CN112084369A - Highlight moment mining method and model based on video live broadcast - Google Patents
Highlight moment mining method and model based on video live broadcast Download PDFInfo
- Publication number
- CN112084369A CN112084369A CN202010767848.2A CN202010767848A CN112084369A CN 112084369 A CN112084369 A CN 112084369A CN 202010767848 A CN202010767848 A CN 202010767848A CN 112084369 A CN112084369 A CN 112084369A
- Authority
- CN
- China
- Prior art keywords
- highlight
- live broadcast
- data
- time
- points
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 26
- 238000005065 mining Methods 0.000 title claims abstract description 24
- 230000000007 visual effect Effects 0.000 claims abstract description 15
- 101100481876 Danio rerio pbk gene Proteins 0.000 claims abstract description 5
- 101100481878 Mus musculus Pbk gene Proteins 0.000 claims abstract description 5
- 238000012800 visualization Methods 0.000 claims description 6
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 claims description 3
- 238000013499 data model Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims 1
- 238000000605 extraction Methods 0.000 claims 1
- 238000010586 diagram Methods 0.000 abstract description 4
- 238000007405 data analysis Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 206010000060 Abdominal distension Diseases 0.000 description 1
- 241000711573 Coronaviridae Species 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 208000024330 bloating Diseases 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000004091 panning Methods 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/70—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
- G06F16/74—Browsing; Visualisation therefor
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
- G06F40/295—Named entity recognition
-
- 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]
- H04N21/20—Servers specifically adapted for the distribution of content, e.g. VOD servers; Operations thereof
- H04N21/21—Server components or server architectures
- H04N21/218—Source of audio or video content, e.g. local disk arrays
- H04N21/2187—Live feed
-
- 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
- H04N21/478—Supplemental services, e.g. displaying phone caller identification, shopping application
- H04N21/4788—Supplemental services, e.g. displaying phone caller identification, shopping application communicating with other users, e.g. chatting
Abstract
The invention discloses a highlight moment mining model and method based on video live broadcast, which can mine highlight moment intervals corresponding to dimension indexes in the whole live broadcast process, and show contents related to the highlight moment intervals by combining with content data related to live broadcast. For example, the method shows the original comment data or word cloud pictures of the user in the live broadcast sales highlight time interval, so that the user can clearly know what happens at the live broadcast time, and further analyze what reason causes the live broadcast sales to be obviously improved. The model firstly acquires all convex points and concave points in a corresponding time sequence diagram through minute-level dimension data, and then acquires highlight time intervals of corresponding dimensions through rules. For example, the distance between adjacent bumps and pits is maximized as the highlight moment criterion, and TopK highlight moments are output. And then the original data of the highlight time interval are combined together for visual display to a user.
Description
Technical Field
The invention relates to the technical field of computer text processing, in particular to a highlight moment mining method and a highlight moment mining model based on video live broadcast.
Background
With the deep development of the internet, the rise of live broadcast such as panning, shaking and the like, and the video live broadcast gradually becomes popular. Particularly, under the influence of 2019-nCov novel coronavirus, live broadcast delivery of various large anchor broadcasters and merchants is more and more common, and a new live broadcast delivery mode is gradually formed. Therefore, a new valuable information demand for mining live video data is generated, and based on the demand, a highlight moment mining model for live video is provided, so that a data analysis user can clearly master the change condition of each dimension index data corresponding to each time period in the live broadcast process, mainly mine and show highlight moments, analyze original comment data of the highlight moments, analyze and process the data by combining the most advanced NLP technology, and visually display in multiple dimensions.
