CN119046481B - Multimedia video stream management system and method based on artificial intelligence - Google Patents

Multimedia video stream management system and method based on artificial intelligence Download PDF

Info

Publication number
CN119046481B
CN119046481B CN202411544997.7A CN202411544997A CN119046481B CN 119046481 B CN119046481 B CN 119046481B CN 202411544997 A CN202411544997 A CN 202411544997A CN 119046481 B CN119046481 B CN 119046481B
Authority
CN
China
Prior art keywords
video
video stream
element information
module
tag
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202411544997.7A
Other languages
Chinese (zh)
Other versions
CN119046481A (en
Inventor
易小武
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Rong'an Cloud Network Beijing Technology Co ltd
Original Assignee
Rong'an Cloud Network Beijing Technology Co ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Rong'an Cloud Network Beijing Technology Co ltd filed Critical Rong'an Cloud Network Beijing Technology Co ltd
Priority to CN202411544997.7A priority Critical patent/CN119046481B/en
Publication of CN119046481A publication Critical patent/CN119046481A/en
Application granted granted Critical
Publication of CN119046481B publication Critical patent/CN119046481B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
    • G06F16/43—Querying
    • G06F16/435—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
    • G06F16/45—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/40—Information retrieval; Database structures therefor; File system structures therefor of multimedia data, e.g. slideshows comprising image and additional audio data
    • G06F16/48—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
    • G06F16/73—Querying
    • G06F16/735—Filtering based on additional data, e.g. user or group profiles
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
    • G06F16/75—Clustering; Classification
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/70—Information retrieval; Database structures therefor; File system structures therefor of video data
    • G06F16/78—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually
    • G06F16/7867—Retrieval characterised by using metadata, e.g. metadata not derived from the content or metadata generated manually using information manually generated, e.g. tags, keywords, comments, title and artist information, manually generated time, location and usage information, user ratings

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Multimedia (AREA)
  • Data Mining & Analysis (AREA)
  • Databases & Information Systems (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Library & Information Science (AREA)
  • Computational Linguistics (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)

Abstract

本发明公开了一种基于人工智能的多媒体视频流管理系统及方法，涉及视频流标签处理技术领域，本发明通过梳理AI模块在对任意视频流进行视频标签种类划分时所依据的参照要素信息，对AI模块在进行区别分类时存在邻近判断的现象基于数据分析提取得到若干邻近标签，本发明通过监测AI模块在互为邻近标签的视频标签范围内输出判断的分类结果是否存在数据偏差，从而为是否需要对AI模块进行优化提供判断依据，本发明能实现对AI模块在多媒体视频流标签处理领域内的性能进行自适应监测，有效提高了AI在多媒体视频内容的可发现性和用户体验上的作用。

The present invention discloses a multimedia video stream management system and method based on artificial intelligence, and relates to the technical field of video stream label processing. The present invention combs the reference element information based on which the AI module divides the video label types of any video stream, and extracts a number of neighboring labels based on data analysis for the phenomenon of neighboring judgment when the AI module performs differential classification. The present invention monitors whether there is data deviation in the classification result output by the AI module within the range of video labels that are mutually neighboring labels, thereby providing a judgment basis for whether the AI module needs to be optimized. The present invention can realize adaptive monitoring of the performance of the AI module in the field of multimedia video stream label processing, and effectively improves the role of AI in the discoverability of multimedia video content and user experience.

