CN102567467A - Method for acquiring hotspot video information based on video tags - Google Patents
Method for acquiring hotspot video information based on video tags Download PDFInfo
- Publication number
- CN102567467A CN102567467A CN2011103965154A CN201110396515A CN102567467A CN 102567467 A CN102567467 A CN 102567467A CN 2011103965154 A CN2011103965154 A CN 2011103965154A CN 201110396515 A CN201110396515 A CN 201110396515A CN 102567467 A CN102567467 A CN 102567467A
- Authority
- CN
- China
- Prior art keywords
- video
- feature description
- focus
- description speech
- video tab
- 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
Images
Landscapes
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
Abstract
The invention discloses a method for acquiring hotspot video information based on video tags, which comprises the following steps of: acquiring a video tag of a video, wherein the video tag comprises time points and tag contents of the video; carrying out Chinese word segmentation on the tag contents of the video tag; filtering the obtained word segmentation results so as to keep nouns and verbs as feature description words of a time point at which the video tag is located; calculating the importance degree values of the feature description words in the video tag; according to the importance degree values, carrying out sequencing on the feature description words, and taking the most important k feature description words as standby hotspot words of the time point; and carrying out statistical modeling on the standby hotspot words so as to generate hotspot video information. By using the method disclosed by the invention, the accuracy of video description can be improved through hotspots, thus the hotspot-based advertisement delivery effect is improved.
Description
Technical field
The present invention relates to the Internet video application, more particularly, the present invention relates to a kind of method of obtaining the video hot information based on video tab.
Background technology
Video tab is meant the phrase that is used to describe video features.Existing video tab technology all is to describe to whole video, can't describe the video highlight fragment.In addition, existing video tab technology also exists description accurate inadequately, cause based on the advertisement pushing of existing video tab technology inadequately precisely, ad content and video content is uncorrelated, the problem of throwing in weak effect.
Summary of the invention
In view of this, the purpose of this invention is to provide and a kind ofly obtain the method for video hot information based on video tab, it can improve the video presentation accuracy through focus, and then improves the advertisement pushing effect based on focus.
The present invention realizes through following technical scheme:
A kind ofly obtain the method for video hot information based on video tab, may further comprise the steps: obtain the video tab of video, video tab comprises the time point and the label substance of video; Label substance to video tab carries out Chinese word segmentation; Filter word segmentation result, to keep noun, verb feature description speech, the significance level value of calculated characteristics descriptor in video tab as video tab place time point; According to the significance level value feature description speech is sorted; And get the focus reserved word of K most important characteristic descriptor as time point, the focus reserved word is carried out statistical modeling, to generate the video hot information.
The step of the significance level value of calculated characteristics descriptor in video tab comprises: the word frequency value of calculated characteristics descriptor, concrete computing formula does
N wherein
I, jBe the number of times that feature description speech i occurs in video tab j, M is the feature description speech sum of video tab j, the reverse file frequency values of calculated characteristics descriptor, and concrete computing formula does
Wherein | D| is the sum of video tab, | { j:t
i∈ d
j| for comprising the video tab sum of feature description speech i, according to the reverse file frequency values of the word frequency value * feature description speech of the significance level value=feature description speech of following formula calculated characteristics descriptor.
The focus reserved word is carried out statistical modeling; Step to generate the video hot information comprises: take out all time points that have the focus reserved word in the video; From all time points, select N maximum time point of number of clicks as interim focus; All focus reserved words in before and after the interim focus 10 seconds are deposited in the interim focus, to generate the video hot information.
The step of the significance level value of calculated characteristics descriptor in video tab is to adopt the TF-IDF algorithm.
The value of K equals 5.
The value of N equals length/300 second of video.
The invention has the advantages that:
(1) video segment is described: through the video focus that extracts, can obtain the wonderful of video, thereby to the fragment pushed advertisement;
(2) the focus feature description is accurate: the feature description speech that each wonderful has, push the relevant advertisement of theme to the feature description speech, and cause user's sympathetic response more easily, improve advertisement delivery effect.
Description of drawings
Fig. 1 the present invention is based on the process flow diagram that video tab obtains the method for video hot information.
Fig. 2 is the refinement process flow diagram of step in the inventive method (3).
Fig. 3 is the refinement process flow diagram of step in the inventive method (5).
Embodiment
As shown in Figure 1, the present invention is based on the method that video tab obtains the video hot information and may further comprise the steps:
(1) obtain the video tab of video, video tab comprises the time point and the label substance of video;
(2) label substance to video tab carries out Chinese word segmentation, filters word segmentation result, to keep noun, verb belong to time point as video tab feature description speech;
(3) the significance level value of calculated characteristics descriptor in video tab;
In this step, adopt TF-IDF algorithm computation significance level value, specifically comprise following substep:
(31) word frequency of calculated characteristics descriptor (Term Frequency is called for short TF) value, the number of times that on behalf of certain speech, TF in this video tab, occur, concrete computing formula does
N wherein
I, jBe the number of times that feature description speech i occurs in video tab j, M is the feature description speech sum of video tab j;
(32) the reverse file frequency of calculated characteristics descriptor (Inverse Document Frequency is called for short IDF) value, IDF represents the tolerance of the general importance of certain word, and computing formula does
Wherein | D| is the sum of video tab, | { j:t
i∈ d
j| for comprising the video tab sum of feature description speech i;
(33) according to the IDF value of the TF value * feature description speech of the significance level value=feature description speech of following formula calculated characteristics descriptor.
(4) according to the significance level value feature description speech is sorted, and get the focus reserved word of K most important characteristic descriptor as this time point, in this embodiment, the value of K is 5;
(5) the focus reserved word is carried out statistical modeling,, specifically comprises following substep to generate the video hot information:
(51) take out all time points that have the focus reserved word in the video;
(52) get the maximum N of number of clicks time point as interim focus, the N value equals video length/300 second;
(53) deposit all the focus reserved words in before and after the interim focus 10 seconds in interim focus, to generate the video hot information.
Instance
Suppose that the video tab content in the current video is that " dunk shot of this penalty line take-off of Jordon is very handsome! "; through semantic participle can obtain " Jordon ", " this ", " penalty line ", " take-off ", " ", " dunk shot ", " very handsome " several speech; filter through part of speech and to obtain " Jordon ", " penalty line ", " dunk shot " two speech; calculate the TF-IDF value of two speech, suppose that here the TF-IDF value of " Jordon " is 0.4, the TF-IDF value of " penalty line " is 0.45, the TF-IDF value of " dunk shot " is 0.3, obtain " penalty line ", " Jordon ", " dunk shot " three feature description speech after the ordering.The number of tags of each time point in the statistics whole video; get the maximum k of a label time point (the k value equals video length divided by 3 quotient); this k time point is defined as interim focus; feature description speech of 10 seconds before and after each interim focus is merged in the interim focus, after the merging newly temporarily focus be the video focus.
Claims (6)
1. one kind is obtained the method for video hot information based on video tab, may further comprise the steps:
Obtain the video tab of video, said video tab comprises the time point and the label substance of said video;
Label substance to said video tab carries out Chinese word segmentation, filters word segmentation result, to keep noun, verb belong to time point as said video tab feature description speech;
Calculate the significance level value of said feature description speech in said video tab;
According to said significance level value said feature description speech is sorted, and get the focus reserved word of K most important characteristic descriptor as said time point;
Said focus reserved word is carried out statistical modeling, to generate the video hot information.
2. method according to claim 1 is characterized in that, the step of the significance level value of the said feature description speech of said calculating in said video tab comprises:
Calculate the word frequency value of said feature description speech, concrete computing formula does
N wherein
I, jBe the number of times that feature description speech i occurs in video tab j, M is the feature description speech sum of video tab j; Calculate the reverse file frequency values of said feature description speech, concrete computing formula does
Wherein | D| is the sum of video tab, | { j:t
i∈ d
j| for comprising the video tab sum of feature description speech i;
Reverse file frequency values according to the word frequency value * feature description speech of the significance level value=feature description speech of following formula calculated characteristics descriptor.
3. method according to claim 1 is characterized in that, said said focus reserved word is carried out statistical modeling, comprises with the step that generates the video hot information:
Take out all time points that have said focus reserved word in the said video;
From said all time points, select N maximum time point of number of clicks as interim focus;
All focus reserved words in before and after the said interim focus 10 seconds are deposited in the said interim focus, to generate said video hot information.
4. method according to claim 1 is characterized in that, the step of the significance level value of the said feature description speech of said calculating in said video tab is to adopt the TF-IDF algorithm.
5. method according to claim 1 is characterized in that the value of said K equals 5.
6. method according to claim 1 is characterized in that, the value of said N equals the length of said video/300 second.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2011103965154A CN102567467A (en) | 2011-12-02 | 2011-12-02 | Method for acquiring hotspot video information based on video tags |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN2011103965154A CN102567467A (en) | 2011-12-02 | 2011-12-02 | Method for acquiring hotspot video information based on video tags |
Publications (1)
Publication Number | Publication Date |
---|---|
CN102567467A true CN102567467A (en) | 2012-07-11 |
Family
ID=46412877
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN2011103965154A Pending CN102567467A (en) | 2011-12-02 | 2011-12-02 | Method for acquiring hotspot video information based on video tags |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN102567467A (en) |
Cited By (10)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103402124A (en) * | 2013-07-23 | 2013-11-20 | 百度在线网络技术(北京)有限公司 | Method and system for pushing information in video viewing process of user and cloud server |
CN105022797A (en) * | 2015-06-30 | 2015-11-04 | 北京奇艺世纪科技有限公司 | Resource topic processing method and apparatus |
CN105550277A (en) * | 2015-12-10 | 2016-05-04 | 中国传媒大学 | Intelligent movie ranking and evaluation system based on tag popularity |
CN105657575A (en) * | 2015-12-30 | 2016-06-08 | 北京奇艺世纪科技有限公司 | Video annotation methods and apparatuses |
CN106658232A (en) * | 2016-10-26 | 2017-05-10 | 广东小天才科技有限公司 | Label editing method and electronic terminal |
CN107846615A (en) * | 2016-09-20 | 2018-03-27 | 创意引晴(开曼)控股有限公司 | Visualize advertisement delivery system, put-on method and advertisement broadcast method |
CN109391829A (en) * | 2017-08-09 | 2019-02-26 | 创意引晴(开曼)控股有限公司 | Video gets position analysis system, analysis method and storage media ready |
CN110019922A (en) * | 2017-12-07 | 2019-07-16 | 北京雷石天地电子技术有限公司 | A kind of audio climax recognition methods and device |
CN110166811A (en) * | 2019-05-15 | 2019-08-23 | 口碑(上海)信息技术有限公司 | Processing method, device and the equipment of barrage information |
CN113806545A (en) * | 2021-09-24 | 2021-12-17 | 重庆理工大学 | Comment text emotion classification method based on label description generation |
Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101625695A (en) * | 2009-08-20 | 2010-01-13 | 中国科学院计算技术研究所 | Method and system for extracting complex named entities from Web video p ages |
CN101984437A (en) * | 2010-11-23 | 2011-03-09 | 亿览在线网络技术(北京)有限公司 | Music resource individual recommendation method and system thereof |
-
2011
- 2011-12-02 CN CN2011103965154A patent/CN102567467A/en active Pending
Patent Citations (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101625695A (en) * | 2009-08-20 | 2010-01-13 | 中国科学院计算技术研究所 | Method and system for extracting complex named entities from Web video p ages |
CN101984437A (en) * | 2010-11-23 | 2011-03-09 | 亿览在线网络技术(北京)有限公司 | Music resource individual recommendation method and system thereof |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103402124A (en) * | 2013-07-23 | 2013-11-20 | 百度在线网络技术(北京)有限公司 | Method and system for pushing information in video viewing process of user and cloud server |
CN105022797B (en) * | 2015-06-30 | 2018-10-09 | 北京奇艺世纪科技有限公司 | A kind of theme treating method and apparatus of resource |
CN105022797A (en) * | 2015-06-30 | 2015-11-04 | 北京奇艺世纪科技有限公司 | Resource topic processing method and apparatus |
CN105550277A (en) * | 2015-12-10 | 2016-05-04 | 中国传媒大学 | Intelligent movie ranking and evaluation system based on tag popularity |
CN105657575A (en) * | 2015-12-30 | 2016-06-08 | 北京奇艺世纪科技有限公司 | Video annotation methods and apparatuses |
CN105657575B (en) * | 2015-12-30 | 2018-10-19 | 北京奇艺世纪科技有限公司 | Video labeling method and device |
CN107846615A (en) * | 2016-09-20 | 2018-03-27 | 创意引晴(开曼)控股有限公司 | Visualize advertisement delivery system, put-on method and advertisement broadcast method |
CN106658232A (en) * | 2016-10-26 | 2017-05-10 | 广东小天才科技有限公司 | Label editing method and electronic terminal |
CN106658232B (en) * | 2016-10-26 | 2019-10-22 | 广东小天才科技有限公司 | Label editing method and electronic terminal |
CN109391829A (en) * | 2017-08-09 | 2019-02-26 | 创意引晴(开曼)控股有限公司 | Video gets position analysis system, analysis method and storage media ready |
CN110019922A (en) * | 2017-12-07 | 2019-07-16 | 北京雷石天地电子技术有限公司 | A kind of audio climax recognition methods and device |
CN110166811A (en) * | 2019-05-15 | 2019-08-23 | 口碑(上海)信息技术有限公司 | Processing method, device and the equipment of barrage information |
CN113806545A (en) * | 2021-09-24 | 2021-12-17 | 重庆理工大学 | Comment text emotion classification method based on label description generation |
CN113806545B (en) * | 2021-09-24 | 2022-06-17 | 重庆理工大学 | Comment text emotion classification method based on label description generation |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN102567467A (en) | Method for acquiring hotspot video information based on video tags | |
US10726446B2 (en) | Method and apparatus for pushing information | |
WO2019041521A1 (en) | Apparatus and method for extracting user keyword, and computer-readable storage medium | |
US8566866B1 (en) | Web identity to social media identity correlation | |
JP5717858B2 (en) | Text set matching | |
CN105138670B (en) | Audio file label generating method and system | |
EP3021264A1 (en) | Information recommendation method and apparatus in social media | |
CN106528532A (en) | Text error correction method and device and terminal | |
CN109710841A (en) | Comment on recommended method and device | |
CN110263248A (en) | A kind of information-pushing method, device, storage medium and server | |
CN105357586A (en) | Video bullet screen filtering method and device | |
CN106484764A (en) | User's similarity calculating method based on crowd portrayal technology | |
CN103778260A (en) | Individualized microblog information recommending system and method | |
CN104156436A (en) | Social association cloud media collaborative filtering and recommending method | |
CN111259271A (en) | Comment information display method and device, electronic equipment and computer readable medium | |
CN109815386B (en) | User portrait-based construction method and device and storage medium | |
CN103377258A (en) | Method and device for classification display of microblog information | |
CN108021619B (en) | Event description object recommendation method and device | |
CN104111925A (en) | Item recommendation method and device | |
US20140236969A1 (en) | Apparatus and method for recommending contents based on metadata graph | |
US9268861B2 (en) | Method and system for recommending relevant web content to second screen application users | |
CN110070410A (en) | A kind of population social activity analysis method and system based on big data | |
CN106933878B (en) | Information processing method and device | |
CN105005555A (en) | Chatting time-based keyword extraction method and device | |
Tiwari et al. | Multimodal multiplatform social media event summarization |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
C06 | Publication | ||
PB01 | Publication | ||
C10 | Entry into substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
C12 | Rejection of a patent application after its publication | ||
RJ01 | Rejection of invention patent application after publication |
Application publication date: 20120711 |