switches between different exhibition areas, a brand new scene and content of a new exhibition area can be provided through space switching type scene loading, so that the user obtains a brand new experience;
By distinguishing the mobile types of the users and loading the resources at the correct time, unnecessary resource loading and page loading time can be reduced, and the loading speed of the pages can be improved, particularly for large-scale webpage version museum scenes, the resource files are large, and the response speed and user experience of the pages can be improved through accurate resource loading control;
the method has the advantages that the resource loading is carried out according to the mobile type of the user, the resource loading can be optimized, the user experience and the page loading speed are improved, the smooth and coherent scene switching and browsing experience is provided according to the actual demand and the behavior mode of the user through a reasonable scene loading strategy, and the more smooth, natural and efficient museum scene browsing experience is provided for the user.
The step of loading the space transition type scene is as follows:
detecting a movement intention of a user;
loading AssetBundle resource packages of the next scene according to the mobile intention of the user;
and preloading the exhibition hall scene to be browsed by the user according to AssetBundle resource packages of the next scene.
The process of detecting the movement intention of the user comprises the following steps:
Forming a motion vector array through the triggering sequence of the multi-layer triggers, and performing unified calculation to obtain a final movement intention vector;
judging the direction and length of the movement intention vector, and determining the specific movement intention of the user;
when the movement intention vector is in the positive direction and the length reaches the requirement, the movement intention of the user is judged to be forward, when the movement intention vector is in the opposite direction, the movement intention of the user is judged to be backward, and when the movement intention vector is not in the sufficient length, the movement intention of the user is judged to be backward.
The process of loading AssetBundle resource packages of the next scene according to the mobile intention of the user is as follows:
And loading a scene AssetBundle resource package of the next exhibition hall according to the movement intention determined by the user, and loading, decompressing and instantiating the selected AssetBundle resource package to finish loading the scene to be loaded.
The process of preloading the exhibition hall scene to be browsed by the user according to AssetBundle resource packages of the next scene is as follows:
When the scene loading used by the user is completed, the current use requirement of the user is met, the program downloading queue is empty, the bandwidth is in an idle state, the idle bandwidth resource is utilized to predict and calculate the moving direction of the next step of the user, and the exhibition hall which the user goes to next step is preloaded;
predicting the next moving direction of a calculated user by using a pedestrian track tracking technology, calculating a exhibition hall scene covered by a user route, loading a scene model of the exhibition hall which is predicted to be covered according to a prediction result obtained by calculation, and loading adjacent scenes of the current scene by referring to a spatial locality principle in a computer composition principle when the bandwidth is still idle;
When the scene model of the next exhibition hall meets the calculation of an unequal equation set or a Rec function, judging that the scene model of the exhibition hall is covered by a user route, wherein the unequal equation set is as follows:
Wherein { D Path point } is the set of path points for the next movement of the user, r Exhibition hall is the distance from the current scene of the user to the scene model of the next exhibition hall, and l Exhibition hall is the distance from the center point of the scene model of the next exhibition hall to the edge of the scene model;
the Rect function judges whether a scene model of the exhibition hall is covered by a user route according to the position and the size of the exhibition hall;
in the process of loading the space transition type scene, the downloading and memory optimization of the resources are as follows:
and accelerating the downloading of AssetBundle resource packages to terminal equipment through the CDN, caching the resource packages in a IndexDB database, monitoring the memory of the database by using a front-end performance monitoring function, and recycling garbage according to the sequence of the queues when the occupied memory exceeds a preset threshold.
It should be noted that:
By accurately judging the moving direction and the moving intention of the user, the AssetBundle package of the next scene is loaded in advance and the scene is decompressed and instantiated when the user approaches the scene boundary, so that the waiting time of the user can be reduced, the loading speed and the smoothness are improved, and the user can enter the next scene more quickly;
the scene resources can be loaded into the local cache in advance by preloading the scenes which the user possibly passes through next, so that when the user actually needs to browse the next scene, the scene resources are already preloaded, and the resource downloading and loading are not needed, so that the loading time is further shortened, and the user experience is improved;
When a scene which a user needs to use is loaded, a program downloading queue is empty, the bandwidth is in an idle state, the idle bandwidth can be utilized, the scene which the user possibly passes through is preloaded through prediction calculation, idle bandwidth resources are fully utilized, meanwhile, the front-end performance monitoring function is used for monitoring the browser memory, garbage collection can be timely carried out, dynamic load optimization of the browser memory is realized, and the management and utilization efficiency of the resources are improved;
through the space transition type scene loading technology, smooth transition of scene switching and loading in the moving process of a user can be realized, and when the user browses the museum scene, the scene switching can be more coherent and natural, so that the overall efficiency and performance of the page are improved.
