CN114185431A - Intelligent media interaction method based on MR technology - Google Patents
Intelligent media interaction method based on MR technologyDownload PDFInfo
- Publication number
- CN114185431A CN114185431ACN202111398973.1ACN202111398973ACN114185431ACN 114185431 ACN114185431 ACN 114185431ACN 202111398973 ACN202111398973 ACN 202111398973ACN 114185431 ACN114185431 ACN 114185431A
- Authority
- CN
- China
- Prior art keywords
- image
- real
- time
- intelligent terminal
- intelligent
- 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.)
- Granted
Links
Images
Classifications
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F3/00—Input arrangements for transferring data to be processed into a form capable of being handled by the computer; Output arrangements for transferring data from processing unit to output unit, e.g. interface arrangements
- G06F3/01—Input arrangements or combined input and output arrangements for interaction between user and computer
- G06F3/011—Arrangements for interaction with the human body, e.g. for user immersion in virtual reality
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/22—Matching criteria, e.g. proximity measures
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
- G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2203/00—Indexing scheme relating to G06F3/00 - G06F3/048
- G06F2203/01—Indexing scheme relating to G06F3/01
- G06F2203/012—Walk-in-place systems for allowing a user to walk in a virtual environment while constraining him to a given position in the physical environment
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Evolutionary Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Human Computer Interaction (AREA)
- Processing Or Creating Images (AREA)
Abstract
Description
Technical Field
The invention relates to an intelligent media interaction method based on an MR (magnetic resonance) technology, belonging to the technical field of media interaction.
Background
MR technology refers to mixed reality technology. Mixed Reality (MR) is a further development of virtual reality technology that builds an interactive feedback information loop between the real world, the virtual world and the user by presenting virtual scene information in the real scene to enhance the realism of the user experience.
Mixed reality is a combination of technologies that provide not only new viewing methods but also new input methods, and all methods are combined with each other, thereby promoting innovation. The combination of input and output is a key differentiation advantage for small and medium-sized enterprises. Therefore, mixed reality can directly influence your workflow and help your staff to improve work efficiency and innovation capability. Let us look at some feasible schemes to understand the working principle and what benefits it has.
Mixed reality (MR, including both augmented reality and augmented virtual) refers to a new visualization environment created by the merging of real and virtual worlds. Physical and digital objects coexist in the new visualization environment and interact in real time. The system generally employs three main features:
1. it combines virtual and reality; 2. in virtual three dimensions (3D registration); 3. and (4) running in real time.
Mixed Reality (MR) implementations need to be in an environment where real-world objects can interact with each other. It is the domain of VR if everything is virtual. If the displayed virtual information can only be simply superposed on the real things, the AR is the virtual reality (AR). The key point of MR is the interaction with the real world and the timely acquisition of information.
Currently, representative devices of MR technology, MR glasses, whose classical applications are shared MR glasses used in scenic spots, are accepted by an increasing number of scenic spots and visitors as wearing devices for MR applications. However, many disadvantages are reflected in the practical application of the existing equipment. Such as: the degree of intellectualization of part of MR glasses in a virtual scene is not high, and corresponding operations such as clicking, selecting and the like need to be carried out by using a handheld remote controller in the actual operation process; and implementing the operation on the desktop type graphic menu in an auxiliary voice input mode.
In the triggering method, the MR glasses are required to be connected with a peripheral through the application of the handheld remote controller or the pressure gloves, the man-machine interaction degree is limited, and the dynamic interesting effect is limited. On the other hand, the voice input corresponding command is limited by the existing voice input technology, especially the technical barrier of the voice input accuracy is still not good.
The above two points limit the development of the existing MR technology in the media interaction field.
Disclosure of Invention
Aiming at the defects in the prior art, the invention provides an intelligent media interaction method based on an MR technology, which has the following specific technical scheme:
an intelligent media interaction method based on MR technology comprises the following steps:
step S1, recognizing a front scene by using a position recognition module in the MR intelligent terminal and calculating to obtain a real-time geographic position coordinate and a real-time marking image corresponding to the real-time geographic position coordinate;
step S2, identifying whether the current motion state of the user is suitable for pushing appointed information by using a motion state module in the MR intelligent terminal;
s3, recognizing the real-time geographic position coordinates and the real-time marking images and the specimen images stored in the local database in the MR intelligent terminal by using an AI algorithm in the MR intelligent terminal, and screening information to be played from the designated information when the recognition result is in accordance with the preset; when the motion state module identifies that the current motion state of the user is suitable for pushing the information to be played, the display area of the MR intelligent terminal can display the information to be played.
As an improvement of the above technical solution, the operating method of the location identification module includes:
the position recognition module photographs a front scene to generate a real-time image, compares the real-time image with the similarity of the sample image stored in a local database in the MR intelligent terminal, selects the sample image with the maximum similarity with the real-time image and marks the sample image as a module image; the specimen image and the module image both carry a position calibration point and a rectangular calibration pattern block taking the position calibration point as a center, the real-time image is subjected to gridding processing to obtain a gridding pattern block set, the calibration pattern block and the rectangular gridding pattern block in the gridding pattern block set are subjected to traversal comparison, and a comparison algorithm adopts a similarity algorithm; when the similarity between the calibration pattern block and the rectangular gridding pattern block in the gridding pattern block set is larger than or equal to a preset threshold value, the position information of the position calibration point load corresponding to the calibration pattern block is a real-time geographic position coordinate, and the real-time marking image is a rectangular picture which is intercepted at the position corresponding to the real-time image by taking the real-time geographic position coordinate as the center.
As an improvement of the above technical solution, the operating method of the motion state module in step S2 includes:
and recording the real-time motion state data set of the user by utilizing an acceleration sensor and a gyroscope which are installed on the MR intelligent terminal, and comparing the real-time motion state data set with a preset data set in a local database in the MR intelligent terminal so as to judge the corresponding motion state.
As an improvement of the above technical solution, the preset threshold is obtained by performing an AI algorithm, and the method includes:
the template wall is used for assisting training and is formed by arraying a plurality of rectangular module units, each module unit comprises a transparent box, a thermochromic layer is arranged in each transparent box, the front surface of each thermochromic layer is marked with a position mark, the back surface of each thermochromic layer is fixedly provided with an electric heating plate, and a heat insulation layer is filled between the edge of each electric heating plate and the inner wall of each transparent box; heating by using an electric heating plate to change the color of the thermochromic layer so as to change the pattern displayed by the template wall, photographing the template wall by using an MR intelligent terminal to obtain a first image set, and performing gray processing on each first image in the first image set to obtain a second image set;
constructing and training a twin neural network, wherein the twin neural network comprises two symmetrically arranged Convolutional Neural Networks (CNN), and two symmetrically arranged hidden Markov models are introduced into a generator of the Convolutional Neural Networks (CNN) for countertraining; inputting the second image set into a twin neural network, utilizing a defect random generation module to perform defect processing on the second image in the second image set to obtain a third image, receiving the third image by a generator of a Convolutional Neural Network (CNN), extracting a feature image, performing similarity comparison on the feature image and the real-time mark image, performing counterstudy, and regressing a preset threshold value according to a counterstudy result.
As an improvement of the technical scheme, the MR intelligent terminal comprises MR intelligent glasses and an MR intelligent helmet.
As an improvement of the technical scheme, the transparent box is made of a glass material.
As an improvement of the above technical solution, the thermochromic layer is made of a reversible thermochromic material.
As an improvement of the technical scheme, the heat insulation layer is made of heat insulation epoxy resin through filling and curing.
As an improvement of the technical scheme, the position mark is one of a cross shape, a meter shape and an shape.
The invention has the beneficial effects that:
the invention utilizes MR technology to improve the interaction mode with intelligent media, and provides more diversified information for users; according to the invention, the image recognition technology is used for triggering the push content appointed by the merchant, and the deep learning algorithm is used for improving the defects of the existing image recognition technology, so that the triggering accuracy and flexibility are improved, and the human-computer interaction effect is more friendly and convenient; the conditions that the triggering accuracy is low and accurate triggering cannot be achieved due to objective factors such as seasons, temperature, illumination and shooting angles are reduced, and the user experience is remarkably improved.
Drawings
FIG. 1 is a schematic view of the construction of the formwork wall of the present invention;
fig. 2 is a schematic structural diagram of the module unit according to the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the present invention more apparent, the present invention is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the invention and are not intended to limit the invention.
Example 1
The intelligent media interaction method based on the MR technology comprises the following steps:
step S1, recognizing a front scene by using a position recognition module in the MR intelligent terminal and calculating to obtain a real-time geographic position coordinate and a real-time marking image corresponding to the real-time geographic position coordinate;
step S2, identifying whether the current motion state of the user is suitable for pushing appointed information by using a motion state module in the MR intelligent terminal;
s3, recognizing the real-time geographic position coordinates and the real-time marking images and the specimen images stored in the local database in the MR intelligent terminal by using an AI algorithm in the MR intelligent terminal, and screening information to be played from the designated information when the recognition result is in accordance with the preset; when the motion state module identifies that the current motion state of the user is suitable for pushing the information to be played, the display area of the MR intelligent terminal can display the information to be played.
Wherein, MR intelligent terminal includes MR intelligent glasses, MR intelligence helmet. In this embodiment, the MR smart terminal is preferably MR smart glasses.
Example 2
Based on theembodiment 1, the working method of the position identification module comprises the following steps:
the position recognition module photographs a front scene to generate a real-time image, compares the real-time image with the similarity of the sample image stored in a local database in the MR intelligent terminal, selects the sample image with the maximum similarity with the real-time image and marks the sample image as a module image; the specimen image and the module image both carry a position calibration point and a rectangular calibration pattern block taking the position calibration point as a center, the real-time image is subjected to gridding processing to obtain a gridding pattern block set, the calibration pattern block and the rectangular gridding pattern block in the gridding pattern block set are subjected to traversal comparison, and a comparison algorithm adopts a similarity algorithm; when the similarity between the calibration pattern block and the rectangular gridding pattern block in the gridding pattern block set is larger than or equal to a preset threshold value, the position information of the position calibration point load corresponding to the calibration pattern block is a real-time geographic position coordinate, and the real-time marking image is a rectangular picture which is intercepted at the position corresponding to the real-time image by taking the real-time geographic position coordinate as the center.
The reason for this is that in practical application scenes, for example, scenes in a scenic spot are affected by factors such as seasons, temperature, illumination, shooting angles, and the like, thereby causing great difficulty in comparison between the real-time image and the specimen image. The specimen image is only a picture taken by a merchant in advance, for example, the merchant uses MR smart glasses to push corresponding MR content, such as video, images, and the like, to a user. Based on commercial needs, when a user wears the MR smart glasses at a specified position and needs to trigger advertisements preset by the merchant in time, if the user is in a common commercial city, a sensor can be installed at the specified position of the commercial city in the conventional technology, and when the MR smart glasses move to the position near the sensor, an advertisement push command can be triggered; this technique is too limited, limited to the field, and very inflexible.
In an application scene of a scenic spot, when a user moves to a specified position, such as a position near a yellow mountain guest-welcoming pine tree, a merchant has a requirement that when the user is within a range of 20-100 meters near the yellow mountain guest-welcoming pine tree, an advertisement of the 'guest-welcoming pine' card cigarette can be triggered to be pushed to the user for 1-3 times. Since the installation of the corresponding sensor is not allowed near the yellow mountain guest-welcoming pine, the technique of installing the sensor is not practical. At present, a mode of shooting a picture of a yellow mountain welcoming pine tree in real time is also adopted, the picture is compared with a picture shot in advance in a database, and when the similarity reaches a certain degree, an advertisement push command can be triggered; this technique is theoretically very flexible. However, there are several problems:
firstly, the scenery of the scenic spot can be affected by factors such as season, temperature, illumination and shooting angle, so that great difficulty is caused in comparison between the real-time image and the specimen image.
Secondly, because the number of the specimen images is limited, how can the similarity between the real-time image and the specimen image be grasped to satisfy the trigger advertisement push command and ensure the accuracy of identification?
In the embodiment, the sample image with the maximum similarity to the real-time image is found and is calibrated as the module image, so that the most similar sample image can be found even if a large defect exists between the real-time image and the sample image due to the influences of factors such as seasons, air temperature, illumination, shooting angles and the like, and the rest images are excluded, so that the sample amount of calculation is reduced to the maximum extent. When the calibration image block is rectangular, the calibration image block is matched with the subsequent gridding technology. When the similarity between the calibration pattern block and the rectangular gridding pattern block in the gridding pattern block set is greater than or equal to a preset threshold value, the result shows that a 'preset point' with a merchant is found, because a gridding technology is adopted, the position information (different positions and image contents are different, so different images can represent different position information and are equivalent to a coordinate point) of the position calibration point load corresponding to the calibration pattern block is a real-time geographic position coordinate, and the real-time marking image is convenient for subsequent deep learning and checking and is used for checking the accuracy of an algorithm.
Therefore, in the embodiment, even if the use environment is affected by factors such as seasons, air temperature, illumination, shooting angle and the like, the comparison between the real-time image and the specimen image is not affected, so that the algorithm provided by the invention is remarkably improved, the application limitation is small, and the application range is wider.
Example 3
Based on embodiment 2, the working method of the motion state module in step S2 includes:
and recording the real-time motion state data set of the user by utilizing an acceleration sensor and a gyroscope which are installed on the MR intelligent terminal, and comparing the real-time motion state data set with a preset data set in a local database in the MR intelligent terminal so as to judge the corresponding motion state.
Application scenarios: when an acceleration sensor and a gyroscope of the MR intelligent terminal record a real-time motion state data set of a user, if a common male walks, the acceleration sensor and the gyroscope collect corresponding data, the data is compared with a standard database when the common male walks, if the data is in accordance with the standard database, the user can be judged to be in a walking state, and in the state, an advertisement is not recommended to be pushed to the user. If in the leisure and rest state, the advertisement can be properly pushed to the user.
Example 4
Based on embodiment 3, the preset threshold is obtained by performing an AI algorithm, and the method includes:
the template wall 10 is used for auxiliary training, as shown in fig. 1 and 2, the template wall 10 is formed by arranging a plurality of rectangular module units 11 in an array, each module unit 11 comprises atransparent box 111, athermochromic layer 112 is arranged inside eachtransparent box 111, a position mark is marked on the front surface of eachthermochromic layer 112, anelectric heating plate 113 is fixedly installed on the back surface of eachthermochromic layer 112, and aheat insulation layer 114 is filled between the edge of eachelectric heating plate 113 and the inner wall of eachtransparent box 111; the color of thethermochromic layer 112 is changed by heating with theelectric heating plate 113 to change the pattern displayed on the template wall 10, the template wall 10 is photographed by the MR smart terminal to obtain a first image set (set of first images), and each of the first images in the first image set is subjected to grayscale processing to obtain a second image set (set of second images).
Constructing and training a twin neural network, wherein the twin neural network comprises two symmetrically arranged Convolutional Neural Networks (CNN), and two symmetrically arranged hidden Markov models are introduced into a generator of the Convolutional Neural Networks (CNN) for countertraining; inputting the second image set into a twin neural network, utilizing a defect random generation module to perform defect processing on the second image in the second image set to obtain a third image, receiving the third image by a generator of a Convolutional Neural Network (CNN), extracting a feature image, performing similarity comparison on the feature image and the real-time mark image, performing counterstudy, and regressing a preset threshold value according to a counterstudy result.
And the defect random generation module performs defect processing (such as cropping) on the second image in the second image set by using a random forest mode to obtain a third image.
In the invention, the preset threshold value is not set by designers at will, and is obtained by deep learning algorithm operation according to a large number of samples, so that the method is more practical and more accurate. The accuracy of the preset threshold value is poor, and the direct consequence is that the humanization degree is very poor, so that the user experience feeling is reduced linearly.
In this embodiment, the template wall 10 is used for auxiliary training, and different module units 11 are controlled to change color into corresponding color images, so that the module unit 11 with a rectangular structure is equivalent to a "pixel color block", and finally, different color patterns are displayed on the template wall 10. The mode of randomly generating different patterns is a solid pattern, so that a scene real object image can be simulated. Compared with the electronic display screen for displaying images, the images displayed by the electronic display screen are flatter and more two-dimensional, the shooting effect is extremely poor, and the burden of subsequent image processing can be increased.
By using the pattern wall 10 to randomly generate different patterns, image defects caused by factors such as season, temperature, illumination, and shooting angle can be simulated, thereby generating corresponding samples.
In the embodiment, deep learning is performed by constructing a twin neural network, and the two symmetrically arranged convolutional neural networks CNN can perform cross validation, so that the sample size processing rate is remarkably improved; the double random characteristics of the hidden Markov models are utilized, and then the two hidden Markov models are symmetrically arranged, so that 2 can be generated for the same imagenThe defect samples can be subjected to cross comparison and regression for checking in countercheck learning, and accuracy of the regression preset threshold value is improved. Therefore, two symmetrically arranged hidden markov models are introduced into the two symmetrically arranged convolutional neural networks CNN for countertraining, which also helps to control the robustness of the system.In the deep learning process, a large number of samples do not need to be shot, the problem of small sample amount in the practical application process can be solved, the problem of low accuracy caused by small sample amount in the conventional deep learning process is solved, and the phenomenon of overfitting is small.
That is to say, in this embodiment, even the quantity of sample image is limited, can guarantee to trigger the accuracy and the promptness of advertisement propelling movement order to show improvement user experience and feel, avoid taking place the condition emergence of awkwardness such as spurious triggering, not triggering.
Example 5
Based on embodiment 4, thetransparent case 111 is made of a glass material. Thethermochromic layer 112 is made of a reversible thermochromic material. Thethermal insulation layer 114 is made of thermal insulation epoxy resin through filling and curing; the provision of thethermal insulation layer 114 can prevent the adjacent two module units 11 from interfering with each other.
The position mark is one of a cross, a meter, and an , and in this embodiment, a cross is preferable. The position mark must have a 'cross point', so that the accuracy of image identification is facilitated; the structure of the position mark cannot be too complex, which is convenient for image recognition.
Example 6
In the invention, the interaction method with the media further comprises the step of setting an advertisement platform for receiving the advertisement delivery of the merchant, wherein the merchant can set advertisement contents such as introduction contents, coupons, discount information and the like on the advertisement platform.
The interaction method between the user and the media can also be as follows: eye movement interaction, gesture recognition interaction, voiced/unvoiced speech recognition, head movement interaction, etc. When the hand makes a specific motion, receiving motion information of the hand of the user, or identifying a specific shape of the hand, and matching the specific motion with control options such as up-down, left-right sliding, zooming-in, zooming-out, clicking, closing and the like, so as to control related contents to perform corresponding motions. The accuracy of recognizing voiced/unvoiced is higher compared to speech input.
In the embodiment, the MR technology is used for improving the interaction mode with the intelligent media, so that more diversified information is provided for the user; according to the invention, the image recognition technology is used for triggering the push content appointed by the merchant, and the deep learning algorithm is used for improving the defects of the existing image recognition technology, so that the triggering accuracy and flexibility are improved, and the human-computer interaction effect is more friendly and convenient; the conditions that the triggering accuracy is low and accurate triggering cannot be achieved due to objective factors such as seasons, temperature, illumination and shooting angles are reduced, and the user experience is remarkably improved.
The above description is only for the purpose of illustrating the preferred embodiments of the present invention and is not to be construed as limiting the invention, and any modifications, equivalents and improvements made within the spirit and principle of the present invention are intended to be included within the scope of the present invention.
Claims (9)
1. An intelligent media interaction method based on MR technology is characterized by comprising the following steps:
step S1, recognizing a front scene by using a position recognition module in the MR intelligent terminal and calculating to obtain a real-time geographic position coordinate and a real-time marking image corresponding to the real-time geographic position coordinate;
step S2, identifying whether the current motion state of the user is suitable for pushing appointed information by using a motion state module in the MR intelligent terminal;
s3, recognizing the real-time geographic position coordinates and the real-time marking images and the specimen images stored in the local database in the MR intelligent terminal by using an AI algorithm in the MR intelligent terminal, and screening information to be played from the designated information when the recognition result is in accordance with the preset; when the motion state module identifies that the current motion state of the user is suitable for pushing the information to be played, the display area of the MR intelligent terminal can display the information to be played.
2. The intelligent media interaction method based on MR technology as claimed in claim 1, wherein the working method of the location identification module comprises:
the position recognition module photographs a front scene to generate a real-time image, compares the real-time image with the similarity of the sample image stored in a local database in the MR intelligent terminal, selects the sample image with the maximum similarity with the real-time image and marks the sample image as a module image; the specimen image and the module image both carry a position calibration point and a rectangular calibration pattern block taking the position calibration point as a center, the real-time image is subjected to gridding processing to obtain a gridding pattern block set, the calibration pattern block and the rectangular gridding pattern block in the gridding pattern block set are subjected to traversal comparison, and a comparison algorithm adopts a similarity algorithm; when the similarity between the calibration pattern block and the rectangular gridding pattern block in the gridding pattern block set is larger than or equal to a preset threshold value, the position information of the position calibration point load corresponding to the calibration pattern block is a real-time geographic position coordinate, and the real-time marking image is a rectangular picture which is intercepted at the position corresponding to the real-time image by taking the real-time geographic position coordinate as the center.
3. The MR-technology-based smart media interaction method as claimed in claim 1, wherein the operation method of the motion state module in step S2 includes:
and recording the real-time motion state data set of the user by utilizing an acceleration sensor and a gyroscope which are installed on the MR intelligent terminal, and comparing the real-time motion state data set with a preset data set in a local database in the MR intelligent terminal so as to judge the corresponding motion state.
4. An intelligent media interaction method based on MR technology as claimed in claim 3, wherein the preset threshold is obtained by using AI algorithm, and the method comprises:
the template wall is used for assisting training and is formed by arraying a plurality of rectangular module units, each module unit comprises a transparent box, a thermochromic layer is arranged in each transparent box, the front surface of each thermochromic layer is marked with a position mark, the back surface of each thermochromic layer is fixedly provided with an electric heating plate, and a heat insulation layer is filled between the edge of each electric heating plate and the inner wall of each transparent box; heating by using an electric heating plate to change the color of the thermochromic layer so as to change the pattern displayed by the template wall, photographing the template wall by using an MR intelligent terminal to obtain a first image set, and performing gray processing on each first image in the first image set to obtain a second image set;
constructing and training a twin neural network, wherein the twin neural network comprises two symmetrically arranged Convolutional Neural Networks (CNN), and two symmetrically arranged hidden Markov models are introduced into a generator of the Convolutional Neural Networks (CNN) for countertraining; inputting the second image set into a twin neural network, utilizing a defect random generation module to perform defect processing on the second image in the second image set to obtain a third image, receiving the third image by a generator of a Convolutional Neural Network (CNN), extracting a feature image, performing similarity comparison on the feature image and the real-time mark image, performing counterstudy, and regressing a preset threshold value according to a counterstudy result.
5. The intelligent media interaction method based on MR technology as claimed in claim 1, wherein: the MR intelligent terminal comprises MR intelligent glasses and an MR intelligent helmet.
6. The intelligent media interaction method based on MR technology as claimed in claim 4, wherein: the transparent box is made of a glass material.
7. The intelligent media interaction method based on MR technology as claimed in claim 4, wherein: the thermochromic layer is made of a reversible thermochromic material.
8. The intelligent media interaction method based on MR technology as claimed in claim 4, wherein: the heat insulation layer is made by filling and curing heat insulation epoxy resin.
9. The intelligent media interaction method based on MR technology as claimed in claim 4, wherein: the position mark is one of a cross shape, a meter shape and an shape.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111398973.1ACN114185431B (en) | 2021-11-24 | 2021-11-24 | Intelligent media interaction method based on MR technology |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111398973.1ACN114185431B (en) | 2021-11-24 | 2021-11-24 | Intelligent media interaction method based on MR technology |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114185431Atrue CN114185431A (en) | 2022-03-15 |
| CN114185431B CN114185431B (en) | 2024-04-02 |
Family
ID=80541346
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111398973.1AActiveCN114185431B (en) | 2021-11-24 | 2021-11-24 | Intelligent media interaction method based on MR technology |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114185431B (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114816057A (en)* | 2022-04-20 | 2022-07-29 | 广州瀚信通信科技股份有限公司 | Somatosensory intelligent terminal interaction method, device, equipment and storage medium |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015207168A (en)* | 2014-04-21 | 2015-11-19 | Kddi株式会社 | Information presentation system, method, and program |
| CN109344699A (en)* | 2018-08-22 | 2019-02-15 | 天津科技大学 | Disease identification method of winter jujube based on hierarchical deep convolutional neural network |
| CN111741287A (en)* | 2020-07-10 | 2020-10-02 | 南京新研协同定位导航研究院有限公司 | A method for MR glasses to trigger content using location information |
| CN112084849A (en)* | 2020-07-31 | 2020-12-15 | 华为技术有限公司 | Image recognition method and device |
| CN112181152A (en)* | 2020-11-13 | 2021-01-05 | 幻蝎科技(武汉)有限公司 | Advertisement push management method, equipment and application based on MR glasses |
| US20210303854A1 (en)* | 2020-03-26 | 2021-09-30 | Varjo Technologies Oy | Imaging System and Method for Producing Images with Virtually-Superimposed Functional Elements |
- 2021
- 2021-11-24CNCN202111398973.1Apatent/CN114185431B/enactiveActive
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2015207168A (en)* | 2014-04-21 | 2015-11-19 | Kddi株式会社 | Information presentation system, method, and program |
| CN109344699A (en)* | 2018-08-22 | 2019-02-15 | 天津科技大学 | Disease identification method of winter jujube based on hierarchical deep convolutional neural network |
| US20210303854A1 (en)* | 2020-03-26 | 2021-09-30 | Varjo Technologies Oy | Imaging System and Method for Producing Images with Virtually-Superimposed Functional Elements |
| CN111741287A (en)* | 2020-07-10 | 2020-10-02 | 南京新研协同定位导航研究院有限公司 | A method for MR glasses to trigger content using location information |
| CN112084849A (en)* | 2020-07-31 | 2020-12-15 | 华为技术有限公司 | Image recognition method and device |
| CN112181152A (en)* | 2020-11-13 | 2021-01-05 | 幻蝎科技(武汉)有限公司 | Advertisement push management method, equipment and application based on MR glasses |
Non-Patent Citations (1)
| Title |
|---|
| 丁德菊: "基于混合现实的人机交互系统设计", 《西部广播电视》, pages 237 - 238* |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114816057A (en)* | 2022-04-20 | 2022-07-29 | 广州瀚信通信科技股份有限公司 | Somatosensory intelligent terminal interaction method, device, equipment and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114185431B (en) | 2024-04-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Li et al. | Layoutgan: Synthesizing graphic layouts with vector-wireframe adversarial networks | |
| CN107045844B (en) | A Landscape Guide Method Based on Augmented Reality Technology | |
| CN100407798C (en) | 3D geometric modeling system and method | |
| CN110310175A (en) | System and method for mobile augmented reality | |
| CN110503074A (en) | Information labeling method, apparatus, equipment and the storage medium of video frame | |
| JP2011022984A (en) | Stereoscopic video interactive system | |
| CN115100742B (en) | Meta universe exhibition and demonstration experience system based on space-apart gesture operation | |
| CN117369233B (en) | Holographic display method, device, equipment and storage medium | |
| Yao et al. | Neural radiance field-based visual rendering: A comprehensive review | |
| CN117611774A (en) | Multimedia display system and method based on augmented reality technology | |
| CN108628455A (en) | A kind of virtual husky picture method for drafting based on touch-screen gesture identification | |
| Bhakar et al. | A review on classifications of tracking systems in augmented reality | |
| CN114185431B (en) | Intelligent media interaction method based on MR technology | |
| CN116935008B (en) | A display interaction method and device based on mixed reality | |
| CN111383343B (en) | Home decoration design-oriented augmented reality image rendering coloring method based on generation countermeasure network technology | |
| Tao | A VR/AR-based display system for arts and crafts museum | |
| Gross et al. | Gesture Modelling: Using Video to Capture Freehand Modeling Commands | |
| CN116974416A (en) | Data processing method, device, equipment and readable storage medium | |
| CN100593175C (en) | Method and system for realizing organ animation | |
| Weng et al. | Green landscape 3D reconstruction and VR interactive art design experience using digital entertainment technology and entertainment gesture robots | |
| CN120472377B (en) | A fast moving small target tracking method based on dual-modal fusion | |
| Wang et al. | Virtual piano system based on monocular camera | |
| Tu et al. | An automatic base expression selection algorithm based on local blendshape model | |
| CN119383425B (en) | An intelligent video animation generation system based on AIGC | |
| CN119148849B (en) | A transparent display interaction method for augmented reality |
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 | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |