CN117082309A - Meteorological service short video processing method and platform system based on artificial intelligence - Google Patents
Meteorological service short video processing method and platform system based on artificial intelligence Download PDFInfo
- Publication number
- CN117082309A CN117082309A CN202310911134.8A CN202310911134A CN117082309A CN 117082309 A CN117082309 A CN 117082309A CN 202310911134 A CN202310911134 A CN 202310911134A CN 117082309 A CN117082309 A CN 117082309A
- Authority
- CN
- China
- Prior art keywords
- image
- short video
- description
- weather
- enhancement material
- 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
- 238000013473 artificial intelligence Methods 0.000 title claims abstract description 28
- 238000003672 processing method Methods 0.000 title claims abstract description 9
- 239000000463 material Substances 0.000 claims abstract description 582
- 239000013598 vector Substances 0.000 claims abstract description 412
- 238000004220 aggregation Methods 0.000 claims abstract description 155
- 230000002776 aggregation Effects 0.000 claims abstract description 155
- 238000012545 processing Methods 0.000 claims abstract description 59
- 238000005065 mining Methods 0.000 claims description 207
- 238000004458 analytical method Methods 0.000 claims description 37
- 238000013507 mapping Methods 0.000 claims description 30
- 230000004927 fusion Effects 0.000 claims description 28
- 238000000034 method Methods 0.000 claims description 18
- 238000009826 distribution Methods 0.000 claims description 17
- 238000013441 quality evaluation Methods 0.000 claims description 16
- 238000011144 upstream manufacturing Methods 0.000 claims description 11
- 238000013508 migration Methods 0.000 claims description 5
- 230000005012 migration Effects 0.000 claims description 5
- 230000004931 aggregating effect Effects 0.000 claims description 4
- 238000005457 optimization Methods 0.000 abstract description 3
- 230000002688 persistence Effects 0.000 abstract description 3
- 230000000007 visual effect Effects 0.000 abstract description 3
- 230000002787 reinforcement Effects 0.000 description 14
- 238000004590 computer program Methods 0.000 description 8
- 238000004422 calculation algorithm Methods 0.000 description 6
- 230000006870 function Effects 0.000 description 4
- 238000001303 quality assessment method Methods 0.000 description 3
- 230000002441 reversible effect Effects 0.000 description 3
- 238000004891 communication Methods 0.000 description 2
- 238000004519 manufacturing process Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000008569 process Effects 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 238000012935 Averaging Methods 0.000 description 1
- 238000012098 association analyses Methods 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 230000006866 deterioration Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 230000009191 jumping Effects 0.000 description 1
- 238000013139 quantization Methods 0.000 description 1
- 238000005728 strengthening Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/85—Assembly of content; Generation of multimedia applications
- H04N21/854—Content authoring
-
- 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
- G06V10/80—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level
- G06V10/806—Fusion, i.e. combining data from various sources at the sensor level, preprocessing level, feature extraction level or classification level of extracted features
-
- 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
- G06V20/46—Extracting features or characteristics from the video content, e.g. video fingerprints, representative shots or key frames
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
- H04N21/44008—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving operations for analysing video streams, e.g. detecting features or characteristics in the video stream
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/43—Processing of content or additional data, e.g. demultiplexing additional data from a digital video stream; Elementary client operations, e.g. monitoring of home network or synchronising decoder's clock; Client middleware
- H04N21/44—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs
- H04N21/44016—Processing of video elementary streams, e.g. splicing a video clip retrieved from local storage with an incoming video stream or rendering scenes according to encoded video stream scene graphs involving splicing one content stream with another content stream, e.g. for substituting a video clip
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/40—Client devices specifically adapted for the reception of or interaction with content, e.g. set-top-box [STB]; Operations thereof
- H04N21/47—End-user applications
- H04N21/488—Data services, e.g. news ticker
- H04N21/4886—Data services, e.g. news ticker for displaying a ticker, e.g. scrolling banner for news, stock exchange, weather data
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/81—Monomedia components thereof
- H04N21/8126—Monomedia components thereof involving additional data, e.g. news, sports, stocks, weather forecasts
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/81—Monomedia components thereof
- H04N21/8146—Monomedia components thereof involving graphical data, e.g. 3D object, 2D graphics
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N21/00—Selective content distribution, e.g. interactive television or video on demand [VOD]
- H04N21/80—Generation or processing of content or additional data by content creator independently of the distribution process; Content per se
- H04N21/81—Monomedia components thereof
- H04N21/816—Monomedia components thereof involving special video data, e.g 3D video
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Computer Graphics (AREA)
- Computer Security & Cryptography (AREA)
- Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Computing Systems (AREA)
- Databases & Information Systems (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Medical Informatics (AREA)
- Software Systems (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
According to the weather service short video processing method and platform system based on artificial intelligence, under the task of carrying out short video image enhancement on the weather subscription service short video to be processed, short video image enhancement can be carried out through the target aggregation description vector obtained through aggregation, and the characteristic connection between the weather subscription service short video image and the enhancement materials and the image layer involving condition between the enhancement materials in the target aggregation description vector can be utilized, so that the characteristic enhancement processing quality and pertinence of the weather subscription service short video image are improved, and the individualized image enhancement processing of the weather subscription service short video is realized. Therefore, the content optimization processing of the weather service short video can be realized, and the visual understanding degree and the pushing persistence of the weather subscription service short video are improved.
Description
Technical Field
The invention relates to the technical field of artificial intelligence and weather service, in particular to a weather service short video processing method and platform system based on artificial intelligence.
Background
With the gradual deterioration of global environment problems, weather services and social production/public life have close relations, influence the production and management of enterprises and the daily life of the public, and timely carry out relevant weather early warning and service information pushing and conveying, which is particularly important in the current information-based society. However, when the weather service is pushed, related weather information is pushed in a hard and mechanized mode in the form of images or texts in most cases.
Disclosure of Invention
In order to solve the problems, the invention provides a weather service short video processing method and a platform system based on artificial intelligence.
In a first aspect, a weather service short video processing method based on artificial intelligence is provided, and is applied to an artificial intelligence platform system, and the method comprises the following steps:
acquiring short videos and short video image enhancement material libraries of weather subscription services to be processed;
loading the short video of the weather subscription service to be processed into an image description mining model for image description mining to obtain short video image description vectors of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed;
Performing material description mining on the short video image enhancement material library through a material description mining model to obtain an image enhancement material description vector of the short video image enhancement material library, wherein the image enhancement material description vector reflects a material attribute relation network of the short video image enhancement material library;
performing aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector to obtain a target aggregation description vector;
and loading the target aggregate description vector to a short video image enhancement model to enhance the short video image, so as to obtain an enhanced weather subscription service short video image corresponding to the to-be-processed weather subscription service short video.
Preferably, the material description mining model includes a plurality of description vector sorting components corresponding to a plurality of enhancement material clusters in the short video image enhancement material library one by one, and the performing material description mining on the short video image enhancement material library by the material description mining model to obtain an image enhancement material description vector of the short video image enhancement material library includes:
performing image description mining on the enhancement material information of each enhancement material cluster in the plurality of enhancement material clusters to obtain an enhancement material cluster description vector corresponding to each enhancement material cluster;
Loading an enhancement material cluster description vector corresponding to a selected enhancement material cluster of the short video image enhancement material library to a description vector arrangement component corresponding to the selected enhancement material cluster to obtain enhancement material cluster fusion vector data corresponding to the selected enhancement material cluster;
the plurality of enhancement material clusters are walked according to the upstream enhancement material clusters of the selected enhancement material clusters;
loading the fusion vector data of the enhancement material cluster corresponding to the downstream enhancement material cluster of the current traveling enhancement material cluster and the description vector of the enhancement material cluster corresponding to the current traveling enhancement material cluster into a description vector arrangement component corresponding to the current traveling enhancement material cluster to obtain the fusion vector data of the enhancement material cluster corresponding to the current traveling enhancement material cluster;
and after finishing the migration of the plurality of enhancement material clusters, taking the fusion vector data of the enhancement material clusters corresponding to the top enhancement material cluster in the short video image enhancement material library as the image enhancement material description vector.
Preferably, the image description mining model includes an image block disassembly component, an image description mining component and an image description aggregation component, and the loading the short video of the to-be-processed weather subscription service into the image description mining model to perform image description mining to obtain a short video image description vector of the short video of the to-be-processed weather subscription service and weather service page description vectors of a plurality of weather service page image blocks of the short video of the to-be-processed weather subscription service includes:
Loading the short video of the weather subscription service to be processed into the image block disassembly component to disassemble the image blocks, so as to obtain the image blocks of the plurality of weather service pages;
loading the plurality of weather service page image blocks into the image description mining assembly to perform image description mining to obtain basic image descriptors, distribution feature data and image association features corresponding to the plurality of weather service page image blocks;
and loading the basic image descriptors, the distribution characteristic data and the image association characteristics to the image description aggregation component to perform association processing on the plurality of weather service page image blocks, so as to obtain the short video image description vector and the weather service page description vector.
Preferably, the aggregating the image enhancement material description vector, the weather service page description vector and the short video image description vector to obtain a target aggregate description vector includes:
loading the image enhancement material description vector and the meteorological service page description vector into an aggregation sub-model for aggregation operation to obtain a basic aggregation description vector;
and loading the basic aggregate description vector, the short video image description vector and the image enhancement material description vector into an enhancement sub-model for feature enhancement processing to obtain the target aggregate description vector.
Preferably, the aggregation sub-model includes a linkage analysis component, an interval value mapping component and a weight processing component, and loading the image enhancement material description vector and the weather service page description vector into the aggregation sub-model for aggregation operation, where obtaining a basic aggregation description vector includes:
loading the meteorological service page description vector and the image enhancement material description vector to the linkage analysis component for linkage analysis to obtain a target linkage analysis result;
loading the target linkage analysis result to the interval value mapping component for interval value mapping processing to obtain a linkage weighting coefficient of the weather service page description vector;
and loading the linkage weighting coefficient and the weather service page description vector to the weight processing component to perform weight-based aggregation operation to obtain the basic aggregation description vector.
Preferably, the method further comprises:
acquiring a past weather subscription service short video and a priori enhanced weather subscription service short video image corresponding to the past weather subscription service short video;
loading the past weather subscription service short video to an initial image description mining model for image description mining to obtain a past short video image description vector of the past weather subscription service short video and past weather service page description vectors of a plurality of past weather service page image blocks of the past weather subscription service short video;
Performing material description mining on the short video image enhancement material library based on an initial material description mining model to obtain a past image enhancement material description vector of the short video image enhancement material library;
loading the past image enhancement material description vector and the past weather service page description vector into an initial aggregation sub-model for aggregation operation to obtain a past basic aggregation description vector;
loading the past basic aggregate description vector, the past short video image description vector and the past image enhancement material description vector into an initial enhancement sub-model for feature enhancement processing to obtain a target aggregate description vector example;
loading the target aggregate description vector example to an initial short video image enhancement model for short video image enhancement to obtain a predicted enhanced weather subscription service short video image corresponding to the past weather subscription service short video;
determining target model quality evaluation data according to the prior enhanced weather subscription service short video image and the predicted enhanced weather subscription service short video image;
and debugging the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial enhancement sub-model and the initial short video image enhancement model according to the target model quality evaluation data to obtain the image description mining model, the material description mining model, the aggregation sub-model, the enhancement sub-model and the short video image enhancement model.
Preferably, the enhanced weather subscription service short video image is image data corresponding to any selected enhanced material cluster feature in the short video image enhanced material library, after the target aggregate description vector is loaded into a short video image enhancement model to enhance the short video image, the method further includes:
acquiring an enhancement material matching instruction, wherein the enhancement material matching instruction reflects the image data and the projection characteristics of the multi-layer image data corresponding to the image data;
and determining target multi-layer image data corresponding to the enhanced weather subscription service short video image according to the enhanced material matching instruction.
In a second aspect, an artificial intelligence platform system is provided, the artificial intelligence platform system comprising:
the data acquisition module is used for acquiring short videos and short video image enhancement material libraries of weather subscription services to be processed;
the first mining module is used for loading the short video of the weather subscription service to be processed into an image description mining model to carry out image description mining to obtain short video image description vectors of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed;
The second mining module is used for carrying out material description mining on the short video image enhancement material library through a material description mining model to obtain an image enhancement material description vector of the short video image enhancement material library, wherein the image enhancement material description vector reflects a material attribute relation network of the short video image enhancement material library;
the vector aggregation module is used for carrying out aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector to obtain a target aggregation description vector;
and the video processing module is used for loading the target aggregate description vector to a short video image enhancement model to enhance the short video image, so as to obtain an enhanced weather subscription service short video image corresponding to the to-be-processed weather subscription service short video.
Preferably, the material description mining model includes a plurality of description vector sorting components corresponding to a plurality of enhancement material clusters in the short video image enhancement material library one by one, and the second mining module performs material description mining on the short video image enhancement material library through the material description mining model, and obtaining the image enhancement material description vector of the short video image enhancement material library includes:
Performing image description mining on the enhancement material information of each enhancement material cluster in the plurality of enhancement material clusters to obtain an enhancement material cluster description vector corresponding to each enhancement material cluster;
loading an enhancement material cluster description vector corresponding to a selected enhancement material cluster of the short video image enhancement material library to a description vector arrangement component corresponding to the selected enhancement material cluster to obtain enhancement material cluster fusion vector data corresponding to the selected enhancement material cluster;
the plurality of enhancement material clusters are walked according to the upstream enhancement material clusters of the selected enhancement material clusters;
loading the fusion vector data of the enhancement material cluster corresponding to the downstream enhancement material cluster of the current traveling enhancement material cluster and the description vector of the enhancement material cluster corresponding to the current traveling enhancement material cluster into a description vector arrangement component corresponding to the current traveling enhancement material cluster to obtain the fusion vector data of the enhancement material cluster corresponding to the current traveling enhancement material cluster;
and after finishing the migration of the plurality of enhancement material clusters, taking the fusion vector data of the enhancement material clusters corresponding to the top enhancement material cluster in the short video image enhancement material library as the image enhancement material description vector.
Preferably, the image description mining model includes an image block disassembly component, an image description mining component and an image description aggregation component, and the first mining module loads the short video of the weather subscription service to be processed into the image description mining model to perform image description mining, and obtaining a short video image description vector of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed includes:
loading the short video of the weather subscription service to be processed into the image block disassembly component to disassemble the image blocks, so as to obtain the image blocks of the plurality of weather service pages;
loading the plurality of weather service page image blocks into the image description mining assembly to perform image description mining to obtain basic image descriptors, distribution feature data and image association features corresponding to the plurality of weather service page image blocks;
and loading the basic image descriptors, the distribution characteristic data and the image association characteristics to the image description aggregation component to perform association processing on the plurality of weather service page image blocks, so as to obtain the short video image description vector and the weather service page description vector.
Preferably, the vector aggregation module performs an aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector, and the obtaining a target aggregate description vector includes:
loading the image enhancement material description vector and the meteorological service page description vector into an aggregation sub-model for aggregation operation to obtain a basic aggregation description vector;
loading the basic aggregate description vector, the short video image description vector and the image enhancement material description vector into an enhancement sub-model for feature enhancement processing to obtain the target aggregate description vector;
the aggregation sub-model comprises a linkage analysis component, an interval numerical mapping component and a weight processing component, wherein the loading of the image enhancement material description vector and the weather service page description vector into the aggregation sub-model for aggregation operation, the obtaining of a basic aggregation description vector comprises the following steps: loading the meteorological service page description vector and the image enhancement material description vector to the linkage analysis component for linkage analysis to obtain a target linkage analysis result; loading the target linkage analysis result to the interval value mapping component for interval value mapping processing to obtain a linkage weighting coefficient of the weather service page description vector; loading the linkage weighting coefficient and the weather service page description vector to the weight processing component to perform weight-based aggregation operation to obtain the basic aggregation description vector;
Wherein, the artificial intelligence platform system also includes a model debugging module for: acquiring a past weather subscription service short video and a priori enhanced weather subscription service short video image corresponding to the past weather subscription service short video; loading the past weather subscription service short video to an initial image description mining model for image description mining to obtain a past short video image description vector of the past weather subscription service short video and past weather service page description vectors of a plurality of past weather service page image blocks of the past weather subscription service short video; performing material description mining on the short video image enhancement material library based on an initial material description mining model to obtain a past image enhancement material description vector of the short video image enhancement material library; loading the past image enhancement material description vector and the past weather service page description vector into an initial aggregation sub-model for aggregation operation to obtain a past basic aggregation description vector; loading the past basic aggregate description vector, the past short video image description vector and the past image enhancement material description vector into an initial enhancement sub-model for feature enhancement processing to obtain a target aggregate description vector example; loading the target aggregate description vector example to an initial short video image enhancement model for short video image enhancement to obtain a predicted enhanced weather subscription service short video image corresponding to the past weather subscription service short video; determining target model quality evaluation data according to the prior enhanced weather subscription service short video image and the predicted enhanced weather subscription service short video image; and debugging the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial enhancement sub-model and the initial short video image enhancement model according to the target model quality evaluation data to obtain the image description mining model, the material description mining model, the aggregation sub-model, the enhancement sub-model and the short video image enhancement model.