The space transition type scene loading technology can improve loading speed and fluency, optimize user experience, fully utilize idle bandwidth and memory resources, improve overall efficiency and performance of pages, and provide a user with a more rapid, fluent and efficient museum scene browsing experience.
Taking a path point array which a user possibly moves to sequentially carry into an unequal equation set consisting of information such as the position, the size and the like of the exhibition hall to calculate, and proving that the exhibition hall scene is covered by the moving path as long as the unequal equation set is established, otherwise, not being covered by the moving path;
Defining a exhibition hall scene in terms of distance, size and position through two inequality, and preloading the exhibition hall scene as long as a user possible moving path point number group has points meeting the distance, size and position conditions of a certain exhibition hall scene, wherein the exhibition hall scene is covered by a user moving route;
whether the scene model of the exhibition hall is covered by the user route can be judged through the Rect matrix information, wherein the Rect matrix information is about the description of the characteristics and the attributes of rectangular functions, is a function commonly used in mathematics and signal processing, and can be expressed as accurately describing the shape and the attributes of a rectangle in a two-dimensional coordinate system;
and then if the bandwidth is still in an idle state, loading adjacent scenes of the current scene as much as possible according to the principle of spatial locality in the computer composition principle so as to cope with the randomness of browsing by a user.
Step S4 comprises the steps of:
When a user switches the exhibition areas, the scene resource file of the exhibition area is directly downloaded from the CDN according to the exhibition area selected by the user, decompression and instantiation operations are carried out on terminal equipment, and the scene resource of the selected exhibition area is loaded into the memory;
and setting a preloading rule of space switching type scene loading according to the preset knowledge point node number sequence.
The step of setting a preloading rule of space switching type scene loading according to a preset knowledge point node tree sequence, and further comprises the following steps:
Loading the exhibition halls in the same knowledge point or the same knowledge field as a preloading target according to the preset sequence of knowledge point node trees in the exhibition hall, and carrying out knowledge node association labeling for AssetBundle resource packages when the AssetBundle resource packages of the scene model are uploaded, and inquiring the scene model needing to inquire unified knowledge points according to the knowledge points; and for a scene model in the field of related knowledge to be queried, tracing back to an upper-level root knowledge node through a knowledge point node tree, traversing downwards to obtain all sub-knowledge nodes of the root knowledge node, and performing query preloading according to the sub-knowledge nodes.
It should be noted that:
the loading time of scene switching can be reduced by directly downloading the scene resource file to be loaded from the CDN and carrying out decompression and instantiation operation, and a user can enter a new exhibition area more quickly and begin browsing new scene content quickly when switching the exhibition areas;
Through space switching type scene loading, a brand new scene model and content of a new exhibition area can be loaded, so that brand new exhibition environment and exhibition content are provided for users, and diversity and interestingness of virtual visit of the museum are increased;
through carrying out knowledge point association labeling on AssetBundle resource packages of a scene model, in space switching type scene loading, preloading is carried out according to the sequence of knowledge point node trees preset by an exhibition hall, and the exhibition hall in the same knowledge point or the same knowledge field is used as a preloading target to load so as to enrich venue browsing experience of users and provide a more organized knowledge acquisition mode;
By rapidly loading the scene model of the new exhibition area, the overall loading speed of the page can be increased, the responsiveness and fluency of the page are improved, and a user can enter the new scene more rapidly when switching the exhibition area, so that the overall efficiency and performance of the page are improved.