Disclosure of Invention
Aiming at the problems, the invention provides a highlight moment mining method based on live video, which comprises the following steps:
s1, acquiring minute-level data of each dimension index of single-field live video to obtain time sequence data of each corresponding dimension;
s2, calculating the convex points of the time series data, and acquiring concave points corresponding to the convex points according to the convex point addition custom rule;
s3, calculating the judgment standard of the highlight time of each dimension data based on the convex points and the concave points acquired in the step S2, and defining the highlight time interval of each dimension data;
s4: acquiring and processing data corresponding to highlight time intervals, acquiring a related original data set in the highlight time interval corresponding to live broadcast based on data such as live broadcast comments, and processing and analyzing the original data by combining with an NLP (non line of sight) correlation technique to obtain visual display information;
s5: and displaying highlight moments in a multi-dimensional visualization mode. And displaying the highlight time by multidimensional visual application based on the obtained highlight time interval and the corresponding visual display information.
As a further explanation of the present invention, the time-series data in step S1 is the corresponding number of praise, pollen, sales or comment in a unit minute (1 minute or every 5 minutes).
Furthermore, in step S2, the salient point is obtained from peakuils in python, the pit is obtained by comparing with the salient point, and when a time point value on the salient point is smaller than the current point, the comparison is continued until the last point value is found to be larger than the current point value, and then the current point is considered to be the pit.
Furthermore, smooth points exist between the convex points and the concave points, namely when the previous point is not obviously raised compared with the current point, the corresponding concave points are not used for comparison, the number of the smooth points is counted, and when the number of the smooth points is greater than the specified number of the smooth points, the comparison is stopped.
Further, in step S3, the highlight moments with the maximum difference between each group of bumps and its corresponding pits are output as the criteria for measuring the highlight moments, wherein the top TopK differences are maximized.
Further, step S3 includes setting a time window, and defining the number of highlight moments in a time window.
Further, the NLP correlation technique processing on the raw data set in step S4 includes NER-extracting star names, keyword-matching words related to price, and the like on the comment data.
Furthermore, when keywords of live comment data are obtained, the combination of traditional code table matching and an advanced entity recognition algorithm is adopted. For comment data, firstly, a large number of comments without key words are filtered through code table matching, and then the comment data related to the key words are input into an entity recognition algorithm and an entity is extracted by combining the code table. The advantage of this is to avoid the situation that the entity identification algorithm is only used to extract the entity data with too much resource shortage and the situation that the code table is only used for matching but the matched key word may not be the corresponding entity key word. For example, if "russia" appears in live comments and the star code table contains "russia", matching is hit, and the situation needs to be corrected through a named entity recognition algorithm based on a Bert model.
In another aspect of the present invention, a highlight moment mining model based on live video is provided, which includes:
and (3) data model: the system is used for acquiring the minute-level data of the corresponding dimensionality of the live broadcast to form time sequence data of the corresponding dimensionality;
time series data concave-convex point calculation model: the device comprises a plurality of convex points and concave points which are used for calculating time sequence data, and a maximized difference value between each group of convex points and the corresponding concave points;
NLP analytical model: the system is used for analyzing the original data of the live video and calculating visual information;
multidimensional visualization application: and displaying the highlight time in a multi-dimensional manner based on the obtained highlight time interval and the visual information thereof.
The invention has the beneficial effects that:
the highlight time interval corresponding to the dimension indexes (such as the live broadcast sales number, the praise number, the comment number and the comment content keyword …) in the whole live broadcast process can be mined, and the content related to the highlight time interval is displayed by combining with the content data related to the live broadcast. For example, the method shows the original comment data or word cloud pictures of the user in the live broadcast sales volume highlight time interval, so that the user can clearly know what happens at the highlight time, and further analyze what causes cause the live broadcast sales volume to be obviously improved (such as anchor humorous and good product quality …). The model firstly obtains all convex points and concave points (namely local minimum values and local maximum values) in a corresponding time sequence diagram for minute-level dimension data, and then obtains highlight time intervals of corresponding dimensions through rules. For example, the distance between adjacent bumps and pits is maximized as the highlight moment criterion, and TopK highlight moments are output. And then the original data of the highlight time interval are combined together for visual display to a user.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
FIG. 2 is a keyword bubble display diagram of dimensions corresponding to highlight moments according to the present invention;
fig. 3 is a word cloud graph and a keyword word frequency highlight moment graph corresponding to the keyword dimension in the bubble graph according to the present invention.