Preferably, the enhanced weather subscription service short video image is image data corresponding to any selected enhanced material cluster feature in the short video image enhanced material library, after the target aggregate description vector is loaded to a short video image enhancement model to perform short video image enhancement, the artificial intelligent platform system further comprises an image matching module, configured to:
acquiring an enhancement material matching instruction, wherein the enhancement material matching instruction reflects the image data and the projection characteristics of the multi-layer image data corresponding to the image data;
and determining target multi-layer image data corresponding to the enhanced weather subscription service short video image according to the enhanced material matching instruction.
In a third aspect, an artificial intelligence platform system is provided, comprising a processor and a memory in communication with each other, the processor being adapted to retrieve a computer program from the memory and to implement the method of the first aspect by running the computer program.
In a fourth aspect, a computer readable storage medium is provided, on which a computer program is stored, which computer program, when run, implements the method of the first aspect.
The embodiment of the invention acquires short videos and short video image enhancement material libraries of weather subscription service to be processed; loading the short video of the weather subscription service to be processed into an image description mining model for image description mining to obtain short video image description vectors of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed; performing material description mining on the short video image enhancement material library through a material description mining model to obtain image enhancement material description vectors of the short video image enhancement material library, wherein the image enhancement material description vectors represent a material attribute relation network of the short video image enhancement material library; then, carrying out aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector to obtain a target aggregation description vector; and finally, loading the target aggregate description vector into a short video image enhancement model to enhance the short video image, so as to obtain an enhanced weather subscription service short video image corresponding to the to-be-processed weather subscription service short video.
Under the task of carrying out short video image enhancement on short video of weather subscription service to be processed, the embodiment of the invention aggregates the image enhancement material description vector containing the material attribute relation network of the short video image enhancement material library, the short video image description vector of the short video of weather subscription service to be processed and the weather service page description vectors of a plurality of weather service page image blocks of the short video of weather subscription service to be processed to obtain the target aggregate description vector, and carries out short video image enhancement on the target aggregate description vector, so that the characteristic enhancement processing quality and pertinence of the weather subscription service short video image can be improved by utilizing the characteristic connection between the weather subscription service short video image and enhancement material in the target aggregate description vector and the image layer involving condition between the enhancement material, thereby realizing the individualized image enhancement processing of the weather subscription service short video. Therefore, the content optimization processing of the weather service short video can be realized, and the visual understanding degree and the pushing persistence of the weather subscription service short video are improved.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments will be briefly described below, it being understood that the following drawings only illustrate some embodiments of the present invention and therefore should not be considered as limiting the scope, and other related 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 weather service short video processing method based on artificial intelligence according to an embodiment of the invention.
Detailed Description
In order to better understand the above technical solutions, the following detailed description of the technical solutions of the present invention is made by using the accompanying drawings and specific embodiments, and it should be understood that the specific features of the embodiments and the embodiments of the present invention are detailed descriptions of the technical solutions of the present invention, and not limiting the technical solutions of the present invention, and the technical features of the embodiments and the embodiments of the present invention may be combined with each other without conflict.
FIG. 1 illustrates an artificial intelligence based weather service short video processing method for an artificial intelligence platform system, the method comprising the following steps 101-105.
And step 101, acquiring short videos and short video image enhancement material libraries of weather subscription services to be processed.
In the embodiment of the invention, the short video of the weather subscription service to be processed can comprise short videos of the weather subscription service to be processed of a plurality of weather service items, and the short video of the weather subscription service to be processed can comprise weather forecast images, regional weather disaster early warning images and the like by way of example.
In the embodiment of the invention, the short video image enhancement material library can comprise a short video image enhancement material library of a plurality of weather service projects, wherein the short video image enhancement material library of each weather service project can be the same as providing material information for performing image enhancement processing on short videos of weather subscription services to be processed, and the short video image enhancement material library can record the material information in a hierarchical or tree-shaped form, for example, the provided material information can perform VR/AR/MR layer enhancement processing on the short videos of the weather subscription services to be processed, so that three-dimensional display of the short videos of the weather subscription services to be processed is realized.
In some examples, weather service item information corresponding to a short video of a weather subscription service to be processed may be annotated in advance to obtain a short video image enhancement material library matching the weather service item information.
Step 102, loading the short video of the weather subscription service to be processed into an image description mining model to perform image description mining, and obtaining short video image description vectors of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed.
In the embodiment of the invention, the short video image description vector can represent the image content characteristics of the short video image of the weather subscription service, and the weather service page description vector can represent the image detail characteristics of the corresponding weather service page image block in the short video image of the weather subscription service. The image description mining model can perform image description mining by combining a plurality of image blocks with correlation of the short video of the weather subscription service to be processed after the short video of the weather subscription service to be processed is loaded into the image description mining model, so that a short video image description vector and a weather service page description vector are obtained.
In some examples, the short video image description vector and the weather service page description vector may be characterized by quantization vectors.
In some examples, the image description mining model may be obtained after image description mining and debugging is performed on an initial image description mining model, where the initial image description mining model may include an image block disassembly component, an image description mining component and an image description aggregation component, and by way of example, the foregoing loading a short video of a weather subscription service to be processed into the image description mining model to perform image description mining to obtain a short video image description vector of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed may include steps 201-203.
Step 201, loading the short video of the weather subscription service to be processed into an image block disassembly component to disassemble the image blocks, and obtaining a plurality of weather service page image blocks.
In some examples, a weather forecast image, a regional weather disaster early warning image and the like in a short weather subscription service video to be processed can be subjected to characteristic strengthening processing (splicing processing) to obtain a short weather subscription service video image to be disassembled, the short weather subscription service video image to be disassembled is loaded into an image block disassembling component to disassemble an image block, a weather service page image block queue is obtained, and the weather service page image block queue can comprise a plurality of weather service page image blocks.
Step 202, loading a plurality of weather service page image blocks into an image description mining component for image description mining to obtain basic image descriptors, distribution feature data and image association features corresponding to the plurality of weather service page image blocks.
The basic image descriptors can represent image detail characteristics of the image blocks of the corresponding weather service pages, the image association characteristics can represent associated tag vectors of the image areas of the image blocks of the corresponding weather service pages in short videos of the weather subscription service to be processed, and the distribution characteristic data can represent positioning vectors of the image blocks of the corresponding weather service pages in the image areas.
Under some examples, the image description mining component may include a first feature mapping unit, a second feature mapping unit, and a third feature mapping unit, load a plurality of weather service page image blocks to the first feature mapping unit for performing image material description mining to obtain a basic image descriptor, load a plurality of weather service page image blocks to the second feature mapping unit for performing material tag description mining to obtain image association features, and load a plurality of weather service page image blocks to the third feature mapping unit for performing material distribution description mining to obtain distribution feature data.
And 203, loading the basic image descriptors, the distributed feature data and the image association features into an image description aggregation component to perform association processing on a plurality of weather service page image blocks, so as to obtain short video image description vectors and weather service page description vectors.
In some examples, the image description aggregation component may include a feature reversible operator (bi-directional encoding operator), load the base image descriptor, the distribution feature data, and the image-associated features into the feature reversible operator, and aggregate the image detail features of the associated image blocks of each weather service page image block by the feature reversible operator to obtain a weather service page description vector for each weather service page image block.
In some examples, considering that the above several weather service page image blocks may further include a distinguishing identifier, the base image descriptors, the distribution feature data, and the image association features corresponding to the above several weather service page image blocks may include: distinguishing basic image descriptors, distribution feature data and image association features corresponding to the identifiers; correspondingly, the loading the basic image description sub, the distributed feature data and the image association features into the image description aggregation component to perform association processing on a plurality of weather service page image blocks, and the obtaining of the short video image description vector and the weather service page description vector may include: the basic image descriptors, the distributed feature data and the image association features corresponding to the distinguishing marks are loaded to the image description aggregation component to carry out association processing on a plurality of weather service page image blocks, so that weather service page description vectors of the distinguishing marks are obtained, and the weather service page description vectors of the distinguishing marks can represent image content features in actual implementation, so that the weather service page description vectors of the distinguishing marks can be used as short video image description vectors.
Under other design ideas, the image detail characteristic stitching processing can be carried out on the weather service page description vector of each weather service page image block to obtain a short video image description vector. Optionally, performing image detail feature stitching processing on the weather service page description vector of each weather service page image block to obtain a short video image description vector may include averaging the weather service page description vector of each weather service page image block to obtain a short video image description vector.
Therefore, the image description mining model comprising the image block disassembly component, the image description mining component and the image description aggregation component is used for carrying out characteristic analysis of the associated image blocks on the short video of the weather subscription service to be processed, so that aggregation of image content characteristics of the short video of the weather subscription service to be processed is realized, and richness and accuracy of image detail characteristics of the short video of the weather subscription service to be processed can be ensured.