Description

Multimedia video stream management system and method based on artificial intelligence
Technical Field
The invention relates to the technical field of video stream label processing, in particular to a multimedia video stream management system and method based on artificial intelligence.
Background
‌ AI in the multimedia video stream tagging ‌ refers to a process of automatically tagging and classifying video content using artificial intelligence and machine learning algorithms, specifically, AI in the multimedia video stream tagging process, firstly analyzes video content through computer vision and deep learning techniques, identifies key elements (such as characters, objects, scenes, etc.) in video, and then generates tags according to these elements. These tags may cover aspects of the content description, emotional color, topic classification, etc. of the video.
AI is very widely used in the field of tagging of multimedia video streams. Firstly, in the aspect of personalized recommendation, by analyzing video content to generate a label and combining interest preference of a user, accurate personalized recommendation can be realized. Secondly, in the field of content auditing, the AI can help to detect illegal contents, and ensure the safety and compliance of the platform. Finally, in the aspect of media asset management, the automatically generated labels can greatly improve the retrieval efficiency of the media assets, and facilitate users to quickly find the required video data. For example, the repeated or similar segments in the video can be identified and searched by using the AI technology, and the method is suitable for scenes such as original identification, video duplicate checking and the like.
Therefore, it can be known that the accuracy of determining the attribute tag of the AI on the video stream directly affects the effective utilization of the AI processing result in reality, but some data deviation problems often occur in the process of performing the AI on the video stream, mainly because the data set used in training the AI model has an imbalance or incomplete condition in a specific aspect, so that the model has an error on the prediction result of some specific groups or conditions, and the imbalance or incomplete condition is a representation of lack of diversity of sample data.
Disclosure of Invention
The invention aims to provide a multimedia video stream management system and method based on artificial intelligence, which are used for solving the problems in the prior art.
In order to achieve the above purpose, the invention provides a multimedia video stream management method based on artificial intelligence, which comprises the following steps:
step S1, collecting process data generated when each video stream segment received from a multimedia port generates a corresponding video tag by calling an AI module on a multimedia video stream management cloud platform, and combing reference element information according to which the AI module classifies the video tag type of any video stream segment;
s2, judging and identifying any two video tags with adjacent judgment when the AI module is subjected to differential classification by comparing deviation distribution conditions of reference element information presented between any two different video tags, and extracting to obtain a plurality of pairs of video tag groups which are mutually adjacent tags;
Step S3, extracting distribution conditions of the presented distinguishing reference element information in the process of judging the video labels of the corresponding video stream fragments within the range of the corresponding mutually adjacent labels by the AI module, and calculating the mutually adjacent degree value of each pair of mutually adjacent label video label groups;
and S4, monitoring calling rate distribution of the video stream fragments stored in each storage area by a user in the multimedia video stream management cloud platform, adjusting video labels of the video stream fragments with abnormal calling rates by referring to corresponding adjacent labels, and judging whether to send early warning prompts needing to optimize the AI module according to corresponding calling rate change conditions brought by the video labels after adjustment.
Preferably, step S1 includes:
S1-1, acquiring a multimedia video stream management cloud platform, and before an AI module is called to generate a corresponding video tag for each video stream segment received from a multimedia port, acquiring all feature element information extracted from each video stream segment after image identification and audio analysis, and collecting and generating a feature element information set corresponding to each video stream segment;
As can be seen from the above, the feature element information set corresponding to each video stream segment is the information set that ultimately causes the AI module to analyze the information set that is referenced when generating a certain exact video tag for that video stream segment;
Step S1-2, feature element information sets of all video stream fragments which are stored in a multimedia video stream management cloud platform and have the same corresponding video labels are respectively collected to respectively obtain a plurality of reference element information sets according to which an AI module performs video label type classification on any video stream fragment, wherein the video stream fragments with the same video labels are placed in the same storage area in the multimedia video stream management cloud platform for data storage, and one reference element information set corresponds to one video label.
Preferably, the step S2 comprises the following steps:
S2-1, extracting overlapped characteristic element information from reference element information sets corresponding to any two different video tags one by one to respectively generate overlapped reference element information sets corresponding to any two different video tags;
And step S2-2, acquiring the total value of the characteristic element information contained in each superposition reference element information set corresponding to each video tag and other types of video tags, and judging that the AI module has adjacent judgment when classifying any two different video tags, and judging that the any two different video tags are a pair of video tag groups mutually adjacent tags if the total value of the characteristic element information contained in the superposition reference element information set between any two different video tags is larger than a total threshold value.
From the above, it can be seen that, when a certain video stream segment is determined to be a specific video tag of the mutually adjacent tags, the AI module determines the final output tag division result by using relatively tiny distinguishing feature elements, that is, when a video stream segment containing the corresponding overlapping feature elements is determined to be a specific type of the mutually adjacent tags, the AI module has a higher requirement on identification accuracy, and if the AI module has identification data deviation, an error is more likely to occur when the AI module performs distinguishing classification between the mutually adjacent video tags.
Preferably, step S3 includes:
Step S3-1, extracting a reference element information set P (A) of the video tag A, a reference element information set P (B) of the video tag B and a superposition reference element information set U a,b between the video tag A and the video tag B if the video tag A and the video tag B are adjacent tags, extracting a difference reference element information set C 1=P(A)-Ua,b from the video tag A, and extracting a difference reference element information set C 2=P(B)-Ua,b from the video tag B;