Detailed Description
The following detailed description of specific embodiments of the invention, taken in conjunction with the accompanying drawings, will make apparent that the described embodiments are only some, but not all embodiments of the invention.
A highlight moment mining model based on live video comprises the following steps:
and (3) data model: the system is used for acquiring the minute-level data of the corresponding dimensionality of the live broadcast to form time sequence data of the corresponding dimensionality;
time series data concave-convex point calculation model: the device comprises a plurality of convex points and concave points which are used for calculating time sequence data, and a maximized difference value between each group of convex points and the corresponding concave points;
NLP analytical model: the system is used for analyzing the original data of the live video and calculating visual information;
multidimensional visualization application: and displaying the highlight time in a multi-dimensional manner based on the obtained highlight time interval and the visual information thereof.
As shown in fig. 1, a flow chart of the method of the present invention is a flow chart of a video live broadcast highlight moment mining method, and specifically includes the following steps:
and S1, acquiring the minute-level data of the corresponding dimension of the live broadcast. The dimension here is understood to be an index related to live broadcast, the minute level represents that the unit is minute, and the interval may be 1 minute, 5 minutes, etc., for example, the number of praise (or bloat, sales, and comments) corresponding to every 1 minute (or every 5 minutes) of a certain live broadcast is obtained, and the time series data of the corresponding dimension is obtained. Taking how to obtain the live sales per minute as an example, first, all the sales of the live broadcast are collected in seconds (collected and obtained by a crawler or other methods), for example, the sales data corresponding to all the commodities of the live broadcast during the live broadcast are collected every 5 seconds, and then the collected sales data in seconds are combined into minute-level data, that is, the sales of the minute-level are obtained by adding the second-level sales increments contained in the minute.
And S2, acquiring the convex points of the time sequence data, and acquiring the concave points corresponding to the convex points according to the convex point addition custom rule. The time sequence salient points are obtained in a plurality of ways, for example, the time sequence salient points can be easily obtained by utilizing packets such as peakuils in python, the seeking of the concave points starts from the salient points, when the value of a time point on the salient point is smaller than that of the current point, the comparison is continued until the value of the previous point is found to be larger than that of the current point, and the current point is considered to be the concave point. In some embodiments, a smooth point may be set, i.e., not as a corresponding pit when the last point is not significantly elevated compared to the current point, the comparison is continued, and the number of occurrences of the smooth point is counted, and the comparison is stopped when the number of occurrences is greater than a specified number of occurrences.
And S3, acquiring a highlight time interval. Based on the bumps and the pits obtained in S2, the highlight time interval corresponding to each dimension can be obtained, for example, by using the maximized difference between each group of bumps and its corresponding pits as the standard for measuring highlight time, the highlight time of the corresponding dimension with the maximized TopK differences before the specification can be output. To avoid the highlight moments concentrated in certain time periods, we can also specify time windows, such as half an hour as a time window, and only allow one highlight moment to occur in each time window (i.e. each window only obtains the first one with maximized difference).
S4: and acquiring and processing data corresponding to the highlight time interval. Based on data such as live barrage comments, related original data sets in the highlight moment interval corresponding to live broadcasting can be obtained, and the original data are processed and analyzed by combining with NLP (non line segment) related technology. For example, NER is performed on the comment data to extract star names, keywords are matched with words related to price, and information such as keywords corresponding to related fields (such as star mentions, price talking and product talking …) and times of keywords is analyzed statistically and used as next step visual display. Of course, the original data set is not limited to live barrage comment data, and may also be live related dimension data, such as live online number of people, number of bloating, and the like. And finally, visually displaying the highlight time interval describing the corresponding dimension by analyzing and processing the original data. For example, statistical results such as analysis results of corresponding comment data (understanding what the user talks about affects sales volume change), real-time online number of people, number of blogs in the market, and the like in highlight time intervals for displaying sales volume are obtained.