And step 103, carrying out material description mining on the short video image enhancement material library through a material description mining model to obtain image enhancement material description vectors of the short video image enhancement material library, wherein the image enhancement material description vectors represent a material attribute relation network of the short video image enhancement material library.
In the embodiment of the invention, the image enhancement material description vector can represent a material attribute relation network (relation among attributes of different materials) of a short video image enhancement material library, the material attribute relation network of the short video image enhancement material library can comprise material relation characteristics and image detail characteristics of the short video image enhancement material library, and the material relation characteristics of the short video image enhancement material library can be as follows: the upstream and downstream subordinate conditions of the enhancement materials in the short video image enhancement material library and the involving conditions among the enhancement materials of each layer, and the image detail characteristics of the short video image enhancement material library can be fusion characteristics from the selected enhancement material cluster to the top enhancement material cluster.
In some examples, the material description mining model may be obtained after performing material description mining on an initial material description mining model, where the initial material description mining model may include a plurality of description vector sorting components corresponding to a plurality of enhancement material clusters in a preset short video image enhancement material library one by one, and by using the material description mining model, for example, performing material description mining on the short video image enhancement material library to obtain image enhancement material description vectors of the short video image enhancement material library, and the method may include steps 301-305.
Step 301, performing image description mining on the enhancement material information of each enhancement material cluster in the plurality of enhancement material clusters to obtain an enhancement material cluster description vector corresponding to each enhancement material cluster.
For example, the enhancement material cluster description vector may characterize image detail characteristics of the enhancement material information itself for the corresponding enhancement material cluster.
In some examples, the material description mining model may further include an enhancement material image description mining component, and the enhancement material information of each enhancement material cluster is loaded into the enhancement material image description mining component to perform image description mining, so as to obtain an enhancement material cluster description vector corresponding to each enhancement material cluster.
Step 302, the description vector of the enhancement material cluster corresponding to the selected enhancement material cluster of the short video image enhancement material library is loaded to the description vector arrangement component corresponding to the selected enhancement material cluster, and the fusion vector data of the enhancement material cluster corresponding to the selected enhancement material cluster is obtained.
Step 303, starting from the upstream enhancement material cluster of the selected enhancement material clusters, wandering a number of enhancement material clusters.
And step 304, loading the fusion vector data of the enhancement material cluster corresponding to the downstream enhancement material cluster of the current walking enhancement material cluster and the description vector of the enhancement material cluster corresponding to the current walking enhancement material cluster to a description vector arrangement component corresponding to the current walking enhancement material cluster to obtain the fusion vector data of the enhancement material cluster corresponding to the current walking enhancement material cluster.
For example, the enhancement material cluster fusion vector data may characterize the upstream and downstream membership and fusion characteristics from a selected enhancement material cluster to a corresponding enhancement material cluster.
And 305, after the plurality of reinforcement clusters are walked, taking the reinforcement cluster fusion vector data corresponding to the top reinforcement cluster in the short video image reinforcement library as an image reinforcement description vector.
In some examples, the output mode of the enhanced material cluster description vector may be an enhanced material cluster feature variable, the output mode of the enhanced material cluster fusion vector data may be an enhanced material cluster aggregation feature variable, and correspondingly, the output mode of the image enhanced material description vector may be an enhanced material feature variable.
Further, the performing material description mining on the short video image enhancement material library through the material description mining model to obtain the enhancement material feature variable of the short video image enhancement material library may include: performing image description mining on the enhancement material information of each enhancement material cluster in the plurality of enhancement material clusters to obtain an enhancement material cluster characteristic variable corresponding to each enhancement material cluster; loading the feature variable of the enhancement material cluster corresponding to the selected enhancement material cluster of the short video image enhancement material library into a description vector sorting component corresponding to the selected enhancement material cluster in a material description mining model to obtain the aggregation feature variable of the enhancement material cluster corresponding to the selected enhancement material cluster; starting from an upstream enhancement material cluster of the selected enhancement material cluster, sequentially (e.g., from top to bottom) walking (e.g., traversing) through a plurality of enhancement material clusters; loading an enhancement material cluster aggregation characteristic variable corresponding to a downstream enhancement material cluster of the current traveling enhancement material cluster and an enhancement material cluster characteristic variable corresponding to the current traveling enhancement material cluster into a description vector arrangement component corresponding to the current traveling enhancement material cluster in a material description mining model to obtain the enhancement material cluster aggregation characteristic variable corresponding to the current traveling enhancement material cluster; and after the plurality of the enhancement material clusters are moved, taking the aggregation characteristic variable of the enhancement material clusters corresponding to the top enhancement material cluster in the short video image enhancement material library as the enhancement material characteristic variable.
The material description mining model in the embodiment of the invention is not limited to the initial material description mining model, and can be a depth residual error model, a full convolution model and the like.
Therefore, the short video image enhancement material library is subjected to material description mining through the material description mining network comprising a plurality of description vector sorting components corresponding to a plurality of enhancement material clusters in the short video image enhancement material library one by one, so that the upstream and downstream subordinate conditions and image detail characteristics of the short video image enhancement material library from the selected enhancement material cluster to the top enhancement material cluster are fused, and the richness and the accuracy of the characteristics of the short video image enhancement material library can be improved.
And 104, performing aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector to obtain a target aggregation description vector.
In the embodiment of the invention, the target aggregate description vector can be a weather subscription service short video image representation obtained by aggregating the image detail characteristics of the weather subscription service short video to be processed with the material relation characteristics and the image detail characteristics of the short video image enhancement material library, and can accurately and completely record the involvement between the weather subscription service short video to be processed and the short video image enhancement material library characteristics.
In some examples, the aggregating operation is performed on the image enhancement material description vector, the weather service page description vector and the short video image description vector, and the obtaining the target aggregate description vector may include step 401 and step 402.
And step 401, loading the image enhancement material description vector and the weather service page description vector into an aggregation sub-model to carry out aggregation operation, so as to obtain a basic aggregation description vector.
Illustratively, the base aggregate description vector may be a weather subscription service short video image representation of a material attribute relationship network of an embedded short video image enhancement material library.
In some examples, the aggregation sub-model may be obtained after the aggregation operation is debugged on the initial aggregation sub-model, where the initial aggregation sub-model may include a linkage analysis component, an interval value mapping component, and a weight processing component, and for example, loading the image enhancement material description vector and the weather service page description vector into the aggregation sub-model to perform the aggregation operation, to obtain the basic aggregation description vector may include steps 501-503.
And step 501, loading the weather service page description vector and the image enhancement material description vector to a linkage analysis component for linkage analysis (association analysis) to obtain a target linkage analysis result.
For example, the target linkage analysis result may represent a correlation coefficient between the weather service page description vector and the image enhancement material description vector.
In some examples, the linkage analysis component may include a correlation coefficient determination algorithm, and the weather service page description vector and the image enhancement material description vector are loaded to the correlation coefficient determination algorithm to determine the correlation coefficient, so as to obtain the target linkage analysis result.
Step 502, loading the target linkage analysis result to the interval value mapping component for interval value mapping processing to obtain the linkage weighting coefficient of the weather service page description vector.
In some examples, the interval value mapping component may include an interval value mapping algorithm, and load the target linkage analysis result to the interval value mapping algorithm to perform interval value mapping calculation, so as to obtain a linkage weighting coefficient. Further, the interval value mapping algorithm may be a softmax algorithm.
Step 503, loading the linkage weighting coefficient and the weather service page description vector to the weight processing component to perform weight-based aggregation operation, so as to obtain a basic aggregation description vector.
Through the embodiment, the aggregation sub-model comprising the linkage analysis component, the interval value mapping component and the weight processing component is used for accurately adding the material attribute relation network of the short video image enhancement material library into the short video image representation of the weather subscription service by mining the involvement conditions between the short video image enhancement material library features of the weather subscription service to be processed, so that the richness and the accuracy of the short video image features of the weather subscription service are improved.
And step 402, loading the basic aggregate description vector, the short video image description vector and the image enhancement material description vector into an enhancement sub-model for feature enhancement processing to obtain a target aggregate description vector.
In some examples, the reinforcement sub-model (splice sub-model) may be obtained after feature reinforcement process debugging of the initial reinforcement sub-model.
Therefore, the basic aggregate description vector, the short video image description vector and the image enhancement material description vector are subjected to characteristic enhancement processing, and the characteristics of the short video of the weather subscription service to be processed and the short video image enhancement material library are further spliced to obtain the target aggregate description vector, so that the short video image enhancement is conveniently performed through the characteristic connection between the short video image of the weather subscription service and the enhancement material in the target aggregate description vector and the layer-order involving condition among the enhancement materials, and the accuracy and pertinence of the short video image enhancement are improved.