The space switching type scene loading technology can accelerate scene switching speed, provides brand new experience of a new exhibition area, enriches venue browsing experience of users through preloaded rules based on knowledge point association, and provides more rapid, diversified and efficient museum scene switching and browsing experience for the users.
The construction process of the knowledge point node tree comprises the following steps:
Establishing a first-level root knowledge node according to the correspondence of the exhibition halls, establishing a second-level root knowledge node according to the correspondence of different exhibition halls in the exhibition halls, establishing a sub-knowledge node according to the correspondence of different exhibition halls in each exhibition halls, and planning and forming the first-level knowledge node, the second-level knowledge node and the sub-node according to the professional level design of the museum;
Extracting image features of the scene model of the sub-knowledge node, and performing special visual compatibility labeling adjustment on the sub-knowledge node based on the image features, and labeling the sub-knowledge point;
similar label aggregation is carried out on the sub-knowledge nodes according to the labels, and sub-knowledge nodes with the same or similar labels are associated;
When a scene model is inquired, the first-level root knowledge nodes and the corresponding second-level root knowledge nodes are mutually associated, the second-level root knowledge nodes and the corresponding sub-knowledge nodes are mutually associated, and the first-level root knowledge nodes, the second-level root knowledge nodes and the sub-knowledge nodes of the same kind are mutually associated;
When a user uses an exhibition area scene model, preloading sub-knowledge nodes, secondary root knowledge nodes and scene models corresponding to the primary root knowledge nodes which are associated with the exhibition area scene model, and traversing the sub-knowledge nodes to obtain a next selected scene model according to the primary root knowledge nodes to the secondary root knowledge nodes.
It should be noted that:
The museum is divided into different exhibition halls, the exhibition halls are generally divided based on times, topics, geographic positions, artistic forms, disciplines and the like, different exhibition halls in each exhibition hall are subdivided according to the categories of the exhibition halls, scene models corresponding to different exhibition halls in each exhibition hall are further subdivided on the basis of the categories of the exhibition halls, so that a first-level root knowledge node-second-level root knowledge-child knowledge node tree structure is formed by the exhibition halls-exhibition halls, and a high-dimensional characteristic knowledge node tree is formed by the exhibition halls-exhibition halls structure;
The design and the planning of the professional level of the museum take the music museum as an example, wherein the first-level root knowledge nodes correspond to the music museum, the second-level root knowledge points correspond to exhibition halls such as singing, dancing, rapling, musical instrument playing and the like, and the sub-knowledge nodes are such as national dance exhibition areas and modern dance exhibition areas under the dancing exhibition halls;
The special visual compatibility labeling adjustment is priority adjustment for ensuring that space switching type scene loading also has certain visual consistency and venue uniformity, and specific indexes comprise two aspects of color and geometric structure consistency.
When the scene model is queried, the scene model of a certain exhibition area can be traversed and queried according to the sequence of the exhibition hall and the exhibition area, and the scene model can be directly queried and positioned to a specific exhibition area to be loaded;
Extracting image features of the scene model of the sub-knowledge node, classifying the scene model of the sub-knowledge node based on the image features, and labeling the sub-knowledge node comprises the following steps:
Constructing a multidimensional prediction model, fusing the features of different dimensions of the scene model through the root semantic information of the high-dimensional features and the sub-semantic information of the low-dimensional features, and carrying out feature detection on the fused feature map;
Classifying the scene model of the sub knowledge node based on the detected characteristic data, and labeling according to the classification result;
The multi-dimensional prediction model performs target detection independently in two dimensions, the root semantic information of the high-dimensional features is an original knowledge point node tree of the scene model, and the sub-semantic information of the low-dimensional features comprises extracted visual image features of the scene model.
It should be noted that:
Because a specific scene model of a exhibition area is often not summarized and clear only by one category, that is, a plurality of categories or keywords exist, image feature extraction is required to be performed on the scene model, and the scene model is classified according to the extracted features;
In this embodiment, the scene model class of each exhibition is defined as a specific one, but a plurality of labels are marked, scene models with the same or similar labels are associated according to the labels, meanwhile, exhibition halls and exhibition halls with similar classes are associated with the scene models, and when the scene models are queried, the exhibition halls and exhibition halls with similar classes and the scene models of the exhibition halls with the same or similar labels are preloaded;
performing target detection on the scene model by adopting a convolutional neural network model:
The feature importance degree in the scene model is different, so that the scene model is divided into high-dimensional features and low-dimensional features according to expert experience, wherein root semantic information of the high-dimensional features comprises an original knowledge point node tree of the scene model, sub-semantic information of the low-dimensional features comprises extracted visual image features of the scene model, and simultaneously, two feature images with different sizes are adopted to detect images of the scene model respectively and independently; the classification of the low-dimensional features is more specific, and the detail requirement on feature identification is higher, so that the feature diagram has smaller requirement size and needs smaller receptive field;
The feature map of the low-dimensional feature is set smaller in size and the feature map of the high-dimensional feature is set larger in size;
and the fusion is carried out according to the detection results of the two feature maps, so that the detection accuracy is improved.
The process of associating sub-knowledge nodes with the same or similar labels according to the labels is as follows:
extracting a semantic vector of a current tag, presetting a semantic similarity threshold, and associating the tag with the semantic similarity of the current tag within the threshold with the current tag;
When a scene model is inquired, preloading the scene model corresponding to the sub-knowledge node corresponding to the label associated with the labels according to the inquired labels of the scene model.
It should be noted that: by associating sub-knowledge nodes with the same or similar labels, the accuracy and efficiency of knowledge retrieval are improved, and more accurate knowledge matching is realized. And the scene model of the sub-knowledge node associated with the query tag is preloaded, so that related knowledge can be prepared in advance, the query time and the consumption of computing resources are reduced, the response speed of knowledge retrieval can be improved, and the user experience is improved.
For example, in a music museum, when the size of the venue model is very large, if the overall loading is performed, not only the memory of the browser overflows, but also the loading time is very long, and the user experience is very poor. And the scene model is segmented and loaded as required by using a AssetBundle-based webpage version museum scene loading method. If the user enters the exhibition hall of the museum, the movement track of the user is analyzed and predicted to load the model of the exhibition hall which is possibly going to next, and the step is repeated when the user enters the exhibition hall which is next. Therefore, the pressure on the memory of the browser is reduced, the idle bandwidth is fully utilized, the proportion of the scene loading time in the whole browsing time is reduced, and the user experience is optimized.
The foregoing description of the preferred embodiments of the invention is not intended to be limiting, but rather is intended to cover all modifications, equivalents, alternatives, and improvements that fall within the spirit and scope of the invention.
Claims (9)
1. A method for loading a scene of a webpage version museum based on AssetBundle is characterized by comprising the following steps:
s1, packing museum scene resources, generating AssetBundle files and storing the AssetBundle files into a database;
S2, distinguishing scene loading types according to user selection, performing space transition type scene loading when the user moves across the scene, and performing space switching type scene loading when the user moves across the exhibition area;
S3, when the user loads the space transition type scene, carrying out scene model instantiation according to AssetBundle files, predicting the moving route of the user, and preloading the scene passed by the user according to the moving route of the user;
s4, when the user loads the space switching type scene, the scene model is instantiated according to AssetBundle files, similar scenes in the system knowledge nodes or knowledge field are searched based on the scene model, and the similar scenes are preloaded;
step S4 comprises the steps of:
When a user switches the exhibition areas, the scene resource file of the exhibition area is directly downloaded from the CDN according to the exhibition area selected by the user, decompression and instantiation operations are carried out on terminal equipment, and the scene resource of the selected exhibition area is loaded into the memory;
according to the preset knowledge point node number sequence, a preloading rule of space switching type scene loading is set;
Loading the exhibition halls in the same knowledge point or the same knowledge field as a preloading target according to the preset sequence of knowledge point node trees in the exhibition hall, and carrying out knowledge node association labeling for AssetBundle resource packages when the AssetBundle resource packages of the scene model are uploaded, and inquiring the scene model needing to inquire unified knowledge points according to the knowledge points; for a scene model in the field of related knowledge to be queried, tracing back to an upper-level root knowledge node through a knowledge point node tree, traversing downwards to obtain all sub-knowledge nodes of the root knowledge node, and querying and preloading according to the sub-knowledge nodes;
The preloading operation is performed by using idle bandwidth resources only when the bandwidth is in an idle state.