The above-mentioned differential reference element information sets obtained by obtaining the video tag a and the video tag B respectively are often video stream segments of overlapping reference element information of the video tag a and the video tag B included in the corresponding feature element information sets by the AI module, and in determining whether the video stream segment is finally the video tag a or the video tag B, feature element information for narrowing the matching degree with the video tag a or the video tag B is used, so to speak, feature element information included in the overlapping reference element information set U a,b represents common element information between the video tag a and the video tag B, feature element information included in the differential reference element information set obtained by obtaining the video tag a represents individual element information of the video tag a, and feature element information included in the differential reference element information set obtained by obtaining the video tag B represents individual element information of the video tag B;
Step S3-2, collecting all video stream fragments of marked video labels A from a multimedia video stream management cloud platform to obtain a first video stream fragment set Y1, collecting all video stream fragments of marked video labels B to obtain a second video stream fragment set Y2, and if a certain video stream fragment exists in the first video stream fragment set Y1 or the second video stream fragment set Y2, and a characteristic element information set R extracted from a certain video stream fragment satisfies R and U a,b = Q not equal to R and U35 = Q not equal to Q @ Wherein Q represents an intersection between the feature element information set R and the superposition reference element information set U a,b, and a certain video stream segment is used as a feature marker, and a target distinguishing element information set Q' =r-Q is extracted from a certain video stream segment;
S3-3, collecting a target distinguishing element information set extracted from all video stream fragments subjected to characteristic marking in a first video stream fragment set Y1, collecting the types eta 1 of the accumulated characteristic element information, collecting a target distinguishing element information set extracted from all video stream fragments subjected to characteristic marking in a second video stream fragment set Y2, collecting the types eta 2 of the accumulated characteristic element information, and calculating to obtain a first adjacent index beta 1=[η1/card(C1)+η2/card(C2)/2 between a video tag A and a video tag B;
η 1/card(C1) or η 2/card(C2) that indicates that when a video stream segment containing overlapping reference element information is determined as a video tag a or a video tag B, the distribution of the related distinguishing reference element information is more uniform, that is, the AI module makes a distinction proximity division less difficult in the process of excluding a conclusion belonging to the video tag B or the video tag a based on the distinguishing reference element information of the video tag a or the video tag B, that is, the video stream segment having the distinguishing reference element information of the video tag a or the video tag B is sufficiently separated from a video feature gap between the video stream segment corresponding to the video tag B or the video tag a, that is, the AI module makes a difference in the actual proximity degree on the data determination analysis level when the video stream segment operates the classification of the video tag a and the video tag B, that is, the AI module makes a distinction classification between the video tag a and the video tag B less likely to occur;
Step S3-4, obtaining a ratio alpha 1 of the number of video stream fragments which are marked by the features in the first video stream fragment set Y1 and a ratio alpha 2 of the number of video stream fragments which are marked by the features in the second video stream fragment set Y2, and calculating to obtain a second adjacent index beta 2 = (alpha 1+ alpha 2)/2 between the video tag A and the video tag B;
the higher the ratio alpha 1 or the ratio alpha 2, the higher the ratio of the video stream fragments containing the superposition reference element information between the video label A and the video label B in the video stream fragments judged to correspond to the video label A or the video label B, and the more frequently the phenomenon that the AI module distinguishes the adjacent division between the video label A and the video label B occurs;
Step S3-5, performing similarity calculation on the feature element information sets corresponding to the video stream fragments in the first video stream fragment set Y1 and the feature element information sets corresponding to the video stream fragments in the second video stream fragment set Y2 one by one, capturing the highest similarity value delta, and calculating the proximity value zeta= (1/beta 1+β2) x delta between the video tag A and the video tag B.
Preferably, step S4 includes:
Step S4-1, monitoring the average calling rate of a user to all stored video stream fragments in each storage area, if a certain video stream fragment with the corresponding calling rate lower than the average calling rate exists in a certain storage area corresponding to the video tag a, capturing a video tag b with the highest adjacent degree value with the video tag a, and adjusting the video tag of the certain video stream fragment from a to b;
Step S4-2, monitoring the calling rate of a user for presenting a certain video stream segment in unit period time after the video label of the certain video stream segment is adjusted from a to b, generating an early warning signal optimized for an AI module if the calling rate is increased compared with the value presented before adjustment, and canceling adjustment if the calling rate is increased compared with the value not presented before adjustment;
And S4-3, when the accumulated times of the early warning signals generated by optimizing the AI module are larger than a time threshold, sending an early warning prompt for optimizing the AI module to the management terminal.
In order to better realize the method, the system also provides a multimedia video stream management system, which comprises a video stream tag processing data management module, a proximity tag judgment management module, a proximity value calculation management module and an AI module optimization prompt management module;
The video stream tag processing data management module is used for collecting process data generated when each video stream segment received from the multimedia port generates a corresponding video tag by calling the AI module for the multimedia video stream management cloud platform, and combing reference element information according to which the AI module carries out video tag type classification on any video stream segment;
The adjacent tag judgment management module is used for judging and identifying any two video tags with adjacent judgment when the AI module is used for distinguishing and classifying by comparing deviation distribution conditions of the reference element information presented between any two different video tags, and extracting a plurality of pairs of video tag groups which are mutually adjacent tags;
The adjacent degree value calculation management module is used for extracting the distribution condition of the presented distinguishing reference element information in the process of judging the video labels of the corresponding video stream fragments within the range of the corresponding mutually adjacent labels by the AI module, and calculating the mutually adjacent degree value of each pair of mutually adjacent label video label groups;
and the AI module optimization prompt management module is used for monitoring the calling rate distribution of the video stream fragments stored in each storage area by a user in the multimedia video stream management cloud platform, carrying out video tag adjustment on the video stream fragments with abnormal calling rate by referring to corresponding adjacent tags, and judging whether to send an early warning prompt for optimizing the AI module according to the corresponding calling rate change condition brought by the video tag adjustment.
Preferably, the adjacent label judging and managing module comprises a characteristic element information carding unit and an adjacent label judging unit;
The characteristic element information carding unit is used for comparing deviation distribution conditions of the reference element information presented between any two different video tags;
The adjacent label judging unit is used for judging and identifying any two video labels with adjacent judgment when the AI module is used for distinguishing and classifying, and extracting a plurality of pairs of video label groups which are adjacent labels.
Preferably, the AI module optimization prompt management module comprises a label adjustment monitoring management unit and an optimization early warning prompt management unit;
The label adjustment monitoring management unit is used for monitoring the calling rate distribution of the video stream fragments stored in each storage area by a user in the multimedia video stream management cloud platform, and performing video label adjustment on the video stream fragments with abnormal calling rate by referring to corresponding adjacent labels;
And the optimized early warning prompt management unit is used for judging whether to send an early warning prompt which needs to be optimized to the AI module according to the corresponding calling rate change condition brought by the video tag after adjustment.
Compared with the prior art, the invention has the beneficial effects that: the invention obtains a plurality of adjacent labels based on data analysis and extraction by combing the reference element information according to which the AI module carries out video label type classification on any video stream, the phenomenon of adjacent judgment exists when the AI module carries out differential classification, the invention outputs whether the classification result of judgment has data deviation or not in the range of the video labels which are mutually adjacent labels by monitoring the AI module, therefore, a judgment basis is provided for whether the AI module needs to be optimized, the self-adaptive monitoring of the performance of the AI module in the field of multimedia video stream label processing can be realized, and the discoverability of the AI in the multimedia video content and the effect of the AI in user experience are effectively improved.
Drawings
Fig. 1 is a schematic flow chart of a multimedia video stream management method based on artificial intelligence.
Fig. 2 is a schematic structural diagram of an artificial intelligence-based multimedia video stream management system according to the present invention.
Detailed Description
All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, are intended to be within the scope of the invention.
1-2, The invention provides a technical scheme, namely a multimedia video stream management method based on artificial intelligence, which comprises the following steps:
step S1, collecting process data generated when each video stream segment received from a multimedia port generates a corresponding video tag by calling an AI module on a multimedia video stream management cloud platform, and combing reference element information according to which the AI module classifies the video tag type of any video stream segment;
wherein, step S1 includes:
S1-1, acquiring a multimedia video stream management cloud platform, and before an AI module is called to generate a corresponding video tag for each video stream segment received from a multimedia port, acquiring all feature element information extracted from each video stream segment after image identification and audio analysis, and collecting and generating a feature element information set corresponding to each video stream segment;
S1-2, respectively collecting characteristic element information sets of all video stream fragments which are stored in a multimedia video stream management cloud platform and have the same corresponding video labels, respectively obtaining a plurality of reference element information sets according to which an AI module performs video label type classification on any video stream fragment;
s2, judging and identifying any two video tags with adjacent judgment when the AI module is subjected to differential classification by comparing deviation distribution conditions of reference element information presented between any two different video tags, and extracting to obtain a plurality of pairs of video tag groups which are mutually adjacent tags;
preferably, the step S2 comprises the following steps:
S2-1, extracting overlapped characteristic element information from reference element information sets corresponding to any two different video tags one by one to respectively generate overlapped reference element information sets corresponding to any two different video tags;
for example, the reference element information set corresponding to the first video tag includes { feature element information r1, feature element information r2, feature element information r4, feature element information r7};
in summary, the feature element information overlapped between the first video tag and the second video tag comprises feature element information r1 and feature element information r2, so that the overlapped reference element information set of the first video tag and the second video tag is { feature element information r1 and feature element information r2};
Step S2-2, acquiring the total value of characteristic element information contained in each superposition reference element information set corresponding to each video tag and other types of video tags, judging that an AI module has adjacent judgment when distinguishing and classifying any two types of different video tags if the total value of the characteristic element information contained in the superposition reference element information set between any two types of different video tags is larger than a total threshold value, and judging that the any two types of different video tags are a pair of video tag groups mutually adjacent to each other;
Step S3, extracting distribution conditions of the presented distinguishing reference element information in the process of judging the video labels of the corresponding video stream fragments within the range of the corresponding mutually adjacent labels by the AI module, and calculating the mutually adjacent degree value of each pair of mutually adjacent label video label groups;
wherein, step S3 includes:
Step S3-1, extracting a reference element information set P (A) of the video tag A, a reference element information set P (B) of the video tag B and a superposition reference element information set U a,b between the video tag A and the video tag B if the video tag A and the video tag B are adjacent tags, extracting a difference reference element information set C 1=P(A)-Ua,b from the video tag A, and extracting a difference reference element information set C 2=P(B)-Ua,b from the video tag B;
Step S3-2, collecting all video stream fragments of marked video labels A from a multimedia video stream management cloud platform to obtain a first video stream fragment set Y1, collecting all video stream fragments of marked video labels B to obtain a second video stream fragment set Y2, and if a certain video stream fragment exists in the first video stream fragment set Y1 or the second video stream fragment set Y2, and a characteristic element information set R extracted from a certain video stream fragment satisfies R and U a,b = Q not equal to R and U35 = Q not equal to Q @ Wherein Q represents an intersection between the feature element information set R and the superposition reference element information set U a,b, and a certain video stream segment is used as a feature marker, and a target distinguishing element information set Q' =r-Q is extracted from a certain video stream segment;
S3-3, collecting a target distinguishing element information set extracted from all video stream fragments subjected to characteristic marking in a first video stream fragment set Y1, collecting the types eta 1 of the accumulated characteristic element information, collecting a target distinguishing element information set extracted from all video stream fragments subjected to characteristic marking in a second video stream fragment set Y2, collecting the types eta 2 of the accumulated characteristic element information, and calculating to obtain a first adjacent index beta 1=[η1/card(C1)+η2/card(C2)/2 between a video tag A and a video tag B;
Step S3-4, obtaining a ratio alpha 1 of the number of video stream fragments which are marked by the features in the first video stream fragment set Y1 and a ratio alpha 2 of the number of video stream fragments which are marked by the features in the second video stream fragment set Y2, and calculating to obtain a second adjacent index beta 2 = (alpha 1+ alpha 2)/2 between the video tag A and the video tag B;
the higher the ratio alpha 1 or the ratio alpha 2, the higher the ratio of the video stream fragments containing the superposition reference element information between the video label A and the video label B in the video stream fragments judged to correspond to the video label A or the video label B, and the more frequently the phenomenon that the AI module distinguishes the adjacent division between the video label A and the video label B occurs;
Step S3-5, performing similarity calculation on the feature element information sets corresponding to the video stream fragments in the first video stream fragment set Y1 and the feature element information sets corresponding to the video stream fragments in the second video stream fragment set Y2 one by one, capturing the highest similarity value delta, and calculating the proximity value zeta= (1/beta 1+β2) x delta between the video tag A and the video tag B.
For example, the first video stream segment set Y1 includes a video stream segment w1, a video stream segment w2, and a video stream segment w3;
The second video stream segment set Y2 comprises a video stream segment d1 and a video stream segment d2;
the similarity between the feature element information set corresponding to the video stream segment w1 and the feature element information set corresponding to the video stream segment d1 is Q1;
the similarity between the characteristic element information set corresponding to the video stream fragment w1 and the characteristic element information set corresponding to the video stream fragment d2 is Q2;
the similarity between the characteristic element information set corresponding to the video stream fragment w2 and the characteristic element information set corresponding to the video stream fragment d1 is Q3;
the similarity between the characteristic element information set corresponding to the video stream fragment w2 and the characteristic element information set corresponding to the video stream fragment d2 is Q4;
the similarity between the characteristic element information set corresponding to the video stream fragment w3 and the characteristic element information set corresponding to the video stream fragment d1 is Q5;
the similarity between the characteristic element information set corresponding to the video stream fragment w3 and the characteristic element information set corresponding to the video stream fragment d2 is Q6;
if Q6> Q3> Q2> Q4> Q61> Q5 is satisfied, the highest captured similarity value δ=q6;
Step S4, monitoring calling rate distribution of a user on video stream fragments stored in each storage area in a multimedia video stream management cloud platform, adjusting video labels of the video stream fragments with abnormal calling rates by referring to corresponding adjacent labels, and judging whether to send early warning prompts needing to optimize an AI module according to corresponding calling rate change conditions brought by the video labels after adjustment;
Wherein, step S4 includes:
Step S4-1, monitoring the average calling rate of a user to all stored video stream fragments in each storage area, if a certain video stream fragment with the corresponding calling rate lower than the average calling rate exists in a certain storage area corresponding to the video tag a, capturing a video tag b with the highest adjacent degree value with the video tag a, and adjusting the video tag of the certain video stream fragment from a to b;
Step S4-2, monitoring the calling rate of a user for presenting a certain video stream segment in unit period time after the video label of the certain video stream segment is adjusted from a to b, generating an early warning signal optimized for an AI module if the calling rate is increased compared with the value presented before adjustment, and canceling adjustment if the calling rate is increased compared with the value not presented before adjustment;
And S4-3, when the accumulated times of the early warning signals generated by optimizing the AI module are larger than a time threshold, sending an early warning prompt for optimizing the AI module to the management terminal.
In order to better realize the method, the system also provides a multimedia video stream management system, which comprises a video stream tag processing data management module, a proximity tag judgment management module, a proximity value calculation management module and an AI module optimization prompt management module;
The video stream tag processing data management module is used for collecting process data generated when each video stream segment received from the multimedia port generates a corresponding video tag by calling the AI module for the multimedia video stream management cloud platform, and combing reference element information according to which the AI module carries out video tag type classification on any video stream segment;
The adjacent tag judgment management module is used for judging and identifying any two video tags with adjacent judgment when the AI module is used for distinguishing and classifying by comparing deviation distribution conditions of the reference element information presented between any two different video tags, and extracting a plurality of pairs of video tag groups which are mutually adjacent tags;
The adjacent label judging and managing module comprises a characteristic element information carding unit and an adjacent label judging unit;
The characteristic element information carding unit is used for comparing deviation distribution conditions of the reference element information presented between any two different video tags;
The adjacent label judging unit is used for judging and identifying any two video labels with adjacent judgment when the AI module is used for distinguishing and classifying, and extracting a plurality of pairs of video label groups which are mutually adjacent labels;
The adjacent degree value calculation management module is used for extracting the distribution condition of the presented distinguishing reference element information in the process of judging the video labels of the corresponding video stream fragments within the range of the corresponding mutually adjacent labels by the AI module, and calculating the mutually adjacent degree value of each pair of mutually adjacent label video label groups;
the AI module optimizing prompt management module is used for monitoring the calling rate distribution of the video stream fragments stored in each storage area by a user in the multimedia video stream management cloud platform, carrying out video tag adjustment on the video stream fragments with abnormal calling rate by referring to corresponding adjacent tags, and judging whether to send early warning prompts needing to optimize the AI module according to the corresponding calling rate change conditions brought by the video tag adjustment;
the AI module optimizing prompt management module comprises a label adjusting and monitoring management unit and an optimizing early warning prompt management unit;
The label adjustment monitoring management unit is used for monitoring the calling rate distribution of the video stream fragments stored in each storage area by a user in the multimedia video stream management cloud platform, and performing video label adjustment on the video stream fragments with abnormal calling rate by referring to corresponding adjacent labels;
And the optimized early warning prompt management unit is used for judging whether to send an early warning prompt which needs to be optimized to the AI module according to the corresponding calling rate change condition brought by the video tag after adjustment.
It should be noted that the above-mentioned embodiments are merely preferred embodiments of the present invention, and the present invention is not limited thereto, but may be modified or substituted for some of the technical features thereof by those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims (7)

1.一种基于人工智能的多媒体视频流管理方法，其特征在于：所述方法包括：1. A multimedia video stream management method based on artificial intelligence, characterized in that: the method comprises: 步骤S1：对多媒体视频流管理云平台通过调用AI模块，对从多媒体端口接收到的每一视频流片段生成相应视频标签时所产生的过程数据进行采集，梳理AI模块在对任意视频流片段进行视频标签种类划分时所依据的参照要素信息；Step S1: The multimedia video stream management cloud platform calls the AI module to collect process data generated when each video stream segment received from the multimedia port generates a corresponding video tag, and sorts out the reference element information based on which the AI module classifies the video tag types of any video stream segment; 步骤S2：通过比对在任意两种不同视频标签之间所呈现的参照要素信息的偏差分布情况，对AI模块在进行区别分类时存在邻近判断的任意两种视频标签进行判断识别，提取得到若干对互为邻近标签的视频标签组；Step S2: By comparing the deviation distribution of the reference element information presented between any two different video labels, any two video labels that have adjacent judgments when the AI module performs differential classification are judged and identified, and several pairs of video label groups that are adjacent to each other are extracted; 步骤S3：提取AI模块对相应视频流片段，在相应互为邻近标签的范围内进行视频标签判断的过程中，所呈现的区别参照要素信息的分布情况，对每一对互为邻近标签的视频标签组计算相互之间的邻近程度值；Step S3: extracting the distribution of the distinguishing reference element information presented in the process of the AI module judging the video label of the corresponding video stream segment within the range of the corresponding mutually adjacent labels, and calculating the mutual proximity value for each pair of video label groups of mutually adjacent labels; 步骤S4：在多媒体视频流管理云平台内，监测用户对各存储区域内存储的视频流片段的调用率分布，对调用率呈现异常的视频流片段参考相应的邻近标签作视频标签调整，根据视频标签调整后带来的相应调用率变化情况，判断是否发送需对AI模块进行优化的预警提示；Step S4: In the multimedia video stream management cloud platform, the call rate distribution of the video stream segments stored in each storage area by the user is monitored, and the video tags of the video stream segments with abnormal call rates are adjusted by referring to the corresponding adjacent tags. According to the corresponding call rate changes brought about by the video tag adjustment, it is determined whether to send an early warning prompt that the AI module needs to be optimized; 所述步骤S4包括：The step S4 comprises: 步骤S4-1：在各存储区域内，监测用户对存储的所有视频流片段的平均调用率；若在对应视频标签为a的某存储区域内，存在有相应的调用率低于平均调用率的某视频流片段，捕捉与视频标签a之间的邻近程度值最高的视频标签b，将所述某视频流片段的视频标签由a调整为b；Step S4-1: In each storage area, the average call rate of all stored video stream segments by users is monitored; if in a storage area corresponding to a video tag a, there is a video stream segment whose corresponding call rate is lower than the average call rate, the video tag b with the highest proximity value to the video tag a is captured, and the video tag of the video stream segment is adjusted from a to b; 步骤S4-2：监测在将所述某视频流片段的视频标签由a调整为b后，用户在单位周期时间内对所述某视频流片段呈现的调用率，若调用率较于调整前呈现数值增长，生成一次针对AI模块优化的预警信号，若调用率较于调整前没有呈现数值增长，撤销调整；Step S4-2: monitoring the call rate of the video stream segment presented by the user within a unit period of time after the video tag of the video stream segment is adjusted from a to b, and if the call rate shows a numerical increase compared with before the adjustment, generating a warning signal for AI module optimization; if the call rate does not show a numerical increase compared with before the adjustment, canceling the adjustment; 步骤S4-3：当累计针对AI模块优化所生成的预警信号次数大于次数阈值，向管理终端发送需对AI模块进行优化的预警提示。Step S4-3: When the cumulative number of warning signals generated for the AI module optimization is greater than the number threshold, a warning prompt that the AI module needs to be optimized is sent to the management terminal. 2.根据权利要求1所述的一种基于人工智能的多媒体视频流管理方法，其特征在于：所述步骤S1包括：2. The method for managing multimedia video streams based on artificial intelligence according to claim 1, wherein step S1 comprises: 步骤S1-1：采集多媒体视频流管理云平台，在调用AI模块对从多媒体端口接收到的每一个视频流片段生成相应的视频标签之前，AI模块对每一个视频流片段在进行图像识别和音频分析后所提取得到的所有特征要素信息，并汇集生成对应所述每一个视频流片段的特征要素信息集合；Step S1-1: The multimedia video stream management cloud platform collects all feature element information extracted after image recognition and audio analysis for each video stream segment before calling the AI module to generate a corresponding video tag for each video stream segment received from the multimedia port, and collects and generates a feature element information set corresponding to each video stream segment; 步骤S1-2：分别将存储于多媒体视频流管理云平台内的，且所对应视频标签相同的所有视频流片段的特征要素信息集合进行汇集，分别得到AI模块在对任意视频流片段进行视频标签种类划分时所依据的若干个参照要素信息集合；其中，视频标签相同的视频流片段被置于多媒体视频流管理云平台内的一个相同存储区域进行数据存储；其中，一个参照要素信息集合对应一种视频标签。Step S1-2: Collect the characteristic element information sets of all video stream segments with the same corresponding video tags stored in the multimedia video stream management cloud platform, and obtain several reference element information sets based on which the AI module classifies the video tag types of any video stream segments; wherein the video stream segments with the same video tags are placed in a same storage area in the multimedia video stream management cloud platform for data storage; wherein one reference element information set corresponds to one video tag. 3.根据权利要求2所述的一种基于人工智能的多媒体视频流管理方法，其特征在于：所述步骤S2包括：3. The method for managing multimedia video streams based on artificial intelligence according to claim 2, wherein step S2 comprises: 步骤S2-1：逐一在任意两种不同视频标签所对应的参照要素信息集合中，对重合的特征要素信息进行提取，分别生成对应任意两种不同视频标签之间的重合参照要素信息集合；Step S2-1: extracting overlapping feature element information from reference element information sets corresponding to any two different video tags one by one, and generating overlapping reference element information sets corresponding to any two different video tags respectively; 步骤S2-2：获取在每一种视频标签与其他种类视频标签之间所对应的每一个重合参照要素信息集合中，所包含的特征要素信息的总数值；若在某任意两种不同视频标签之间的重合参照要素信息集合中，所包含的特征要素信息的总数值大于总数阈值，判断AI模块在对所述某任意两种不同视频标签进行区别分类时存在邻近判断，判断所述某任意两种不同视频标签为一对互为邻近标签的视频标签组。Step S2-2: Obtain the total value of the characteristic element information contained in each overlapping reference element information set corresponding to each type of video label and other types of video labels; if the total value of the characteristic element information contained in the overlapping reference element information set between any two different video labels is greater than the total threshold, it is determined that the AI module has a proximity judgment when distinguishing and classifying the any two different video labels, and it is determined that the any two different video labels are a pair of video label groups that are adjacent to each other. 4.根据权利要求3所述的一种基于人工智能的多媒体视频流管理方法，其特征在于：所述步骤S3包括：4. The method for managing multimedia video streams based on artificial intelligence according to claim 3, wherein step S3 comprises: 步骤S3-1：若视频标签A和视频标签B互为邻近标签，提取视频标签A的参照要素信息集合P(A)，视频标签B的参照要素信息集合P(B)，以及视频标签A和视频标签B之间的重合参照要素信息集合Ua,b；对视频标签A提取得到区别参照要素信息集合C1=P(A)-Ua,b,对视频标签B提取得到区别参照要素信息集合C2=P(B)-Ua,b；Step S3-1: If video tag A and video tag B are adjacent tags, extract the reference element information set P(A) of video tag A, the reference element information set P(B) of video tag B, and the overlapping reference element information set Ua ,b between video tag A and video tag B; extract the distinguishing reference element information set C1 = P(A)-Ua ,b for video tag A, and extract the distinguishing reference element information set C2 = P(B)-Ua ,b for video tag B; 步骤S3-2：从多媒体视频流管理云平台中汇集所有被标记视频标签A的视频流片段，得到第一视频流片段集合Y1，汇集所有被标记视频标签B的视频流片段，得到第二视频流片段集合Y2；若在第一视频流片段集合Y1或第二视频流片段集合Y2中存在某视频流片段，且对所述某视频流片段提取得到的特征要素信息集合R满足R∩Ua,b=Q≠，其中，Q表示特征要素信息集合R与重合参照要素信息集合Ua,b之间的交集，则将所述某视频流片段作特征标记，同时对所述某视频流片段提取得到目标区别要素信息集合Q'=R-Q；Step S3-2: Collect all video stream segments marked with video tag A from the multimedia video stream management cloud platform to obtain a first video stream segment set Y1, and collect all video stream segments marked with video tag B to obtain a second video stream segment set Y2; if a certain video stream segment exists in the first video stream segment set Y1 or the second video stream segment set Y2, and the feature element information set R extracted from the certain video stream segment satisfies R∩U a,b =Q≠ , where Q represents the intersection between the feature element information set R and the coincident reference element information set U a,b , then the certain video stream segment is marked as a feature, and at the same time, the target distinguishing element information set Q'=RQ is extracted from the certain video stream segment; 步骤S3-3：在第一视频流片段集合Y1中，汇集对所有被作特征标记的视频流片段提取得到的目标区别要素信息集合，累计出现的特征要素信息的种类数η1，在第二视频流片段集合Y2中，汇集对所有被作特征标记的视频流片段提取得到的目标区别要素信息集合，累计出现的特征要素信息的种类数η2；计算得到视频标签A和视频标签B之间的第一邻近指数β1=[η1/card(C1)+η2/card(C2)]/2；Step S3-3: In the first video stream segment set Y1, the target distinguishing element information set extracted from all the video stream segments marked as features is collected, and the number of types of the feature element information appearing is accumulated η 1 ; in the second video stream segment set Y2, the target distinguishing element information set extracted from all the video stream segments marked as features is collected, and the number of types of the feature element information appearing is accumulated η 2 ; the first proximity index β 1 =[η 1 /card(C 1 )+η 2 /card(C 2 )]/2 between the video label A and the video label B is calculated; 步骤S3-4：获取在第一视频流片段集合Y1中被作特征标记的视频流片段的数量占比值α1，以及在第二视频流片段集合Y2中被作特征标记的视频流片段的数量占比值α2；计算得到视频标签A和视频标签B之间的第二邻近指数β2=(α1+α2)/2；Step S3-4: obtaining a ratio α1 of the number of video stream segments marked as features in the first video stream segment set Y1 and a ratio α2 of the number of video stream segments marked as features in the second video stream segment set Y2; calculating a second proximity index β 2 =(α1+α2)/2 between the video tag A and the video tag B; 步骤S3-5：逐一将第一视频流片段集合Y1内各视频流片段所对应的特征要素信息集合，与第二视频流片段集合Y2内各视频流片段所对应的特征要素信息集合进行相似度计算，捕捉最高相似度值δ；计算视频标签A和视频标签B之间的邻近程度值ξ=(1/β1+β2)×δ。Step S3-5: Calculate the similarity of the feature element information set corresponding to each video stream segment in the first video stream segment set Y1 with the feature element information set corresponding to each video stream segment in the second video stream segment set Y2 one by one, and capture the highest similarity value δ; calculate the proximity value ξ=(1/β 1 +β 2 )×δ between the video label A and the video label B. 5.一种多媒体视频流管理系统，用于执行权利要求1-4中任意一项所述的一种基于人工智能的多媒体视频流管理方法，其特征在于，所述系统包括：视频流标签处理数据管理模块、邻近标签判断管理模块、邻近程度值计算管理模块、AI模块优化提示管理模块；5. A multimedia video stream management system, used to execute a multimedia video stream management method based on artificial intelligence as described in any one of claims 1 to 4, characterized in that the system comprises: a video stream label processing data management module, a neighboring label judgment management module, a neighboring degree value calculation management module, and an AI module optimization prompt management module; 所述视频流标签处理数据管理模块，用于对多媒体视频流管理云平台通过调用AI模块，对从多媒体端口接收到的每一视频流片段生成相应视频标签时所产生的过程数据进行采集，梳理AI模块在对任意视频流片段进行视频标签种类划分时所依据的参照要素信息；The video stream label processing data management module is used to collect process data generated when the multimedia video stream management cloud platform generates corresponding video labels for each video stream segment received from the multimedia port by calling the AI module, and to sort out the reference element information based on which the AI module classifies the video labels of any video stream segment; 所述邻近标签判断管理模块，用于通过比对在任意两种不同视频标签之间所呈现的参照要素信息的偏差分布情况，对AI模块在进行区别分类时存在邻近判断的任意两种视频标签进行判断识别，提取得到若干对互为邻近标签的视频标签组；The adjacent label judgment management module is used to judge and identify any two video labels that have adjacent judgments when the AI module performs differential classification by comparing the deviation distribution of reference element information presented between any two different video labels, and extract several pairs of video label groups that are adjacent to each other; 所述邻近程度值计算管理模块，用于提取AI模块对相应视频流片段，在相应互为邻近标签的范围内进行视频标签判断的过程中，所呈现的区别参照要素信息的分布情况，对每一对互为邻近标签的视频标签组计算相互之间的邻近程度值；The proximity value calculation management module is used to extract the distribution of the distinguishing reference element information presented in the process of the AI module judging the video label of the corresponding video stream segment within the range of the corresponding mutually adjacent labels, and calculate the mutual proximity value between each pair of video label groups of mutually adjacent labels; 所述AI模块优化提示管理模块，用于在多媒体视频流管理云平台内，监测用户对各存储区域内存储的视频流片段的调用率分布，对调用率呈现异常的视频流片段参考相应的邻近标签作视频标签调整，根据视频标签调整后带来的相应调用率变化情况，判断是否发送需对AI模块进行优化的预警提示。The AI module optimization prompt management module is used to monitor the distribution of user call rates of video stream segments stored in each storage area in the multimedia video stream management cloud platform, adjust the video tags of video stream segments with abnormal call rates with reference to corresponding adjacent tags, and determine whether to send an early warning prompt that the AI module needs to be optimized based on the corresponding call rate changes brought about by the video tag adjustment. 6.根据权利要求5所述的一种多媒体视频流管理系统，其特征在于：所述邻近标签判断管理模块包括：特征要素信息梳理单元、邻近标签判断单元；6. A multimedia video stream management system according to claim 5, characterized in that: the adjacent tag judgment management module comprises: a feature element information combing unit, an adjacent tag judgment unit; 所述特征要素信息梳理单元，用于通过比对在任意两种不同视频标签之间所呈现的参照要素信息的偏差分布情况；The characteristic element information combing unit is used to compare the deviation distribution of the reference element information presented between any two different video tags; 所述邻近标签判断单元，用于对AI模块在进行区别分类时存在邻近判断的任意两种视频标签进行判断识别，提取得到若干对互为邻近标签的视频标签组。The adjacent label judgment unit is used to judge and identify any two video labels that have adjacent judgments when the AI module performs differential classification, and extract several pairs of video label groups that are adjacent to each other. 7.根据权利要求5所述的一种多媒体视频流管理系统，其特征在于：所述AI模块优化提示管理模块包括标签调整监测管理单元、优化预警提示管理单元；7. A multimedia video stream management system according to claim 5, characterized in that: the AI module optimization prompt management module includes a label adjustment monitoring management unit and an optimization early warning prompt management unit; 所述标签调整监测管理单元，用于在多媒体视频流管理云平台内，监测用户对各存储区域内存储的视频流片段的调用率分布，对调用率呈现异常的视频流片段参考相应的邻近标签作视频标签调整；The label adjustment monitoring management unit is used to monitor the distribution of the calling rate of the video stream segments stored in each storage area by users in the multimedia video stream management cloud platform, and to adjust the video labels of the video stream segments with abnormal calling rates by referring to the corresponding adjacent labels; 所述优化预警提示管理单元，用于根据视频标签调整后带来的相应调用率变化情况，判断是否发送需对AI模块进行优化的预警提示。The optimization warning prompt management unit is used to determine whether to send a warning prompt that the AI module needs to be optimized based on the corresponding call rate change caused by the adjustment of the video tag.
CN202411544997.7A 2024-11-01 2024-11-01 Multimedia video stream management system and method based on artificial intelligence Active CN119046481B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202411544997.7A CN119046481B (en) 2024-11-01 2024-11-01 Multimedia video stream management system and method based on artificial intelligence

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202411544997.7A CN119046481B (en) 2024-11-01 2024-11-01 Multimedia video stream management system and method based on artificial intelligence

Publications (2)

Publication Number Publication Date
CN119046481A CN119046481A (en) 2024-11-29
CN119046481B true CN119046481B (en) 2025-01-24

Family

ID=93574420

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202411544997.7A Active CN119046481B (en) 2024-11-01 2024-11-01 Multimedia video stream management system and method based on artificial intelligence

Country Status (1)

Country Link
CN (1) CN119046481B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109660865A (en) * 2018-12-17 2019-04-19 杭州柚子街信息科技有限公司 Make method and device, medium and the electronic equipment of video tab automatically for video
CN110795598A (en) * 2019-10-29 2020-02-14 青岛聚看云科技有限公司 Intelligent column generation method based on VCA video tag analysis

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CA3173977A1 (en) * 2020-03-02 2021-09-10 Visual Supply Company Systems and methods for automating video editing
KR102501723B1 (en) * 2020-11-17 2023-02-17 연세대학교 산학협력단 Method and device for extracting video feature
CN118280385B (en) * 2024-06-04 2024-08-06 国网浙江省电力有限公司营销服务中心 Problem data processing method, device, equipment and medium based on semantic analysis model

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109660865A (en) * 2018-12-17 2019-04-19 杭州柚子街信息科技有限公司 Make method and device, medium and the electronic equipment of video tab automatically for video
CN110795598A (en) * 2019-10-29 2020-02-14 青岛聚看云科技有限公司 Intelligent column generation method based on VCA video tag analysis

Also Published As

Publication number Publication date
CN119046481A (en) 2024-11-29

Similar Documents

Publication Publication Date Title
CN110351244A (en) A kind of network inbreak detection method and system based on multireel product neural network fusion
CN110505179B (en) A method and system for detecting abnormal network traffic
CN112507860A (en) Video annotation method, device, equipment and storage medium
CN114049508B (en) Fraud website identification method and system based on picture clustering and manual research and judgment
WO2019200739A1 (en) Data fraud identification method, apparatus, computer device, and storage medium
Liu et al. Video content analysis for compliance audit in finance and security industry
WO2022114895A1 (en) System and method for providing customized content service by using image information
CN117119253B (en) High-quality video frame extraction method for target object
CN114386931A (en) Image analysis system and method based on AI technology
CN115269438A (en) Automatic testing method and device for image processing algorithm
CN119046481B (en) Multimedia video stream management system and method based on artificial intelligence
CN110795598A (en) Intelligent column generation method based on VCA video tag analysis
CN114528909A (en) Unsupervised anomaly detection method based on flow log feature extraction
Laptev et al. Visualization system for fire detection in the video sequences
CN116991932B (en) Data analysis and management system and method based on artificial intelligence
CN118569617A (en) Intelligent power plant management and control system and method based on computer vision target detection
Chen et al. An improved local descriptor and threshold learning for unsupervised dynamic texture segmentation
CN112487864A (en) Method for detecting small target safety helmet and protective clothing for construction site
CN111126373A (en) Device and method for judging Internet short video violations based on cross-modal recognition technology
CN114092743B (en) Compliance detection method and device for sensitive picture, storage medium and equipment
CN105930430A (en) Non-cumulative attribute based real-time fraud detection method and apparatus
CN113344696B (en) Position forecast evaluation method, device, equipment and storage medium
Prakash-Borah et al. Real-time helmet detection and number plate extraction using computer vision
CN110633466B (en) Short message crime identification method and system based on semantic analysis and readable storage medium
CN115457620A (en) User expression recognition method and device, computer equipment and storage medium

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