And 105, loading the target aggregate description vector into a short video image enhancement model to enhance the short video image, and obtaining an enhanced weather subscription service short video image corresponding to the short video of the weather subscription service to be processed.
In the embodiment of the invention, the enhanced weather subscription service short video image may be a weather subscription service short video image corresponding to the weather subscription service short video to be processed, into which the target enhanced material is integrated, and the target enhanced material may be, by way of example, at least one enhanced material corresponding to the weather subscription service short video to be processed in a plurality of enhanced materials in the short video image enhanced material library.
In some examples, the enhanced weather subscription service short video image may be a weather subscription service VR/AR/MR short video image.
In some examples, the short video image enhancement model may be obtained after short video image enhancement debugging of an initial short video image enhancement model, and illustratively, the initial short video image enhancement model may include fullyconnectiedlayer and Outputlayer.
In some examples, the enhanced weather subscription service short video image may be image data corresponding to any selected enhanced material cluster feature in the short video image enhancement material library, and after the target aggregate description vector is loaded into the short video image enhancement model to perform short video image enhancement, the method may further include: acquiring an enhancement material matching instruction, wherein the enhancement material matching instruction characterizes the projection characteristics of the image data and the multi-layer image data corresponding to the image data; and determining target multi-layer image data corresponding to the enhanced weather subscription service short video image based on the enhanced material matching indication.
In some examples, the image data may include an underlying enhancement material identifier, the multi-layer image data corresponding to the image data may include a multi-layer enhancement material identifier corresponding to the underlying enhancement material identifier, and the corresponding enhancement material matching indication may characterize projection characteristics of the underlying enhancement material identifier and the multi-layer enhancement material identifier corresponding to the underlying enhancement material identifier. For example, the bottom layer enhanced material identifier may be an enhanced material identifier corresponding to any selected enhanced material cluster feature in the short video image enhanced material library, and the multi-layer enhanced material identifier may be a multi-layer enhanced material identifier corresponding to an enhanced material index from the corresponding selected enhanced material cluster feature to the enhanced material of the top layer enhanced material cluster.
In some examples, where the enhanced weather subscription service short video image includes a plurality of bottom layer enhanced material identifications, determining, based on the enhanced material matching indication, target multi-layer image data corresponding to the enhanced weather subscription service short video image may include: determining a plurality of target multi-layer enhanced material identifications corresponding to the plurality of bottom layer enhanced material identifications based on the enhanced material matching indication; after determining the plurality of target multi-layer enhanced material identifications corresponding to the plurality of bottom layer enhanced material identifications based on the enhanced material matching indication, the weather subscription service VR/AR/MR short video image tree may also be generated based on the plurality of target multi-layer enhanced material identifications.
Therefore, the short video image enhancement model is utilized to annotate the enhancement material of the corresponding bottom enhancement material (with highest fine granularity), and then the target multi-layer image data corresponding to the short video image of the enhanced weather subscription service is determined according to the projection characteristics of the multi-layer image data corresponding to the image data, so that the quality of the short video image enhancement is improved.
In the embodiment of the invention, the image description mining model, the material description mining model, the aggregation sub-model, the enhancement sub-model and the short video image enhancement model can be obtained by collaborative debugging of an initial image description mining model, an initial material description mining model, an initial aggregation sub-model, an initial enhancement sub-model and an initial short video image enhancement model.
In some examples, the debugging process for the above-described image description mining model, material description mining model, aggregation sub-model, enhancement sub-model, and short video image enhancement model may be implemented by steps 601-608.
Step 601, acquiring a past weather subscription service short video (weather subscription service short video sample) and a priori enhanced weather subscription service short video image (enhanced weather subscription service short video image sample) corresponding to the past weather subscription service short video.
Before debugging, a debugging sample can be determined, and in an exemplary embodiment of the present invention, a past weather subscription service short video including a priori enhanced weather subscription service short video image can be obtained as the debugging sample.
For example, the prior enhanced weather subscription service short video image may be a preset enhanced material identifier preset for the past weather subscription service short video.
Step 602, loading the past weather subscription service short video into an initial image description mining model to perform image description mining, and obtaining past short video image description vectors (short video image description vector samples) of the past weather subscription service short video and past weather service page description vectors (weather service page description vector samples) of a plurality of past weather service page image blocks of the past weather subscription service short video.
And 603, performing material description mining on the short video image enhancement material library based on the initial material description mining model to obtain past image enhancement material description vectors (image enhancement material description vector samples) of the short video image enhancement material library.
And step 604, loading the past image enhancement material description vector and the past weather service page description vector into an initial aggregation sub-model to perform aggregation operation, and obtaining a past basic aggregation description vector (basic aggregation description vector sample).
Step 605, loading the past basic aggregate description vector, the past short video image description vector and the past image enhancement material description vector into the initial enhancement sub-model for feature enhancement processing, and obtaining a target aggregate description vector example (target aggregate description vector sample).
Step 606, the target aggregate description vector example is loaded to the initial short video image enhancement model to enhance the short video image, and the predicted enhanced weather subscription service short video image corresponding to the past weather subscription service short video is obtained.
In step 607, target model quality assessment data is determined based on the a priori enhanced weather subscription service short video image and the predicted enhanced weather subscription service short video image.
Step 608, debugging the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial reinforcement sub-model and the initial short video image enhancement model based on the target model quality evaluation data to obtain the image description mining model, the material description mining model, the aggregation sub-model, the reinforcement sub-model and the short video image enhancement model.
In some examples, the predictive enhanced weather subscription service short video image may include an enhanced material identification sample of past weather subscription service short videos; the target model quality assessment data may include an enhanced material identification cost (enhanced material identification loss);
further, determining the target model quality assessment data based on the prior enhanced weather subscription service short video image and the predicted enhanced weather subscription service short video image may include: and determining the identification cost of the enhancement material according to the preset enhancement material identification and the enhancement material identification sample.
In some examples, determining the enhancement material identification cost according to the preset enhancement material identification and the enhancement material identification sample may include determining the enhancement material identification cost between the preset enhancement material identification and the enhancement material identification sample based on a preset cost function.
In some examples, the enhancement material identification cost may characterize a distinction between a preset enhancement material identification and an enhancement material identification sample.
In some examples, the preset cost function may include, but is not limited to, a cross entropy cost function, or the like.
In some examples, debugging the initial image description mining model, the initial material description mining model, the initial aggregate sub-model, the initial reinforcement sub-model, and the initial short video image enhancement model based on the target model quality evaluation data, the obtaining the image description mining model, the material description mining model, the aggregate sub-model, the reinforcement sub-model, and the short video image enhancement model may include: updating model variables of an initial image description mining model, an initial material description mining model, an initial aggregation sub-model, an initial enhancement sub-model and an initial short video image enhancement model based on the target model quality evaluation data; based on the updated initial image description mining model, initial material description mining model, initial aggregation sub-model, initial enhancement sub-model and initial short video image enhancement model, jumping to step 602 to short video image enhancement debugging cycle processing of updating model variables of the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial enhancement sub-model and the initial short video image enhancement model based on the target model quality evaluation data, until short video image enhancement debugging requirements are met; and taking the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial reinforcement sub-model and the initial short video image enhancement model which are obtained under the condition of meeting the short video image enhancement debugging requirement as the image description mining model, the material description mining model, the aggregation sub-model, the reinforcement sub-model and the short video image enhancement model.
In some examples, the model variables of updating the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial enhancement sub-model, and the initial short video image enhancement model based on the target model quality evaluation data may include learning rates, and initial values thereof may be adjusted according to actual conditions.
In some examples, the achieving the short video image enhanced debug requirement may be a number of debug cycles up to a preset number of debug cycles. Alternatively, the target model quality evaluation data (target loss function value) may be smaller than the set loss value when the short video image enhancement debugging requirement is met.
Therefore, the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial enhancement sub-model and the initial short video image enhancement model are cooperatively debugged, so that the debugging timeliness can be improved, and the short video image enhancement quality of the model application stage can be ensured.
In some examples, an AI network including the image description mining model, the material description mining model, the aggregation sub-model, the initial enhancement sub-model, and the short video image enhancement model may be generated, and the short video of the weather subscription service to be processed and the short video image enhancement material library are loaded into the AI network to perform short video image enhancement, so as to obtain an enhanced short video image of the weather subscription service corresponding to the short video of the weather subscription service to be processed.
According to the embodiment of the invention, under the task of carrying out short video image enhancement on the short video of the weather subscription service to be processed, the feature analysis of the associated image blocks can be carried out on the short video of the weather subscription service to be processed through the image description mining model comprising the image block disassembly component, the image description mining component and the image description aggregation component, so that the aggregation of the image content features of the short video of the weather subscription service to be processed is realized, and the richness and the accuracy of the image detail features can be further improved. In addition, the short video image enhancement material library can be subjected to material description mining through a material description mining network comprising a plurality of description vector sorting components which are one by one corresponding to a plurality of enhancement material clusters in the short video image enhancement material library, so that the upstream and downstream subordinate conditions and image detail characteristics of the short video image enhancement material library from the selected enhancement material cluster to the top enhancement material cluster are spliced, and the richness and the accuracy of the characteristics of the short video image enhancement material library are improved. And secondly, the short video image enhancement material library and the features of the short video of the weather subscription service to be processed are spliced in two rounds through the aggregation sub-model and the enhancement sub-model, so that the richness and the accuracy of the target aggregation description vector to the image content are further improved, the image enhancement is carried out by utilizing the feature connection between the short video image of the weather subscription service and the enhancement material in the target aggregation description vector and the layer-order involving condition between the enhancement materials, and the enhancement quality of the short video image can be ensured.
In summary, the embodiment of the invention carries out short video image enhancement through the target aggregation description vector obtained by aggregation, and can improve the characteristic enhancement processing quality and pertinence of the short video image of the weather subscription service by utilizing the characteristic connection between the short video image of the weather subscription service and the enhancement material and the layer involvement situation between the enhancement materials in the target aggregation description vector so as to realize the image enhancement processing of the individualized short video of the weather subscription service. Therefore, the content optimization processing of the weather service short video can be realized, the visual understanding degree and pushing persistence of the weather subscription service short video are improved, and the hard and mechanical traditional pushing mode is effectively improved.
On the basis of the above, there is provided an artificial intelligence platform system, comprising:
the data acquisition module is used for acquiring short videos and short video image enhancement material libraries of weather subscription services to be processed;
the first mining module is used for loading the short video of the weather subscription service to be processed into an image description mining model to carry out image description mining to obtain short video image description vectors of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed;
The second mining module is used for carrying out material description mining on the short video image enhancement material library through a material description mining model to obtain an image enhancement material description vector of the short video image enhancement material library, wherein the image enhancement material description vector reflects a material attribute relation network of the short video image enhancement material library;
the vector aggregation module is used for carrying out aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector to obtain a target aggregation description vector;
and the video processing module is used for loading the target aggregate description vector to a short video image enhancement model to enhance the short video image, so as to obtain an enhanced weather subscription service short video image corresponding to the to-be-processed weather subscription service short video.
Preferably, the material description mining model includes a plurality of description vector sorting components corresponding to a plurality of enhancement material clusters in the short video image enhancement material library one by one, and the second mining module performs material description mining on the short video image enhancement material library through the material description mining model, and obtaining the image enhancement material description vector of the short video image enhancement material library includes:
Performing image description mining on the enhancement material information of each enhancement material cluster in the plurality of enhancement material clusters to obtain an enhancement material cluster description vector corresponding to each enhancement material cluster;
loading an enhancement material cluster description vector corresponding to a selected enhancement material cluster of the short video image enhancement material library to a description vector arrangement component corresponding to the selected enhancement material cluster to obtain enhancement material cluster fusion vector data corresponding to the selected enhancement material cluster;
the plurality of enhancement material clusters are walked according to the upstream enhancement material clusters of the selected enhancement material clusters;
loading the fusion vector data of the enhancement material cluster corresponding to the downstream enhancement material cluster of the current traveling enhancement material cluster and the description vector of the enhancement material cluster corresponding to the current traveling enhancement material cluster into a description vector arrangement component corresponding to the current traveling enhancement material cluster to obtain the fusion vector data of the enhancement material cluster corresponding to the current traveling enhancement material cluster;
and after finishing the migration of the plurality of enhancement material clusters, taking the fusion vector data of the enhancement material clusters corresponding to the top enhancement material cluster in the short video image enhancement material library as the image enhancement material description vector.
Preferably, the image description mining model includes an image block disassembly component, an image description mining component and an image description aggregation component, and the first mining module loads the short video of the weather subscription service to be processed into the image description mining model to perform image description mining, and obtaining a short video image description vector of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed includes:
loading the short video of the weather subscription service to be processed into the image block disassembly component to disassemble the image blocks, so as to obtain the image blocks of the plurality of weather service pages;
loading the plurality of weather service page image blocks into the image description mining assembly to perform image description mining to obtain basic image descriptors, distribution feature data and image association features corresponding to the plurality of weather service page image blocks;
and loading the basic image descriptors, the distribution characteristic data and the image association characteristics to the image description aggregation component to perform association processing on the plurality of weather service page image blocks, so as to obtain the short video image description vector and the weather service page description vector.
Preferably, the vector aggregation module performs an aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector, and the obtaining a target aggregate description vector includes:
loading the image enhancement material description vector and the meteorological service page description vector into an aggregation sub-model for aggregation operation to obtain a basic aggregation description vector;
loading the basic aggregate description vector, the short video image description vector and the image enhancement material description vector into an enhancement sub-model for feature enhancement processing to obtain the target aggregate description vector;
the aggregation sub-model comprises a linkage analysis component, an interval numerical mapping component and a weight processing component, wherein the loading of the image enhancement material description vector and the weather service page description vector into the aggregation sub-model for aggregation operation, the obtaining of a basic aggregation description vector comprises the following steps: loading the meteorological service page description vector and the image enhancement material description vector to the linkage analysis component for linkage analysis to obtain a target linkage analysis result; loading the target linkage analysis result to the interval value mapping component for interval value mapping processing to obtain a linkage weighting coefficient of the weather service page description vector; loading the linkage weighting coefficient and the weather service page description vector to the weight processing component to perform weight-based aggregation operation to obtain the basic aggregation description vector;
Wherein, the artificial intelligence platform system also includes a model debugging module for: acquiring a past weather subscription service short video and a priori enhanced weather subscription service short video image corresponding to the past weather subscription service short video; loading the past weather subscription service short video to an initial image description mining model for image description mining to obtain a past short video image description vector of the past weather subscription service short video and past weather service page description vectors of a plurality of past weather service page image blocks of the past weather subscription service short video; performing material description mining on the short video image enhancement material library based on an initial material description mining model to obtain a past image enhancement material description vector of the short video image enhancement material library; loading the past image enhancement material description vector and the past weather service page description vector into an initial aggregation sub-model for aggregation operation to obtain a past basic aggregation description vector; loading the past basic aggregate description vector, the past short video image description vector and the past image enhancement material description vector into an initial enhancement sub-model for feature enhancement processing to obtain a target aggregate description vector example; loading the target aggregate description vector example to an initial short video image enhancement model for short video image enhancement to obtain a predicted enhanced weather subscription service short video image corresponding to the past weather subscription service short video; determining target model quality evaluation data according to the prior enhanced weather subscription service short video image and the predicted enhanced weather subscription service short video image; and debugging the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial enhancement sub-model and the initial short video image enhancement model according to the target model quality evaluation data to obtain the image description mining model, the material description mining model, the aggregation sub-model, the enhancement sub-model and the short video image enhancement model.
Preferably, the enhanced weather subscription service short video image is image data corresponding to any selected enhanced material cluster feature in the short video image enhanced material library, after the target aggregate description vector is loaded to a short video image enhancement model to perform short video image enhancement, the artificial intelligent platform system further comprises an image matching module, configured to:
acquiring an enhancement material matching instruction, wherein the enhancement material matching instruction reflects the image data and the projection characteristics of the multi-layer image data corresponding to the image data;
and determining target multi-layer image data corresponding to the enhanced weather subscription service short video image according to the enhanced material matching instruction.
Based on the above, an artificial intelligence platform system is provided, comprising a processor and a memory in communication with each other, the processor being adapted to retrieve a computer program from the memory and to implement the above-mentioned method by running the computer program.
On the basis of the above, a computer readable storage medium is provided, on which a computer program is stored, which computer program, when run, implements the method described above.
The foregoing is merely exemplary of the present invention and is not intended to limit the present invention. Various modifications and variations of the present invention will be apparent to those skilled in the art. Any modification, equivalent replacement, improvement, etc. which come within the spirit and principles of the invention are to be included in the scope of the claims of the present invention.
Claims (10)
1. An artificial intelligence based weather service short video processing method, which is characterized by being applied to an artificial intelligence platform system, comprising the following steps:
acquiring short videos and short video image enhancement material libraries of weather subscription services to be processed;
loading the short video of the weather subscription service to be processed into an image description mining model for image description mining to obtain short video image description vectors of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed;
performing material description mining on the short video image enhancement material library through a material description mining model to obtain an image enhancement material description vector of the short video image enhancement material library, wherein the image enhancement material description vector reflects a material attribute relation network of the short video image enhancement material library;
Performing aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector to obtain a target aggregation description vector;
and loading the target aggregate description vector to a short video image enhancement model to enhance the short video image, so as to obtain an enhanced weather subscription service short video image corresponding to the to-be-processed weather subscription service short video.
2. The method according to claim 1, wherein the material description mining model includes a plurality of description vector sorting components corresponding to a plurality of enhancement material clusters in the short video image enhancement material library one by one, and the performing material description mining on the short video image enhancement material library by the material description mining model to obtain an image enhancement material description vector of the short video image enhancement material library includes:
performing image description mining on the enhancement material information of each enhancement material cluster in the plurality of enhancement material clusters to obtain an enhancement material cluster description vector corresponding to each enhancement material cluster;
loading an enhancement material cluster description vector corresponding to a selected enhancement material cluster of the short video image enhancement material library to a description vector arrangement component corresponding to the selected enhancement material cluster to obtain enhancement material cluster fusion vector data corresponding to the selected enhancement material cluster;
The plurality of enhancement material clusters are walked according to the upstream enhancement material clusters of the selected enhancement material clusters;
loading the fusion vector data of the enhancement material cluster corresponding to the downstream enhancement material cluster of the current traveling enhancement material cluster and the description vector of the enhancement material cluster corresponding to the current traveling enhancement material cluster into a description vector arrangement component corresponding to the current traveling enhancement material cluster to obtain the fusion vector data of the enhancement material cluster corresponding to the current traveling enhancement material cluster;
and after finishing the migration of the plurality of enhancement material clusters, taking the fusion vector data of the enhancement material clusters corresponding to the top enhancement material cluster in the short video image enhancement material library as the image enhancement material description vector.
3. The method of claim 1, wherein the image description mining model includes an image block disassembly component, an image description mining component, and an image description aggregation component, the loading the short video of the short weather subscription service to be processed into the image description mining model for image description mining, obtaining a short video image description vector of the short video of the short weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed includes:
Loading the short video of the weather subscription service to be processed into the image block disassembly component to disassemble the image blocks, so as to obtain the image blocks of the plurality of weather service pages;
loading the plurality of weather service page image blocks into the image description mining assembly to perform image description mining to obtain basic image descriptors, distribution feature data and image association features corresponding to the plurality of weather service page image blocks;
and loading the basic image descriptors, the distribution characteristic data and the image association characteristics to the image description aggregation component to perform association processing on the plurality of weather service page image blocks, so as to obtain the short video image description vector and the weather service page description vector.
4. The method of claim 1, wherein aggregating the image enhancement material description vector, the weather service page description vector, and the short video image description vector to obtain a target aggregate description vector comprises:
loading the image enhancement material description vector and the meteorological service page description vector into an aggregation sub-model for aggregation operation to obtain a basic aggregation description vector;
Loading the basic aggregate description vector, the short video image description vector and the image enhancement material description vector into an enhancement sub-model for feature enhancement processing to obtain the target aggregate description vector;
the aggregation sub-model comprises a linkage analysis component, an interval numerical mapping component and a weight processing component, wherein the loading of the image enhancement material description vector and the weather service page description vector into the aggregation sub-model for aggregation operation, the obtaining of a basic aggregation description vector comprises the following steps: loading the meteorological service page description vector and the image enhancement material description vector to the linkage analysis component for linkage analysis to obtain a target linkage analysis result; loading the target linkage analysis result to the interval value mapping component for interval value mapping processing to obtain a linkage weighting coefficient of the weather service page description vector; loading the linkage weighting coefficient and the weather service page description vector to the weight processing component to perform weight-based aggregation operation to obtain the basic aggregation description vector;
wherein the method further comprises: acquiring a past weather subscription service short video and a priori enhanced weather subscription service short video image corresponding to the past weather subscription service short video; loading the past weather subscription service short video to an initial image description mining model for image description mining to obtain a past short video image description vector of the past weather subscription service short video and past weather service page description vectors of a plurality of past weather service page image blocks of the past weather subscription service short video; performing material description mining on the short video image enhancement material library based on an initial material description mining model to obtain a past image enhancement material description vector of the short video image enhancement material library; loading the past image enhancement material description vector and the past weather service page description vector into an initial aggregation sub-model for aggregation operation to obtain a past basic aggregation description vector; loading the past basic aggregate description vector, the past short video image description vector and the past image enhancement material description vector into an initial enhancement sub-model for feature enhancement processing to obtain a target aggregate description vector example; loading the target aggregate description vector example to an initial short video image enhancement model for short video image enhancement to obtain a predicted enhanced weather subscription service short video image corresponding to the past weather subscription service short video; determining target model quality evaluation data according to the prior enhanced weather subscription service short video image and the predicted enhanced weather subscription service short video image; and debugging the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial enhancement sub-model and the initial short video image enhancement model according to the target model quality evaluation data to obtain the image description mining model, the material description mining model, the aggregation sub-model, the enhancement sub-model and the short video image enhancement model.
5. The method according to claim 1, wherein the enhanced weather subscription service short video image is image data corresponding to a feature of any selected enhanced material cluster in the short video image enhancement material library, and after the target aggregate description vector is loaded into a short video image enhancement model to perform short video image enhancement, obtaining the enhanced weather subscription service short video image corresponding to the to-be-processed weather subscription service short video, the method further includes:
acquiring an enhancement material matching instruction, wherein the enhancement material matching instruction reflects the image data and the projection characteristics of the multi-layer image data corresponding to the image data;
and determining target multi-layer image data corresponding to the enhanced weather subscription service short video image according to the enhanced material matching instruction.
6. An artificial intelligence platform system, the artificial intelligence platform system comprising:
the data acquisition module is used for acquiring short videos and short video image enhancement material libraries of weather subscription services to be processed;
the first mining module is used for loading the short video of the weather subscription service to be processed into an image description mining model to carry out image description mining to obtain short video image description vectors of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed;
The second mining module is used for carrying out material description mining on the short video image enhancement material library through a material description mining model to obtain an image enhancement material description vector of the short video image enhancement material library, wherein the image enhancement material description vector reflects a material attribute relation network of the short video image enhancement material library;
the vector aggregation module is used for carrying out aggregation operation on the image enhancement material description vector, the weather service page description vector and the short video image description vector to obtain a target aggregation description vector;
and the video processing module is used for loading the target aggregate description vector to a short video image enhancement model to enhance the short video image, so as to obtain an enhanced weather subscription service short video image corresponding to the to-be-processed weather subscription service short video.
7. The artificial intelligence platform system according to claim 6, wherein the material description mining model includes a plurality of description vector sorting components corresponding one by one to a plurality of enhancement material clusters in the short video image enhancement material library, and the second mining module performs material description mining on the short video image enhancement material library through the material description mining model, and obtaining the image enhancement material description vector of the short video image enhancement material library includes:
Performing image description mining on the enhancement material information of each enhancement material cluster in the plurality of enhancement material clusters to obtain an enhancement material cluster description vector corresponding to each enhancement material cluster;
loading an enhancement material cluster description vector corresponding to a selected enhancement material cluster of the short video image enhancement material library to a description vector arrangement component corresponding to the selected enhancement material cluster to obtain enhancement material cluster fusion vector data corresponding to the selected enhancement material cluster;
the plurality of enhancement material clusters are walked according to the upstream enhancement material clusters of the selected enhancement material clusters;
loading the fusion vector data of the enhancement material cluster corresponding to the downstream enhancement material cluster of the current traveling enhancement material cluster and the description vector of the enhancement material cluster corresponding to the current traveling enhancement material cluster into a description vector arrangement component corresponding to the current traveling enhancement material cluster to obtain the fusion vector data of the enhancement material cluster corresponding to the current traveling enhancement material cluster;
and after finishing the migration of the plurality of enhancement material clusters, taking the fusion vector data of the enhancement material clusters corresponding to the top enhancement material cluster in the short video image enhancement material library as the image enhancement material description vector.
8. The artificial intelligence platform system of claim 6, wherein the image description mining model comprises an image block disassembly component, an image description mining component, and an image description aggregation component, wherein the first mining module loads the short video of the weather subscription service to be processed into the image description mining model for image description mining, and obtaining short video image description vectors of the short video of the weather subscription service to be processed and weather service page description vectors of a plurality of weather service page image blocks of the short video of the weather subscription service to be processed comprises:
loading the short video of the weather subscription service to be processed into the image block disassembly component to disassemble the image blocks, so as to obtain the image blocks of the plurality of weather service pages;
loading the plurality of weather service page image blocks into the image description mining assembly to perform image description mining to obtain basic image descriptors, distribution feature data and image association features corresponding to the plurality of weather service page image blocks;
and loading the basic image descriptors, the distribution characteristic data and the image association characteristics to the image description aggregation component to perform association processing on the plurality of weather service page image blocks, so as to obtain the short video image description vector and the weather service page description vector.
9. The artificial intelligence platform system of claim 6, wherein the vector aggregation module aggregates the image enhancement material description vector, the weather service page description vector, and the short video image description vector to obtain a target aggregate description vector comprising:
loading the image enhancement material description vector and the meteorological service page description vector into an aggregation sub-model for aggregation operation to obtain a basic aggregation description vector;
loading the basic aggregate description vector, the short video image description vector and the image enhancement material description vector into an enhancement sub-model for feature enhancement processing to obtain the target aggregate description vector;
the aggregation sub-model comprises a linkage analysis component, an interval numerical mapping component and a weight processing component, wherein the loading of the image enhancement material description vector and the weather service page description vector into the aggregation sub-model for aggregation operation, the obtaining of a basic aggregation description vector comprises the following steps: loading the meteorological service page description vector and the image enhancement material description vector to the linkage analysis component for linkage analysis to obtain a target linkage analysis result; loading the target linkage analysis result to the interval value mapping component for interval value mapping processing to obtain a linkage weighting coefficient of the weather service page description vector; loading the linkage weighting coefficient and the weather service page description vector to the weight processing component to perform weight-based aggregation operation to obtain the basic aggregation description vector;
Wherein, the artificial intelligence platform system also includes a model debugging module for: acquiring a past weather subscription service short video and a priori enhanced weather subscription service short video image corresponding to the past weather subscription service short video; loading the past weather subscription service short video to an initial image description mining model for image description mining to obtain a past short video image description vector of the past weather subscription service short video and past weather service page description vectors of a plurality of past weather service page image blocks of the past weather subscription service short video; performing material description mining on the short video image enhancement material library based on an initial material description mining model to obtain a past image enhancement material description vector of the short video image enhancement material library; loading the past image enhancement material description vector and the past weather service page description vector into an initial aggregation sub-model for aggregation operation to obtain a past basic aggregation description vector; loading the past basic aggregate description vector, the past short video image description vector and the past image enhancement material description vector into an initial enhancement sub-model for feature enhancement processing to obtain a target aggregate description vector example; loading the target aggregate description vector example to an initial short video image enhancement model for short video image enhancement to obtain a predicted enhanced weather subscription service short video image corresponding to the past weather subscription service short video; determining target model quality evaluation data according to the prior enhanced weather subscription service short video image and the predicted enhanced weather subscription service short video image; and debugging the initial image description mining model, the initial material description mining model, the initial aggregation sub-model, the initial enhancement sub-model and the initial short video image enhancement model according to the target model quality evaluation data to obtain the image description mining model, the material description mining model, the aggregation sub-model, the enhancement sub-model and the short video image enhancement model.
10. The artificial intelligence platform system according to claim 6, wherein the enhanced weather subscription service short video image is image data corresponding to a cluster feature of any selected enhanced material in the short video image enhanced material library, and after the target aggregate description vector is loaded into a short video image enhancement model to perform short video image enhancement, the artificial intelligence platform system further comprises an image matching module for:
acquiring an enhancement material matching instruction, wherein the enhancement material matching instruction reflects the image data and the projection characteristics of the multi-layer image data corresponding to the image data;
and determining target multi-layer image data corresponding to the enhanced weather subscription service short video image according to the enhanced material matching instruction.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310911134.8A CN117082309B (en) | 2023-07-24 | 2023-07-24 | Meteorological service short video processing method and platform system based on artificial intelligence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310911134.8A CN117082309B (en) | 2023-07-24 | 2023-07-24 | Meteorological service short video processing method and platform system based on artificial intelligence |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117082309A true CN117082309A (en) | 2023-11-17 |
| CN117082309B CN117082309B (en) | 2024-05-03 |
Family
ID=88718452
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310911134.8A Active CN117082309B (en) | 2023-07-24 | 2023-07-24 | Meteorological service short video processing method and platform system based on artificial intelligence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117082309B (en) |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018187622A1 (en) * | 2017-04-05 | 2018-10-11 | Lyrical Labs Holdings, Llc | Video processing and encoding |
| CN111078942A (en) * | 2019-12-18 | 2020-04-28 | 腾讯科技（深圳）有限公司 | Method, device and storage medium for recommending videos |
| US20200314356A1 (en) * | 2019-03-29 | 2020-10-01 | Nathaniel Webster Storer | Optimized video review using motion recap images |
| CN111767796A (en) * | 2020-05-29 | 2020-10-13 | 北京奇艺世纪科技有限公司 | Video association method, device, server and readable storage medium |
| CN112330573A (en) * | 2020-12-01 | 2021-02-05 | 厦门美图之家科技有限公司 | Portrait-based image repairing method and device, electronic equipment and storage medium |
| CN113034412A (en) * | 2021-02-25 | 2021-06-25 | 北京达佳互联信息技术有限公司 | Video processing method and device |
| CN115879547A (en) * | 2022-12-14 | 2023-03-31 | 山东省计算中心（国家超级计算济南中心） | Open world knowledge graph complementing method and system based on LSTM and attention mechanism |
| CN115914774A (en) * | 2022-10-31 | 2023-04-04 | 华能伊敏煤电有限责任公司 | Associated short video generation method using alarm information |
-
2023
- 2023-07-24 CN CN202310911134.8A patent/CN117082309B/en active Active
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2018187622A1 (en) * | 2017-04-05 | 2018-10-11 | Lyrical Labs Holdings, Llc | Video processing and encoding |
| US20200314356A1 (en) * | 2019-03-29 | 2020-10-01 | Nathaniel Webster Storer | Optimized video review using motion recap images |
| CN111078942A (en) * | 2019-12-18 | 2020-04-28 | 腾讯科技（深圳）有限公司 | Method, device and storage medium for recommending videos |
| CN111767796A (en) * | 2020-05-29 | 2020-10-13 | 北京奇艺世纪科技有限公司 | Video association method, device, server and readable storage medium |
| CN112330573A (en) * | 2020-12-01 | 2021-02-05 | 厦门美图之家科技有限公司 | Portrait-based image repairing method and device, electronic equipment and storage medium |
| CN113034412A (en) * | 2021-02-25 | 2021-06-25 | 北京达佳互联信息技术有限公司 | Video processing method and device |
| CN115914774A (en) * | 2022-10-31 | 2023-04-04 | 华能伊敏煤电有限责任公司 | Associated short video generation method using alarm information |
| CN115879547A (en) * | 2022-12-14 | 2023-03-31 | 山东省计算中心（国家超级计算济南中心） | Open world knowledge graph complementing method and system based on LSTM and attention mechanism |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117082309B (en) | 2024-05-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109670267B (en) | Data processing method and device | |
| CN102708130B (en) | Calculate the easily extensible engine that fine point of user is mated for offer | |
| CN109086071A (en) | A kind of method and server of management software version information | |
| CN105138601B (en) | A kind of graphic mode matching method for supporting fuzzy constraint relationship | |
| Lin et al. | Semantic classification and hash code accelerated detection of design changes in BIM models | |
| CN108540351B (en) | Automatic testing method for distributed big data service | |
| CN103026334A (en) | Data classification | |
| Lin et al. | Data driven water pipe failure prediction: A bayesian nonparametric approach | |
| KR102668394B1 (en) | System, server and method for providing cooperation solution among multiple workers | |
| CN112860736A (en) | Big data query optimization method and device and readable storage medium | |
| CN109285024A (en) | Online feature determines method, apparatus, electronic equipment and storage medium | |
| CN111311105A (en) | Combined product scoring method, device, equipment and readable storage medium | |
| CN117082309B (en) | Meteorological service short video processing method and platform system based on artificial intelligence | |
| WO2022146433A1 (en) | Smart data warehouse for cloud-based reservoir simulation | |
| JP2003256757A (en) | Space data analysis device, space data analysis method, and space data analysis program | |
| CN116415957A (en) | Abnormal transaction object identification method, device, computer equipment and storage medium | |
| US11227288B1 (en) | Systems and methods for integration of disparate data feeds for unified data monitoring | |
| CN115169578A (en) | AI model production method and system based on meta-space data markers | |
| CN114996228A (en) | Server-unaware-oriented data transmission cost optimization method | |
| CN110399337A (en) | File automating method of servicing and system based on data-driven | |
| CN111985868A (en) | Bill of material maintenance method and system for server product production | |
| CN114548463A (en) | Line information prediction method, line information prediction device, computer equipment and storage medium | |
| CN111414538A (en) | Text recommendation method and device based on artificial intelligence and electronic equipment | |
| CN112529207A (en) | Model optimization method, device, storage medium and equipment | |
| KR102639134B1 (en) | System for providing cooperation service for multiple workers performing a poject |
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 |