2. The method for loading a museum scene based on AssetBundle web pages as set forth in claim 1, wherein the step S1 includes the steps of:
collecting and packaging a grid model, materials and a map of a museum scene;
Creating AssetBundle a file using a Unity engine, the AssetBundle file comprising museum scene resources;
and storing the generated AssetBundle files in a database.
3. The method for loading a museum scene based on AssetBundle web pages as set forth in claim 1, wherein the step S2 includes the steps of:
Distinguishing the user movement type according to the operation of the user movement scene, loading resources according to the user movement type, and loading a space transition scene by the user when the user switches between adjacent exhibition halls; when a user switches between different exhibition areas, the user loads a space switching type scene;
the step of loading the space transition type scene is as follows:
detecting a movement intention of a user;
loading AssetBundle resource packages of the next scene according to the mobile intention of the user;
and preloading the exhibition hall scene to be browsed by the user according to AssetBundle resource packages of the next scene.
4. The method for loading a museum scene based on AssetBundle web pages as set forth in claim 3, wherein the process of detecting the movement intention of the user is:
Forming a motion vector array through the triggering sequence of the multi-layer triggers, and performing unified calculation to obtain a final movement intention vector;
judging the direction and length of the movement intention vector, and determining the specific movement intention of the user;
when the movement intention vector is in the positive direction and the length reaches the requirement, the movement intention of the user is judged to be forward, when the movement intention vector is in the opposite direction, the movement intention of the user is judged to be backward, and when the movement intention vector is not in the sufficient length, the movement intention of the user is judged to be backward.
5. The method for loading a scene of a web-version museum based on AssetBundle as set forth in claim 3, wherein the process of loading AssetBundle resource packages of a next scene according to the mobile intention of the user is as follows:
And loading a scene AssetBundle resource package of the next exhibition hall according to the movement intention determined by the user, and loading, decompressing and instantiating the selected AssetBundle resource package to finish loading the scene to be loaded.
6. The method for loading a museum scene based on AssetBundle web pages as set forth in claim 3, wherein the process of preloading the museum scene to be browsed by the user according to AssetBundle resource package of the next scene is as follows:
When the scene loading used by the user is completed, the current use requirement of the user is met, the program downloading queue is empty, the bandwidth is in an idle state, the idle bandwidth resource is utilized to predict and calculate the moving direction of the next step of the user, and the exhibition hall which the user goes to next step is preloaded;
predicting the next moving direction of a calculated user by using a pedestrian track tracking technology, calculating a exhibition hall scene covered by a user route, loading a scene model of the exhibition hall which is predicted to be covered according to a prediction result obtained by calculation, and loading adjacent scenes of the current scene by referring to a spatial locality principle in a computer composition principle when the bandwidth is still idle;
When the scene model of the next exhibition hall meets the inequality equation set or the Rect function calculation, judging that the scene model of the exhibition hall is covered by the user route, wherein the inequality equation set is as follows:
Wherein { D Path point } is the set of path points for the next movement of the user, r Exhibition hall is the distance from the current scene of the user to the scene model of the next exhibition hall, and l Exhibition hall is the distance from the center point of the scene model of the next exhibition hall to the edge of the scene model;
the Rect function judges whether a scene model of the exhibition hall is covered by a user route according to the position and the size of the exhibition hall;
In the process of loading the space transition type scene, the downloading and memory optimization of the resources are as follows:
and accelerating the downloading of AssetBundle resource packages to terminal equipment through the CDN, caching the resource packages in a IndexDB database, monitoring the memory of the database by using a front-end performance monitoring function, and recycling garbage according to the sequence of the queues when the occupied memory exceeds a preset threshold.
7. The method for loading a museum scene based on AssetBundle web pages as set forth in claim 1, wherein the construction process of the knowledge point node tree is as follows:
Establishing a first-level root knowledge node according to the correspondence of the exhibition halls, establishing a second-level root knowledge node according to the correspondence of different exhibition halls in the exhibition halls, establishing a sub-knowledge node according to the correspondence of different exhibition halls in each exhibition halls, and forming a knowledge point node tree by the first-level knowledge node, the second-level knowledge node and the sub-node according to the design and the planning of the professional level of the museum;
Extracting image features of the scene model of the sub-knowledge node, and carrying out special visual compatibility labeling adjustment on the sub-knowledge node based on the image features;
similar label aggregation is carried out on the sub-knowledge nodes according to the labels, and sub-knowledge nodes with the same or similar labels are associated;
When a scene model is inquired, the first-level root knowledge nodes and the corresponding second-level root knowledge nodes are mutually associated, the second-level root knowledge nodes and the corresponding sub-knowledge nodes are mutually associated, and the first-level root knowledge nodes, the second-level root knowledge nodes and the sub-knowledge nodes of the same kind are mutually associated;
when a user uses an exhibition hall scene model, preloading sub-knowledge nodes, secondary root knowledge nodes and scene models corresponding to the primary root knowledge nodes which are associated with the exhibition hall scene model, and traversing the sub-knowledge nodes to obtain a next selected scene model according to the primary root knowledge nodes to the secondary root knowledge nodes.
8. The method for loading a scene of a web-version museum based on AssetBundle of claim 7, wherein the steps of extracting image features from the scene model of the child knowledge node, classifying the scene model of the child knowledge node based on the image features, and labeling the child knowledge node include:
Constructing a multidimensional prediction model, fusing the features of different dimensions of the scene model through the root semantic information of the high-dimensional features and the sub-semantic information of the low-dimensional features, and carrying out feature detection on the fused feature map;
Classifying the scene model of the sub knowledge node based on the detected characteristic data, and labeling according to the classification result;
The multi-dimensional prediction model performs target detection independently in two dimensions, the root semantic information of the high-dimensional features is an original knowledge point node tree of the scene model, and the sub-semantic information of the low-dimensional features comprises extracted visual image features of the scene model.
9. The method for loading a scene of a museum based on AssetBundle web pages as set forth in claim 8, wherein the process of associating child knowledge nodes with the same or similar labels according to the labels is as follows:
extracting a semantic vector of a current tag, presetting a semantic similarity threshold, and associating the tag with the semantic similarity of the current tag within the threshold with the current tag;
When a scene model is inquired, preloading the scene model corresponding to the sub-knowledge node corresponding to the label associated with the labels according to the inquired labels of the scene model.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311337077.3A CN117331622B (en) | 2023-10-16 | 2023-10-16 | AssetBundle-based webpage version museum scene loading method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311337077.3A CN117331622B (en) | 2023-10-16 | 2023-10-16 | AssetBundle-based webpage version museum scene loading method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117331622A CN117331622A (en) | 2024-01-02 |
| CN117331622B true CN117331622B (en) | 2024-05-17 |
Family
ID=89275230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311337077.3A Active CN117331622B (en) | 2023-10-16 | 2023-10-16 | AssetBundle-based webpage version museum scene loading method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117331622B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104102545A (en) * | 2014-07-04 | 2014-10-15 | 北京理工大学 | Three-dimensional resource allocation and loading optimization method for mobile augmented reality browser |
| CN106052696A (en) * | 2016-08-17 | 2016-10-26 | 西安理工大学 | Museum real-time route guidance method based on mobile terminals |
| CN107122201A (en) * | 2017-03-10 | 2017-09-01 | 腾讯科技（深圳）有限公司 | Resource loading, the generation method of resource file and device |