S5: and displaying highlight moments in a multi-dimensional visualization mode. Based on the obtained highlight time interval and the corresponding information thereof, a very cool and valuable multidimensional visual application display highlight time is designed. As shown in fig. 2, we show keyword bubble charts of all comment contents in the whole live broadcast at corresponding highlight moments of 5 fields of brand negotiation, price negotiation, product negotiation, star promotion and shopping. As shown in fig. 3, by clicking the keyword bubble corresponding to each dimension (for example, the keyword bubble of "no hit" corresponding to the "hit on the figure in the" buy "dimension), a word cloud corresponding to the comment of the corresponding keyword appearing in the whole live broadcast and a highlight moment corresponding to the word frequency are displayed in a cascade manner.
The highlight moment mining model based on the live video can be used for analyzing and mining all dimension index data of the finished live video and showing highlight moments of the data dimensions. Certainly, the scheme of the invention can also be applied to some live broadcasts with goods in non-anchor and merchants, such as game live broadcasts, outdoor live broadcasts and the like, in the live broadcasts, minute-level time sequence data of relevant dimensions, such as real-time online number of people, leaving number of people, number of praise, number of comments and the like, can be used for mining and displaying highlight moments corresponding to the dimensions by using the highlight moment mining model, and further used for data analysis.
The foregoing is illustrative of the preferred embodiments of the present invention only and is not to be construed as limiting the claims. The invention is not limited to the above embodiments, the specific construction of which allows variations, and in any case variations, which are within the scope of the invention as defined in the independent claims.
Claims (9)
1. A highlight moment mining method based on video live broadcast is characterized by comprising the following steps: the method comprises the following steps:
s1, acquiring minute-level data of each dimension index of single-field live video to obtain time sequence data of each corresponding dimension;
s2, calculating the convex points of the time series data, and acquiring concave points corresponding to the convex points according to the convex point addition custom rule;
s3, calculating the judgment standard of the highlight time of each dimension data based on the convex points and the concave points acquired in the step S2, and defining the highlight time interval of each dimension data;
s4: acquiring dimension index data corresponding to the highlight moment interval as an original data set related to the highlight moment interval, and performing NLP processing analysis to obtain visual display information;
s5: and displaying the highlight moments through multidimensional visualization application based on the acquired highlight moment intervals and the corresponding visual display information.
2. The video live broadcast highlight time mining method according to claim 1, characterized in that: in step S1, the time-series data is the corresponding number of praise, pollen expansion, sales or comment in a unit of minute.
3. The video live broadcast highlight time mining method according to claim 1, characterized in that: in step S2, the salient point is obtained from peakuils in python, the pit is obtained by comparing with the salient point, when a time point value on the salient point is smaller than the current point, the comparison is continued until the last point value is found to be larger than the current point value, and the current point is considered to be the pit.
4. The video live broadcast highlight time mining method according to claim 3, characterized in that: and smooth points exist between the convex points and the concave points, namely when the previous point is not obviously raised compared with the current point, the previous point is not taken as the corresponding concave point, the comparison is continued, the occurrence frequency of the smooth points is counted, and the comparison is stopped when the occurrence frequency is greater than the specified occurrence frequency.
5. The video live broadcast highlight time mining method according to claim 1, characterized in that: in step S3, the maximum difference between each group of bumps and the corresponding pits is used as a criterion for measuring highlight moments, and highlight moments with the top TopK differences maximized in the corresponding dimensions are output.
6. The video live broadcast highlight time mining method according to claim 1, characterized in that: step S3 includes setting a time window, defining the number of highlight moments within a time window.
7. The video live broadcast highlight time mining method according to claim 1, characterized in that: the NLP processing analysis in step S4 includes NER extraction of the comment data.
8. The video live broadcast highlight time mining method according to claim 7, characterized in that: and when the NER extracts the keywords of the live comment data, a mode of combining code table matching and an entity recognition algorithm is adopted.
9. A highlight moment mining model based on live video comprises the following steps:
and (3) data model: the system is used for acquiring the minute-level data of the corresponding dimensionality of the live broadcast to form time sequence data of the corresponding dimensionality;
time series data concave-convex point calculation model: the device comprises a plurality of convex points and concave points which are used for calculating time sequence data, and a maximized difference value between each group of convex points and the corresponding concave points;
NLP analytical model: the system is used for analyzing the original data of the live video and calculating visual information;
multidimensional visualization application: and displaying the highlight time in a multi-dimensional manner based on the obtained highlight time interval and the visual information thereof.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010767848.2A CN112084369A (en) | 2020-08-03 | 2020-08-03 | Highlight moment mining method and model based on video live broadcast |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010767848.2A CN112084369A (en) | 2020-08-03 | 2020-08-03 | Highlight moment mining method and model based on video live broadcast |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN112084369A true CN112084369A (en) | 2020-12-15 |
Family
ID=73736151
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010767848.2A Pending CN112084369A (en) | 2020-08-03 | 2020-08-03 | Highlight moment mining method and model based on video live broadcast |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112084369A (en) |
Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104918067A (en) * | 2014-03-12 | 2015-09-16 | 乐视网信息技术（北京）股份有限公司 | Method and system for performing curve processing on video hot degree |
| US20160247540A1 (en) * | 2015-02-25 | 2016-08-25 | Carnegie Technology Investment Limited | Method of recording multiple highlights concurrently |
| CN108259929A (en) * | 2017-12-22 | 2018-07-06 | 北京交通大学 | A kind of prediction of video active period pattern and caching method |
| CN109040796A (en) * | 2018-08-17 | 2018-12-18 | 深圳市迅雷网络技术有限公司 | The calculation method of contents fragment temperature, the playback method of video content and device |
| CN109511006A (en) * | 2018-11-13 | 2019-03-22 | 广州虎牙科技有限公司 | A kind of word cloud drawing generating method, device, equipment and storage medium |
| CN109672899A (en) * | 2018-12-13 | 2019-04-23 | 南京邮电大学 | The Wonderful time of object game live scene identifies and prerecording method in real time |
| CN109685144A (en) * | 2018-12-26 | 2019-04-26 | 上海众源网络有限公司 | The method, apparatus and electronic equipment that a kind of pair of Video Model does to assess |
| US10363488B1 (en) * | 2015-06-29 | 2019-07-30 | Amazon Technologies, Inc. | Determining highlights in a game spectating system |
| CN110418157A (en) * | 2019-08-28 | 2019-11-05 | 广州华多网络科技有限公司 | Live video back method and device, storage medium and electronic equipment |
| CN110475155A (en) * | 2019-08-19 | 2019-11-19 | 北京字节跳动网络技术有限公司 | Live video temperature state identification method, device, equipment and readable medium |
| CN110798716A (en) * | 2019-11-19 | 2020-02-14 | 深圳市迅雷网络技术有限公司 | Video highlight playing method and related device |
| CN111010584A (en) * | 2019-12-05 | 2020-04-14 | 网易（杭州）网络有限公司 | Live broadcast data processing method and device, server and storage medium |
| CN111083515A (en) * | 2019-12-31 | 2020-04-28 | 广州华多网络科技有限公司 | Method, device and system for processing live broadcast content |
| CN111435374A (en) * | 2019-01-11 | 2020-07-21 | 百度在线网络技术（北京）有限公司 | Display device and method for searching statistical data |
| CN111447489A (en) * | 2020-04-02 | 2020-07-24 | 北京字节跳动网络技术有限公司 | Video processing method and device, readable medium and electronic equipment |
-
2020
- 2020-08-03 CN CN202010767848.2A patent/CN112084369A/en active Pending
Patent Citations (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104918067A (en) * | 2014-03-12 | 2015-09-16 | 乐视网信息技术（北京）股份有限公司 | Method and system for performing curve processing on video hot degree |
| US20160247540A1 (en) * | 2015-02-25 | 2016-08-25 | Carnegie Technology Investment Limited | Method of recording multiple highlights concurrently |
| US10363488B1 (en) * | 2015-06-29 | 2019-07-30 | Amazon Technologies, Inc. | Determining highlights in a game spectating system |
| CN108259929A (en) * | 2017-12-22 | 2018-07-06 | 北京交通大学 | A kind of prediction of video active period pattern and caching method |
| CN109040796A (en) * | 2018-08-17 | 2018-12-18 | 深圳市迅雷网络技术有限公司 | The calculation method of contents fragment temperature, the playback method of video content and device |
| CN109511006A (en) * | 2018-11-13 | 2019-03-22 | 广州虎牙科技有限公司 | A kind of word cloud drawing generating method, device, equipment and storage medium |
| CN109672899A (en) * | 2018-12-13 | 2019-04-23 | 南京邮电大学 | The Wonderful time of object game live scene identifies and prerecording method in real time |
| CN109685144A (en) * | 2018-12-26 | 2019-04-26 | 上海众源网络有限公司 | The method, apparatus and electronic equipment that a kind of pair of Video Model does to assess |
| CN111435374A (en) * | 2019-01-11 | 2020-07-21 | 百度在线网络技术（北京）有限公司 | Display device and method for searching statistical data |
| CN110475155A (en) * | 2019-08-19 | 2019-11-19 | 北京字节跳动网络技术有限公司 | Live video temperature state identification method, device, equipment and readable medium |
| CN110418157A (en) * | 2019-08-28 | 2019-11-05 | 广州华多网络科技有限公司 | Live video back method and device, storage medium and electronic equipment |
| CN110798716A (en) * | 2019-11-19 | 2020-02-14 | 深圳市迅雷网络技术有限公司 | Video highlight playing method and related device |
| CN111010584A (en) * | 2019-12-05 | 2020-04-14 | 网易（杭州）网络有限公司 | Live broadcast data processing method and device, server and storage medium |
| CN111083515A (en) * | 2019-12-31 | 2020-04-28 | 广州华多网络科技有限公司 | Method, device and system for processing live broadcast content |
| CN111447489A (en) * | 2020-04-02 | 2020-07-24 | 北京字节跳动网络技术有限公司 | Video processing method and device, readable medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111143610B (en) | Content recommendation method and device, electronic equipment and storage medium | |
| CN108694223B (en) | User portrait database construction method and device | |
| CN102222103B (en) | Method and device for processing matching relationship of video content | |
| CN105357586A (en) | Video bullet screen filtering method and device | |
| CN110582025A (en) | Method and apparatus for processing video | |
| CN104462573A (en) | Method and device for displaying video retrieval results | |
| CN106331778A (en) | Video recommendation method and device | |
| MX2012000724A (en) | Estimating and displaying social interest in time-based media. | |
| CN110188241B (en) | Intelligent manufacturing system and manufacturing method for events | |
| CN113779381B (en) | Resource recommendation method, device, electronic equipment and storage medium | |
| CN111460221A (en) | Comment information processing method and device and electronic equipment | |
| CN111831854A (en) | Video tag generation method and device, electronic equipment and storage medium | |
| CN113315988B (en) | Live video recommendation method and device | |
| EP4086786A1 (en) | Video processing method, video searching method, terminal device, and computer-readable storage medium | |
| US20150113016A1 (en) | Information processing apparatus, information processing method, and program | |
| CN108012192A (en) | A kind of method and system of identification and the polymerization of video resource | |
| CN112507163A (en) | Duration prediction model training method, recommendation method, device, equipment and medium | |
| US20150220940A1 (en) | Trend data counter | |
| CN113301382B (en) | Video processing method, device, medium, and program product | |
| CN109829033B (en) | Data display method and terminal equipment | |
| CN110290403B (en) | Network video paster advertisement playing method and system | |
| CN112182244A (en) | Brand knowledge graph construction method and device and terminal | |
| CN112084369A (en) | Highlight moment mining method and model based on video live broadcast | |
| CN115474093B (en) | Method and device for calculating importance of video elements, storage medium and electronic equipment | |
| CN113365138B (en) | Content display method and device, electronic 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 |