CN117011497A - Remote multiparty video interaction method based on AI universal assistant in AR scene - Google Patents

Remote multiparty video interaction method based on AI universal assistant in AR scene Download PDF

Info

Publication number
CN117011497A
CN117011497A CN202311023569.5A CN202311023569A CN117011497A CN 117011497 A CN117011497 A CN 117011497A CN 202311023569 A CN202311023569 A CN 202311023569A CN 117011497 A CN117011497 A CN 117011497A
Authority
CN
China
Prior art keywords
layer
video
target object
model
assistant
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.)
Pending
Application number
CN202311023569.5A
Other languages
Chinese (zh)
Inventor
李波
田慧云
陈非
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Suxin Iot Solutions Nanjing Co ltd
Original Assignee
Suxin Iot Solutions Nanjing Co ltd
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Suxin Iot Solutions Nanjing Co ltd filed Critical Suxin Iot Solutions Nanjing Co ltd
Priority to CN202311023569.5A priority Critical patent/CN117011497A/en
Publication of CN117011497A publication Critical patent/CN117011497A/en
Pending legal-status Critical Current

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T19/00—Manipulating 3D models or images for computer graphics
    • G06T19/006—Mixed reality
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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
    • G06N3/0464—Convolutional networks [CNN, ConvNet]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR 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; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T17/00—Three dimensional [3D] modelling, e.g. data description of 3D objects
    • 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
    • G06V10/77—Processing image or video features in feature spaces; using data integration or data reduction, e.g. principal component analysis [PCA] or independent component analysis [ICA] or self-organising maps [SOM]; Blind source separation
    • G06V10/774—Generating sets of training patterns; Bootstrap methods, e.g. bagging or boosting
    • 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
    • G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
    • G06V20/00—Scenes; Scene-specific elements
    • G06V20/40—Scenes; Scene-specific elements in video content

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Software Systems (AREA)
  • Evolutionary Computation (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Biomedical Technology (AREA)
  • Databases & Information Systems (AREA)
  • Mathematical Physics (AREA)
  • Computer Graphics (AREA)
  • Medical Informatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Geometry (AREA)
  • Computer Hardware Design (AREA)
  • Processing Or Creating Images (AREA)

Abstract

The invention discloses a remote multiparty video interaction method based on an AI universal assistant in an AR scene. Then, recognizing a three-dimensional model in a model library corresponding to a target object in the current video stream through a target recognition model based on the 3d video, inputting the description of the target object and the user requirement into an AI universal assistant, calling out the three-dimensional model which is most in line with the target object by the AI universal assistant, giving out a corresponding result according to the target requirement, displaying the three-dimensional model in an AR space together, and displaying the three-dimensional model in real time on each interaction party; according to the invention, on the basis of traditional AR multi-party remote video interaction, the target object in the current video stream is focused in real time by fusing the AI universal assistant, and a personalized solution is provided according to the user demand, so that more apparent communication experience can be provided for each party, and the communication efficiency of each party is effectively improved.

Description

Remote multiparty video interaction method based on AI universal assistant in AR scene
Technical Field
The invention relates to the technical field of artificial intelligence, in particular to a remote multiparty video interaction method based on an AI universal assistant in an AR scene.
Background
With the rapid development of internet technology, a remote interaction mode based on network technology is widely applied in more and more fields. Processors of electronic devices such as mobile phones and computers are continually evolving, so that applications of augmented reality (Augmented Reality, AR) are increasingly accepted. The augmented reality technology is used in the field of remote interaction, and is a brand new remote communication, communication and life work mode.
The augmented reality AR technology is a technology of displaying computer-generated virtual information (text, picture, three-dimensional model animation, etc.) superimposed in the real world, which improves the user's perception of the real world and changes the way the user observes the world. The AR technology is widely used in recent years in various fields such as industry, military, medical, education, entertainment, etc., by combining with a three-dimensional tracking technology, calculating the position of virtual information in a real environment by a computer and displaying the virtual information in an output device (such as a mobile phone screen, smart glasses, etc.).
In the industrial multiparty remote interaction process, three-dimensional modeling display and target inquiry are often required to be carried out on a certain target object in a current scene, more real-time feedback and more stereoscopic and vivid model display are required for the requirements of interaction parties, the existing AR technology does not have the function of identifying the target object in real time and providing a solution in real time, and the appearance of a general AI assistant provides higher convenience for multiparty remote interaction.
Disclosure of Invention
The invention aims to: aiming at the problems in the background art, the invention provides a remote multiparty video interaction method based on an AI universal assistant in an AR scene, which comprises the steps of firstly establishing a multiparty real-time interaction communication mechanism in an AR space, extracting audio information in multiparty real-time interaction, extracting semantics, and acquiring target object description and user requirements. And then, recognizing a three-dimensional model in a model library corresponding to the target object in the current video stream through a target recognition model based on the 3d video, inputting the target object description and the user requirement into an AI universal assistant, calling out the three-dimensional model which is most in line with the target object by the AI universal assistant, giving out a corresponding result according to the target requirement, displaying the result in an AR space together, and displaying the result in real time on each interaction party.
The technical scheme is as follows: a remote multiparty video interaction method based on an AI universal assistant in an AR scene comprises the following steps:
step S1, an initiator establishes real-time multi-person communication with a plurality of respondents through an intelligent terminal in a real-time video sharing mode;
step S2, an initiator loads an AR engine through an intelligent terminal, acquires a current environment through a camera, and establishes a reference plane for displaying a space model for displaying target contents to all parties of video communication subsequently;
step S3, in the multiparty video call, the initiator communicates with the responder aiming at the target object in the AR space; carrying out real-time semantic extraction on audio files recorded by each party in the current interaction process, and obtaining a target object aimed by a user and user requirements;
s4, capturing video pictures acquired by an initiator in real time, extracting video streams, identifying a target object aimed at by the current video stream based on a scene target identification model, and selecting a model corresponding to the current target object from a preset model library for displaying on a corresponding target object position in an AR space; specifically, the scene target recognition model structure comprises a plurality of stages of recognition modules, wherein each stage of recognition module comprises 3 sub-blocks, which are marked as a block n_1, a block n_2 and a block n_3, wherein n represents an n-th stage of recognition module; the method comprises the steps that each of the block_1 and the block_2 comprises a 3d convolution layer conv3d, a BN layer and a convolution function layer which are sequentially connected, and the block_3 comprises the BN layer and an activation function layer which are sequentially connected; the input dimension and the output dimension of the blockn_1 and the blockn_2 in each stage of the identification module are the same, and the input dimension and the output dimension of the blockn_1 and the blockn_2 between different stages are increased; adding add of the outputs of each stage of the blockn_1, the blockn_2 and the blockn_3, and inputting the added outputs to a layer of 3d pooling layer maxpool3d; the input sample firstly enters a layer of 3d convolution layer conv3d and a layer of 3d pooling layer maxpool3d, then sequentially enters each stage of identification module, and sequentially connects the output of the last stage of identification module to a layer of flexible layer and a full connection layer, wherein the output of the full connection layer is n, and the output of the full connection layer represents the total number of models in a model library.
Step S5, when the results of the target objects identified in the step S3 and the step S4 are the same, the user requirements and the target object information extracted in the step S3 are simultaneously transmitted to an AI general assistant, and the AI general assistant acquires a corresponding solution;
step S6, in the AR space, the AI general assistant displays the three-dimensional model of the target object which is most in line with the current scene description and is acquired in the step S4 on a corresponding position in the AR space, and simultaneously displays the solution acquired in the step S5 on the three-dimensional model corresponding to the target object in the reference plane; the initiator and the responder acquire related solutions for the target object; the system captures the audio information and the image information in the video interaction process in real time, and then carries out feature semantic extraction again along with the main body of each video interaction party, reselects the model, and provides a solution corresponding to the requirement.
Further, when the initiator shares the audio and video, the current video is uploaded to the own network server through the user datagram protocol or the real-time transmission protocol, and when the responder and the initiator interact in real time, the responder is connected to the own network server through the user datagram protocol, so that the effect of real-time audio and video interaction is achieved.
Further, the specific use method comprises the following steps:
step L1, extracting video stream information in real time, wherein the video stream information comprises a plurality of frames of video images, constructing samples based on a sliding window sampling mode and dividing the samples into a sample set and a test set;
step L2, inputting a sample set into a target recognition model structure for sample training; specifically, the sample set is firstly sequentially input into a conv3d convolution layer and a 3d pooling layer maxpool3d; then inputting the blocks 1_1, 1_2 and 1_3 of the first-stage identification module; adding add of inputs of the block1_1, the block1_2 and the block1_3, and inputting the added inputs to a 3d pooling layer maxpool3d; and so on; and inputting the output of the last stage of identification module to the flat layer for leveling operation, and finally inputting the output to the full-connection layer. The output dimension of the full connection layer is n, and represents the number of the existing model categories in the model library;
step L3, after training all training sets, inputting a test set for model verification until a target recognition model has recognition accuracy meeting the requirements; so far, the three-dimensional model which is most consistent with the target object in the existing model library is found out according to the input video information, and the three-dimensional model is output.
Further, the activation function employs a Relu function.
Compared with the prior art, the technical scheme adopted by the invention has the following beneficial effects:
(1) According to the remote multiparty video interaction method based on the AI universal assistant in the AR scene, on the basis of traditional AR multiparty remote video interaction, the real-time focusing and real-time answering solution is provided, the AI universal assistant is fused, the target object focused by the user and in video focusing in the current video stream is called in real time, personalized answering is provided according to the user requirement, the solution and the corresponding target object three-dimensional model are displayed together, more apparent communication experience can be provided for each party, and communication efficiency of each party is effectively improved.
(2) The invention provides a specific scene target object identification method, identifies the category of the scene target object through an artificial intelligent algorithm, provides a corresponding three-dimensional model, displays the model to the current user in a specific manner, and provides more visual experience for the user.
Drawings
Fig. 1 is a flowchart of a remote multiparty video interaction method based on an AI universal assistant in an AR scene provided by the present invention.
Fig. 2 is a schematic diagram of a target recognition model structure provided by the present invention.
Description of the embodiments
The invention provides a remote multiparty video interaction method based on an AI universal assistant in an AR scene, which comprises the following specific steps as shown in figure 1:
step S1, an initiator establishes real-time multi-person communication with a plurality of respondents through an intelligent terminal in a real-time video sharing mode;
when the initiator shares the audio and video, the current video is uploaded to the own network server through a User Datagram Protocol (UDP) or a real-time transmission protocol (RTP), and when the responder and the initiator interact in real time, the responder is connected to the own network server through the UDP, so that the effect of real-time audio and video interaction is achieved.
And S2, loading an AR engine by the initiator through the intelligent terminal, collecting the current environment through the camera, and establishing a reference plane for displaying the space model by the AR engine. This plane is used to subsequently present targeted content to parties to the video communication. In this embodiment, the content in the AR space is also uploaded to the own network server, so that the initiator and the responder share and perform video interaction synchronously.
Step S3, in the multiparty video call, the initiator communicates with the responder aiming at the target object in the AR space; carrying out real-time semantic extraction on audio files recorded by each party in the current interaction process; the semantic extraction aims at: 1. extracting a target object aimed by a user; 2. user requirements are extracted. In the process of video call, audio information is collected through equipment and translated into text information in real time; based on semantic extraction, locking a target object in the current scene; and simultaneously, the requirements of users at all sides are extracted.
And S4, capturing video pictures acquired by an initiator in real time, extracting a video stream, identifying a target object focused by the current video stream based on a scene target identification model, and selecting a model corresponding to the current target object from a preset model library for displaying on a corresponding target object position in an AR space.
In this embodiment, a specific scene target recognition model structure is provided, and specifically as shown in fig. 2, the main body includes a plurality of stage recognition modules; in this embodiment, a 3-stage identification module is used. Wherein each level identification module comprises 3 sub-blocks, denoted as blockn_1, blockn_2 and blockn_3, where n represents the n-th level identification module. The blockn_1 and the blockn_2 each comprise a 3d convolution layer conv3d, a BN layer and a convolution function layer which are connected in sequence. The blockn_3 comprises a BN layer and an activation function layer which are connected in sequence. The activation function in this embodiment employs a Relu function. The input dimension and the output dimension of the blockn_1 and the blockn_2 in each stage of the identification module are the same, and the input dimension and the output dimension of the blockn_1 and the blockn_2 between different stages are increased. The outputs of each stage of the blockn_1, the blockn_2 and the blockn_3 are added and then input to a layer 3d pooling layer maxpool3d.
The input sample firstly enters a layer of 3d convolution layer conv3d and a layer of 3d pooling layer maxpool3d, then sequentially enters each stage of identification module, and sequentially connects the output of the last stage of identification module to a layer of flexible layer and a full connection layer, wherein the output of the full connection layer is n, and the output of the full connection layer represents the total number of models in a model library.
The following provides a specific example:
when the method is used, video stream information is extracted in real time, wherein the video stream information comprises a plurality of frames of video images, and each 7 frames of images are taken as one sample in the embodiment. The method comprises the steps of constructing samples based on a sliding window sampling mode and dividing the samples into a sample set and a test set.
Inputting the sample set into a target recognition model structure for sample training; specifically, the sample set is first input to a conv3d convolution layer, wherein the input dimension in_channel is 3, the output dimension out_channel is 32, and the convolution kernel size kernel_size= (3, 3); next, a layer 3d pooling layer maxpool3d is input, the 3d pooling layer kernel_size= (1, 2), step size stride= (1, 2).
Then inputs to the block1_1, block1_2, and block1_3 of the first level identification module. The inputs of the block1_1, the block1_2 and the block1_3 are added and then input to the 3d pooling layer maxpool3d. And so on. In this embodiment, the 3d pooling layer parameters of each stage of the identification module are kernel_size= (2, 2), stride= (2, 2). And inputting the output of the last stage of identification module to the flat layer for leveling operation, and finally inputting the output to the full-connection layer. The output dimension of the full connection layer is n, and represents the number of the existing model categories in the model library.
After all training sets are trained, inputting a test set for model verification until a target recognition model has recognition accuracy meeting the requirements. So far, the three-dimensional model which is most consistent with the target object in the existing model library can be found out according to the input video information, and the three-dimensional model is output.
And S5, when the target results identified in the step S3 and the step S4 are the same, simultaneously transmitting the user requirements and the target object information extracted in the step S3 to an AI general assistant, and acquiring corresponding solutions by the AI general assistant.
And S6, in the AR space, the AI general assistant displays the three-dimensional model of the target object which is most suitable for the current scene description and is acquired in the step S4 on a corresponding position in the AR space. And simultaneously, displaying the solution obtained in the step S5 on a three-dimensional model corresponding to the target object in the reference plane. The initiator and the responder acquire relevant solutions for the target object in real time. The system captures the audio information and the image information in the video interaction process in real time, and then carries out feature semantic extraction again along with the main body of each video interaction party, reselects the model, and provides a solution corresponding to the requirement.
The foregoing is merely a preferred embodiment of the present invention and it should be noted that modifications and adaptations to those skilled in the art may be made without departing from the principles of the present invention, which are intended to be comprehended within the scope of the present invention.

Claims (4)

1. A remote multiparty video interaction method based on an AI universal assistant in an AR scene is characterized by comprising the following steps:
step S1, an initiator establishes real-time multi-person communication with a plurality of respondents through an intelligent terminal in a real-time video sharing mode;
step S2, an initiator loads an AR engine through an intelligent terminal, acquires a current environment through a camera, and establishes a reference plane for displaying a space model for displaying target contents to all parties of video communication subsequently;
step S3, in the multiparty video call, the initiator communicates with the responder aiming at the target object in the AR space; carrying out real-time semantic extraction on audio files recorded by each party in the current interaction process, and obtaining a target object aimed by a user and user requirements;
s4, capturing video pictures acquired by an initiator in real time, extracting video streams, identifying a target object aimed at by the current video stream based on a scene target identification model, and selecting a model corresponding to the current target object from a preset model library for displaying on a corresponding target object position in an AR space; specifically, the scene target recognition model structure comprises a plurality of stages of recognition modules, wherein each stage of recognition module comprises 3 sub-blocks, which are marked as a block n_1, a block n_2 and a block n_3, wherein n represents an n-th stage of recognition module; the method comprises the steps that each of the block_1 and the block_2 comprises a 3d convolution layer conv3d, a BN layer and a convolution function layer which are sequentially connected, and the block_3 comprises the BN layer and an activation function layer which are sequentially connected; the input dimension and the output dimension of the blockn_1 and the blockn_2 in each stage of the identification module are the same, and the input dimension and the output dimension of the blockn_1 and the blockn_2 between different stages are increased; adding add of the outputs of each stage of the blockn_1, the blockn_2 and the blockn_3, and inputting the added outputs to a layer of 3d pooling layer maxpool3d; the input sample firstly enters a layer of 3d convolution layer conv3d and a layer of 3d pooling layer maxpool3d, then sequentially enters each stage of identification module, and sequentially connects the output of the last stage of identification module to a layer of flexible layer and a full connection layer, wherein the output of the full connection layer is n, and the output of the full connection layer represents the total number of models in a model library;
step S5, when the results of the target objects identified in the step S3 and the step S4 are the same, the user requirements and the target object information extracted in the step S3 are simultaneously transmitted to an AI general assistant, and the AI general assistant acquires a corresponding solution;
step S6, in the AR space, the AI general assistant displays the three-dimensional model of the target object which is most in line with the current scene description and is acquired in the step S4 on a corresponding position in the AR space, and simultaneously displays the solution acquired in the step S5 on the three-dimensional model corresponding to the target object in the reference plane; the initiator and the responder acquire related solutions for the target object; the system captures the audio information and the image information in the video interaction process in real time, and then carries out feature semantic extraction again along with the main body of each video interaction party, reselects the model, and provides a solution corresponding to the requirement.
2. The remote multiparty video interaction method based on AI universal assistant in AR scene according to claim 1, wherein when the sponsor shares audio/video, the current video is uploaded to the own network server through user datagram protocol or real-time transmission protocol, when the responder interacts with the sponsor in real time, the responder is connected to the own network server through user datagram protocol, thus achieving the effect of real-time audio/video interaction.
3. The remote multiparty video interaction method based on AI universal assistant in AR scene according to claim 1, wherein the specific usage method comprises:
step L1, extracting video stream information in real time, wherein the video stream information comprises a plurality of frames of video images, constructing samples based on a sliding window sampling mode and dividing the samples into a sample set and a test set;
step L2, inputting a sample set into a target recognition model structure for sample training; specifically, the sample set is firstly sequentially input into a conv3d convolution layer and a 3d pooling layer maxpool3d; then inputting the blocks 1_1, 1_2 and 1_3 of the first-stage identification module; adding add of inputs of the block1_1, the block1_2 and the block1_3, and inputting the added inputs to a 3d pooling layer maxpool3d; and so on; and inputting the output of the last stage of identification module to the flat layer for leveling operation, and finally inputting the output to the full-connection layer. The output dimension of the full connection layer is n, and represents the number of the existing model categories in the model library;
step L3, after training all training sets, inputting a test set for model verification until a target recognition model has recognition accuracy meeting the requirements; so far, the three-dimensional model which is most consistent with the target object in the existing model library is found out according to the input video information, and the three-dimensional model is output.
4. The method for remote multiparty video interaction based on AI-generalized assistant in an AR scenario according to claim 1, wherein the activation function employs a Relu function.
CN202311023569.5A 2023-08-15 2023-08-15 Remote multiparty video interaction method based on AI universal assistant in AR scene Pending CN117011497A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202311023569.5A CN117011497A (en) 2023-08-15 2023-08-15 Remote multiparty video interaction method based on AI universal assistant in AR scene

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202311023569.5A CN117011497A (en) 2023-08-15 2023-08-15 Remote multiparty video interaction method based on AI universal assistant in AR scene

Publications (1)

Publication Number Publication Date
CN117011497A true CN117011497A (en) 2023-11-07

Family

ID=88567207

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202311023569.5A Pending CN117011497A (en) 2023-08-15 2023-08-15 Remote multiparty video interaction method based on AI universal assistant in AR scene

Country Status (1)

Country Link
CN (1) CN117011497A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117528131A (en) * 2024-01-05 2024-02-06 青岛美迪康数字工程有限公司 AI integrated display system and method for medical image

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117528131A (en) * 2024-01-05 2024-02-06 青岛美迪康数字工程有限公司 AI integrated display system and method for medical image
CN117528131B (en) * 2024-01-05 2024-04-05 青岛美迪康数字工程有限公司 AI integrated display system and method for medical image

Similar Documents

Publication Publication Date Title
CN111556278B (en) Video processing method, video display device and storage medium
CN107911644B (en) Method and device for carrying out video call based on virtual face expression
CN110517185B (en) Image processing method, device, electronic equipment and storage medium
US11670015B2 (en) Method and apparatus for generating video
CN110266992A (en) A kind of long-distance video interactive system and method based on augmented reality
CN107633441A (en) Commodity in track identification video image and the method and apparatus for showing merchandise news
CN110866977B (en) Augmented reality processing method, device, system, storage medium and electronic equipment
CN112991553B (en) Information display method and device, electronic equipment and storage medium
CN112199016B (en) Image processing method, image processing device, electronic equipment and computer readable storage medium
CN106534757A (en) Face exchange method and device, anchor terminal and audience terminal
CN111368796A (en) Face image processing method and device, electronic equipment and storage medium
CN117011497A (en) Remote multiparty video interaction method based on AI universal assistant in AR scene
WO2021098151A1 (en) Special effect video synthesis method and apparatus, computer device, and storage medium
CN113949808B (en) Video generation method and device, readable medium and electronic equipment
CN112492231B (en) Remote interaction method, device, electronic equipment and computer readable storage medium
CN115379125B (en) Interactive information sending method, device, server and medium
CN112839196A (en) Method, device and storage medium for realizing online conference
CN114445562A (en) Three-dimensional reconstruction method and device, electronic device and storage medium
CN114463470A (en) Virtual space browsing method and device, electronic equipment and readable storage medium
CN114615455A (en) Teleconference processing method, teleconference processing device, teleconference system, and storage medium
CN112131431A (en) Data processing method, data processing equipment and computer readable storage medium
CN108320331B (en) Method and equipment for generating augmented reality video information of user scene
WO2023217138A1 (en) Parameter configuration method and apparatus, device, storage medium and product
CN115278297B (en) Data processing method, device, equipment and storage medium based on drive video
KR20110116116A (en) Method for providing text relation information using mobile terminal and mobile terminal

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