| CN111729305A (en) * | 2020-06-23 | 2020-10-02 | 网易（杭州）网络有限公司 | Map scene preloading method, model training method, device and storage medium |
| CN115373761A (en) * | 2022-09-20 | 2022-11-22 | 江苏赞奇科技股份有限公司 | AI-based dynamic loading method for scene resources in real-time cloud rendering |
| CN116012523A (en) * | 2022-12-05 | 2023-04-25 | 北方华录文化科技（北京）有限公司 | Scene loading method, system and storage medium based on virtual three-dimensional exhibition hall |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US11397462B2 (en) * | 2012-09-28 | 2022-07-26 | Sri International | Real-time human-machine collaboration using big data driven augmented reality technologies |
-
2023
- 2023-10-16 CN CN202311337077.3A patent/CN117331622B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104102545A (en) * | 2014-07-04 | 2014-10-15 | 北京理工大学 | Three-dimensional resource allocation and loading optimization method for mobile augmented reality browser |
| CN106052696A (en) * | 2016-08-17 | 2016-10-26 | 西安理工大学 | Museum real-time route guidance method based on mobile terminals |
| CN107122201A (en) * | 2017-03-10 | 2017-09-01 | 腾讯科技（深圳）有限公司 | Resource loading, the generation method of resource file and device |
| CN111729305A (en) * | 2020-06-23 | 2020-10-02 | 网易（杭州）网络有限公司 | Map scene preloading method, model training method, device and storage medium |
| CN115373761A (en) * | 2022-09-20 | 2022-11-22 | 江苏赞奇科技股份有限公司 | AI-based dynamic loading method for scene resources in real-time cloud rendering |
| CN116012523A (en) * | 2022-12-05 | 2023-04-25 | 北方华录文化科技（北京）有限公司 | Scene loading method, system and storage medium based on virtual three-dimensional exhibition hall |
Non-Patent Citations (1)
| Title |
|---|
| Unity3D环境下虚拟现实场景元素加载策略;黄玮;《景德镇学院学报》;20191231;第34卷(第6期);第5-9页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117331622A (en) | 2024-01-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107256262B (en) | Image retrieval method based on object detection | |
| CN103345645B (en) | Commodity image class prediction method towards net purchase platform | |
| CN102508923B (en) | Automatic video annotation method based on automatic classification and keyword marking | |
| US20190361994A1 (en) | Compositing Aware Digital Image Search | |
| CN106097095B (en) | Determine the method and device of credit | |
| CN111241345A (en) | Video retrieval method and device, electronic equipment and storage medium | |
| CN105117429A (en) | Scenario image annotation method based on active learning and multi-label multi-instance learning | |
| CN107430604A (en) | The semantic expressiveness of picture material | |
| CN110796204A (en) | Video tag determination method and device and server | |
| CN102236714A (en) | Extensible markup language (XML)-based interactive application multimedia information retrieval method | |
| CN113254711A (en) | Interactive image display method and device, computer equipment and storage medium | |
| CN100462978C (en) | Image searching method and system | |
| CN112989169A (en) | Target object identification method, information recommendation method, device, equipment and medium | |
| CN112836487A (en) | Automatic comment method and device, computer equipment and storage medium | |
| CN102819752B (en) | System and method for outdoor large-scale object recognition based on distributed inverted files | |
| Zhuo et al. | Research on personalized image retrieval technology of video stream big data management model | |
| CN117331622B (en) | AssetBundle-based webpage version museum scene loading method | |
| Chen et al. | Research on object detection algorithm based on improved Yolov5 | |
| Yang et al. | Explorations on visual localization from active to passive | |
| CN115455249A (en) | Double-engine driven multi-modal data retrieval method, equipment and system | |
| CN108415950A (en) | A kind of hypernym polymerization and device | |
| Liu et al. | Extended cone-curvature based salient points detection and 3D model retrieval | |
| CN114648762A (en) | Semantic segmentation method and device, electronic equipment and computer-readable storage medium | |
| Liu et al. | Multimodal Wireless Situational Awareness‐Based Tourism Service Scene | |
| Ning et al. | Deep Learning on 3D Point Cloud for Semantic Segmentation |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB02 | Change of applicant information |
Country or region after: China Address after: Unit 202, 3rd Floor, Building 9, West District, No. 9 Linglong Road, Haidian District, Beijing, 100000 RMB Applicant after: Zhongjiao Changxiang Technology Co.,Ltd. Address before: No. 55, Building 119, No. 12 Zhongguancun South Street, Haidian District, Beijing, 100000 Applicant before: Chinese Education Changxiang (Beijing) Technology Co.,Ltd. Country or region before: China |
|
| CB02 | Change of applicant information | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |