CN109408672A - A kind of article generation method, device, server and storage medium - Google Patents
A kind of article generation method, device, server and storage medium Download PDFInfo
- Publication number
- CN109408672A CN109408672A CN201811535231.7A CN201811535231A CN109408672A CN 109408672 A CN109408672 A CN 109408672A CN 201811535231 A CN201811535231 A CN 201811535231A CN 109408672 A CN109408672 A CN 109408672A
- Authority
- CN
- China
- Prior art keywords
- video
- image
- news
- relevant
- article
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Landscapes
- Television Signal Processing For Recording (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The embodiment of the invention discloses a kind of article generation method, device, server and storage mediums, wherein this method comprises: obtaining video image collection from target video；Identify an at least news relevant to target video and its video image；In conjunction with the video knowledge information and an at least news in the knowledge mapping information bank pre-established, summary info relevant to every video image is generated, and generates article relevant to the target video according to the summary info and video image collection.The embodiment of the present invention, which realizes, quickly and efficiently generates that relevant to video content is high-quality and article abundant, can satisfy demand of the user to the article of high-quality hot topic video, promotes the commercial value of Related product.
Description
Technical field
The present embodiments relate to Internet technical field more particularly to a kind of article generation method, device, server and
Storage medium.
Background technique
In the epoch of internet information explosion, is applied around the tendency of network data expansion and product is closed extensively
Note.Stage photo for various types of videos, such as in movie and television play, i.e., the high definition atlas of Highlight, passes through figure in movie and television play
Collect article, movie and television play can be helped to attract user's viewing, whether user is while watching stage photo it can be identified that the video display
Play has viewing interest.Therefore, atlas article relevant to video is to link up most direct bridge between user and these videos.
At present, it is seen that video atlas article, such as movie and television play shines article, the mostly small text for compiling pure manual edit
Then chapter is chosen excellent picture as stage photo picture, is finally carried out again to stage photo such as according to the certain popular TV plays of artificial selection
Story introduction, personage introduction etc. ultimately generate TV play stage photo atlas article.Obviously, entire article generating process requires people
Work participates in, and manual intervention is excessive, program is cumbersome, and time-consuming, low output for article generation.I.e. in addition to collecting, data are cumbersome, count
Except the disadvantages of inadequate resource and poor information timeliness, volume production is also much not achieved and updates in real time.
Summary of the invention
The embodiment of the present invention provides a kind of article generation method, device, server and storage medium, realizes quickly and efficiently
Ground generates that relevant to video content is high-quality and article abundant, can satisfy user's telling to the article of high-quality hot topic video
It asks, promotes the commercial value of Related product.
In a first aspect, the embodiment of the invention provides a kind of article generation methods, comprising:
Video image collection is obtained from target video；
Identify an at least news relevant to target video and its video image；
In conjunction with the video knowledge information and an at least news in the knowledge mapping information bank pre-established, generate with
The relevant summary info of every video image, and generated and the target video phase according to the summary info and video image collection
The article of pass.
Second aspect, the embodiment of the invention also provides a kind of article generating means, comprising:
Image collection module, for obtaining video image collection from target video；
News obtains module, for identification an at least news relevant to target video and its video image；
Article generation module, in conjunction in the knowledge mapping information bank pre-established video knowledge information and it is described at least
One news generates summary info relevant to every video image, and generates according to the summary info and video image collection
Article relevant to the target video.
The third aspect, the embodiment of the invention also provides a kind of server, the server includes:
One or more processors；
Storage device, for storing one or more programs,
When one or more of programs are executed by one or more of processors, so that one or more of processing
Device realizes a kind of article generation method as described in any embodiment of that present invention.
Fourth aspect, it is described the embodiment of the invention also provides a kind of storage medium comprising computer executable instructions
Computer executable instructions as computer processor when being executed for executing a kind of text as described in any embodiment of that present invention
Chapter generation method.
The embodiment of the present invention from target video by obtaining video image collection；Identification and target video and its video image
A relevant at least news；In conjunction with video knowledge information in the knowledge mapping information bank pre-established and at least one described
News generates summary info relevant to every video image, and according to the summary info and the generation of video image collection and institute
The relevant article of target video is stated, realizes and quickly and efficiently generates article relevant to video.Meanwhile the material of article is knot
It closes video image to obtain from news and knowledge mapping information bank, therefore, article content is high-quality and abundant, can satisfy user couple
The demand of high-quality hot topic video article, promotes the commercial value of Related product.
Detailed description of the invention
Fig. 1 is the flow chart of one of embodiment of the present invention one article generation method；
Fig. 2 is the flow chart of one of embodiment of the present invention two article generation method；
The method flow diagram of news is identified in article generation method in Fig. 3 embodiment of the present invention two；
Fig. 4 is the structural schematic diagram of one of embodiment of the present invention three article generating means；
Fig. 5 is the structural schematic diagram of one of the embodiment of the present invention four server.
Specific embodiment
The present invention is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining the present invention rather than limiting the invention.It also should be noted that in order to just
Only the parts related to the present invention are shown in description, attached drawing rather than entire infrastructure.
Embodiment one
Fig. 1 is a kind of flow chart for article generation method that the embodiment of the present invention one provides, and the present embodiment is applicable to text
The generation situation of chapter, this method can be executed by article generating means, which can use hardware and/or software realization,
And can configure in the server, this method specifically includes:
S110, video image collection is obtained from target video.
Target video can be video relevant to the article that will be made or generate, and may include short-sighted frequency, micro- electricity
Shadow, TV play and film etc., video image collection can be the video image extracted from target video and constitute.If it is making
Make stage photo article, then target video can be movie and television play relevant to the stage photo article that will be made or generate.It is exemplary
, stage photo article is made for film B, then film B is target video (target movie and television play).In production movie and television play article
When, it needs first to select video image relevant to target movie and television play, specifically can directly identify from movie and television play library, can such as lead to
It crosses in the video of movie and television play and parses corresponding stage photo.Movie and television play library can be existing Video Reservoir, include a large amount of
The relevant information of movie and television play, such as the title including TV play or film, shooting time, related stage photo and clarity etc..Video
Image set can be the Candidate Set formed for the related stage photo of one or more of target movie and television play.Here, can also be from other
Video image collection is obtained in existing information knowledge database, for example encyclopaedia database, the present embodiment carry out video image collection
Source side is without limitation.
S120, an identification at least news relevant to target video and its video image.
Video image can be one or more image relevant to video of video image concentration.Making related atlas
In addition to image also needs other description informations when article, for example, in the material for making stage photo article in addition to stage photo also needs therewith
Relevant literal expression or the information of description.
In the present embodiment, at least one relevant to video images such as movie and television plays can be identified from news library
News generates text relevant to image in video using news.Wherein, news library includes a large amount of news information, then,
One or more news relevant to video and its video image that can be selected from news library.It is by movie and television play of video
Example, concentrating such as the video image of TV play C includes 5 stage photos, and selects this 5 stage photos as video image, further, can
To identify news information relevant to TV play C and this 5 stage photos from news library.Certainly, identify that the rule of news can
To be the recognition rule or strategy pre-established, illustratively, it can be identification and the character in video image or play the part of
The related news information of role；It is also possible to carry out identification phase using the identification model completed by the training of corresponding sample data
Close news.Here, news can also be obtained from other existing information knowledge databases, the present embodiment carrys out related news
Source is without limitation.
S130, in conjunction with the video knowledge information and an at least news in the knowledge mapping information bank pre-established,
Summary info relevant to every video image is generated, and is generated and the target according to the summary info and video image collection
The relevant article of video.
Wherein, it includes more fully video related information that the knowledge mapping information bank pre-established, which can be pre-establish,
Database, may include the relevant knowledge information of video, knowledge information can be production information and content information of video etc..
For example, the knowledge information of the videos such as movie and television play, micro- film can include at least acute name, plot, performer, role, broadcast time and
Production Time etc..Summary info can be summary or main information for video information.It may further be according to video image
And corresponding summary info combination constitutes article relevant to the video.Illustratively, it is right below each video image to can be
Corresponding summary info should be added, atlas article is ultimately formed.
By taking the generation of the stage photo article of movie and television play as an example, the knowledge information and an at least news of movie and television play can use,
Summary info relevant to stage photo is generated, summary info, which can be, to be extracted or analyzed from the knowledge information and news of movie and television play
Information relevant to stage photo out, such as the description recommended information of stage photo.Of course it is possible to use more news generate corresponding stage photo
Summary info.Then the stage photo article of movie and television play is generated using every stage photo and corresponding summary info.
Optionally, from target video obtain video image collection before, the method also includes:
Popular video is excavated from Video Reservoir, using the hot topic video as the target video.
Popular video can refer to certain time, in a certain range, the video that the public or online friend pay close attention to the most.For example, electric
Audience ratings depending on play A is very big compared with the topic discussion amount of other TV plays of contemporaneity height or TV play A, then TV play A
Popular TV play can be considered as, correspondingly, can be using TV play A as target movie and television play.Using popular video as video correlation
Article can more meet public reading requirement, also can more attract the eyeball of spectators.
It is optionally, described that popular video is excavated from Video Reservoir, comprising:
It obtains in the Video Reservoir to the description information of each video, wherein the description information is used for from different dimensional
Video is introduced in degree；
The selection feature of each video is grabbed from internet according to the description information, wherein the selection feature is used for
Video temperature is evaluated from different perspectives；
The temperature total score that the selection feature is determined to each video by the method for weighted sum, is excavated according to temperature total score
Popular video out.
Specifically, description information can be the specifying information for describing or illustrating video, wherein the description information is extremely
It less include title, shooting time and clarity.For example, can also include: each video for the description information for excavating popular video
Character and do not play person accordingly.Each video is grabbed from internet or other databases according to these description informations
Feature is selected, preset crawl model is can use or existing gripping tool and software etc. carries out the crawl of selection feature.
Select feature for evaluating video temperature from different perspectives, for example, movie and television play temperature, Yan Yuanre can be included at least
The dimensions such as degree and broadcasting temperature.It illustratively, can also be by the reading of corresponding novel if movie and television play is adapted according to novel
Temperature alternatively feature.Using the temperature total score for determining each video to the method for different selection characteristic weighing summations, wherein not
Weighted value with feature can be different type previously according to business demand or video and demand setting.For example, for one
It is larger that the corresponding weighted value of performer's temperature can be set in a little idol TV plays.Heat is finally determined according to the temperature total score of each video
Door video, the temperature such as the higher video of temperature total score is higher, can be considered as popular video.
The technical solution of the embodiment of the present invention is further known in conjunction with video in conjunction with the video image and related news selected
Know information and news generate the summary info of video image, ultimately generates article using video image and corresponding summary info,
It realizes and quickly and efficiently generates article relevant to video.Meanwhile the material of article is from news and knowledge mapping information
It is obtained in library, therefore, content is high-quality and abundant, can satisfy demand of the user to the atlas article of high-quality hot topic video, mentions
Rise the commercial value of Related product.
Embodiment two
Fig. 2 is a kind of flow chart of article generation method provided by Embodiment 2 of the present invention, on the basis of above-described embodiment
Upper further progress optimization, as shown in Fig. 2, this method specifically includes:
S210, the video source of target video is parsed, with the second for singly in all videos frame image analytically obtained
The random video frame images for obtaining preset quantity in position, obtain multiple candidate video frame images of target video.
Video image in order to obtain can be parsed by the video source to target video, obtain all videos frame figure
Then picture therefrom selects multiple candidate video frame images according to certain rules.For example, there are 24 frame video images for per second
Video, by it is per second randomly select 3 in the way of obtain multiple images in video source, to obtain the video of preset quantity
Frame image.
Identification feature in every S220, acquisition candidate video frame image, the high quality image for utilizing preparatory training to obtain are known
Other model and in conjunction with the identification feature, screens multiple described candidate video frame images, obtains video image Candidate Set,
Wherein, the identification feature includes at least one feature for measuring video frame images quality.
After getting multiple video frame images, need to carry out quality identification or filtering to these images.Can specifically it lead to
It crosses the high-quality stage photo identification model that training obtains in advance to identify image, for stage photo article, needs which is identified
It is the high-quality stage photo that can be used a bit.The high-quality stage photo identification model can be based on deep learning RESNET's (resent)
Image classification algorithms, in combination with the priori knowledge of stage photo, such as than such as whether being closed comprising performer, performer's number, performer
The identification features such as whether eye, picture fuzzy are used as sample data, obtain to stage photo two classification learnings of progress.In addition, about play
According to priori knowledge, can be identified and be provided by other third party's services.
When carrying out the identification of high-quality stage photo, different stage photos and corresponding priori knowledge can be input to the identification mould
Type obtains the quality value of every stage photo, when quality value be higher than preset threshold, it may be considered that the stage photo be high-quality stage photo, can
Think that stage photo article is used.
S230, duplicate removal is carried out according to image similarity to the video frame images in the video image Candidate Set, obtains institute
State video image collection.
Due to filtered high quality image may exist repeat or it is similar, this can reduce the reading experience of reader.Cause
This, can carry out deduplication operation to the stage photo after screening again.Illustratively, the mistake of similar image can be carried out using SITF algorithm
Filter.
S240, attribute information relevant to target movie and television play is grabbed from internet, wherein the attribute information at least wraps
Include video name, Dao Yanming, Yan Yuanming, performer's role and plot.
Attribute information can be the relevant information for describing movie and television play, have in further detail in particular for movie and television play itself
The information of body can also include other phases in addition to including video name, Dao Yanming, Yan Yuanming, performer's role and plot
Information is closed, for example can also include the recommended informations such as the corresponding historical background of plot if it is serious drama.It specifically for example, can benefit
The correlation attribute information of crawl with target movie and television play is carried out from internet with existing gripping tool or software.
S250, the name of the performer in conjunction with involved in the video image, according to the news in the attribute information and news library
It is compared, and during comparison, according to the hits of attribute information type, the hit frequency of the attribute information of all types
Secondary and news number of words threshold value, identifies news relevant to target video and its video image.
During identifying news relevant to target video and video image, can first it be related to according in video image
And performer's name, and the attribute information of target video got, by different dimension or feature with it is new in news library
News is compared, for example comparison is the features such as acute name, performer, role the case where.It specifically can choose acute name hits, Yan Yuanming
Middle number, role's hits, the hit frequency, news number of words, in title whether comprising in movie and television play name information, title whether include
The features such as performer's title are as filter condition.Wherein, hits can be the Characteristic Number hit altogether in these features, hit
The frequency can be the number that these features occur altogether in news.
Illustratively, Fig. 3 is the method flow diagram of identification news provided in an embodiment of the present invention, as shown in figure 3, first may be used
To filter out some low-quality news from news library, such as news content is imperfect or the low-qualitys news such as content is unhealthy,
Then it is counted in news for features described above, if obtaining in above-mentioned each feature through statistics, hits are greater than 3 and hit
The frequency is greater than 5, and news number of words is less than 300, it may be considered that the news meets the requirements, can be used as final news.Certainly,
The rule that specific identification news can be preset is also possible to be known using preparatory trained news identification model
Not, news information relevant to movie and television play can be screened from the news data of magnanimity.
S260, the video image for randomly selecting preset quantity is concentrated from the video image, generate video atlas.
Since video image concentrates the image including one or more, a certain number of figures can be therefrom randomly choosed
Picture.For example, being directed to the production of stage photo article, 5 can be needed according to pre-set stage photo quantity or stage photo article template
Stage photo then can concentrate random selection 5 from video image and be used as stage photo atlas, to use when the generation of subsequent stage photo article.
S270, believed according to the identification feature of every image in described image atlas and attribute relevant to target video
Breath obtains video knowledge information relevant to the identification feature and attribute information from the knowledge mapping information bank.
Knowledge mapping information bank can be pre-created, and a large amount of useful information has been described.In the present embodiment, it needs
The article that picture and text combine is generated in conjunction with video image, therefore, can be obtained from knowledge mapping information bank and the knowledge of video image
The attribute information associated video knowledge information of other feature and target video, to match text for atlas article.And knowledge mapping is believed
Information in breath library generally has certain correctness, and covering is more comprehensive, therefore, using in knowledge mapping information bank
Information architecture atlas article can increase the accuracy of article, and it is readable to increase article.It, can be with for example, for single star's stage photo
The phase of the relevant knowledge information of the star and the related current goal movie and television play of the star is obtained from knowledge mapping information bank
Close knowledge information.
S280, believed according to the identification feature of every image in the video atlas and attribute relevant to target video
Breath, and in conjunction with the video knowledge information, it is generated from an at least news related to every image in video atlas
Summary info.
Summary info be as in atlas article with image with Wen Suoyong, can be mentioned from the news having identified that
It takes, further, it is not more than 3 that news quantity, which can also be arranged, i.e., extracts from 1-2 news, so as to most throughout one's life
At summary info have good consistency, increase legibility.For stage photo article, need according to every play when extraction
According to identification feature generate associated summary info, such as in stage photo performer and current target movie and television play it is related
Summary info.Further, it is also possible in conjunction with the movie and television play knowledge information got from knowledge mapping information bank, as to new
The text of news supplements.And as the method for extracting abstract from known text, in embodiments of the present invention, existing skill can be used
Any extraction-type or production summarization generation algorithm in art, therefore which is not described herein again.
S290, every image in video atlas is combined with corresponding summary info, is obtained and target video phase
The article of pass.
Illustratively, if concentrating 8-10 high picture qualities of random selection candidate as stage photo atlas from video image, so
After use abstract corresponding with every stage photo as atlas introduction, combine and obtain final stage photo article.Certainly, if abstract letter
Data volume deficiency is ceased, the video display story introduction of encyclopaedia or performer can also be introduced to the supplement as atlas introduction.
The embodiment of the present invention passes through using default knowledge mapping information bank, while utilizing advanced depth learning technology, from
Largely, good movie and television play is obtained in video data in a jumble to shine while being stage photo life using natural Domestic News advantage
At the summary info for describing, commenting on stage photo and movie and television play, stage photo and summary info are finally combined into the atlas of video
Article can not only automate generation, but also realization is quick, efficient and abundant in content, can satisfy user to high-quality hot topic
The demand of the atlas article of video, promotes the commercial value of Related product.
Embodiment three
Fig. 4 is a kind of structural schematic diagram for article generating means that the embodiment of the present invention three provides, as shown in figure 4, described
Device includes:
Image collection module 410, for obtaining video image collection from target video.News obtains module 420, for knowing
An at least news not relevant to target video and its video image.
Article generation module 430, for combining video knowledge information and institute in the knowledge mapping information bank pre-established
An at least news is stated, generates summary info relevant to every video image, and according to the summary info and video image
Collection generates article relevant to the target video.
Optionally, described device further include:
Target video excavates module, is used in image collection module 410 before obtaining video image collection in target video,
Popular video is excavated from Video Reservoir, using the hot topic video as the target video.
Optionally, the target video excavation module includes:
Description information acquiring unit, for obtaining in the Video Reservoir to the description information of each video, wherein described
Description information is for being introduced video from different dimensions；
Feature picking unit is selected, for the selection feature of each video to be grabbed from internet according to the description information,
Wherein, the selection feature for evaluating video temperature from different perspectives；
Temperature total score computing unit, for the selection feature will to be determined to the heat of each video by the method for weighted sum
Total score is spent, popular video is excavated according to temperature total score.
Optionally, described image acquisition module 410 includes:
Video source resolution unit is parsed for the video source to target video, all videos frame analytically obtained
The video frame images for obtaining preset quantity in image at random in seconds, obtain multiple candidate video frame figures of target video
Picture；
Video Candidate Set screening unit is obtained, for obtaining the identification feature in every candidate video frame image, using preparatory
The obtained high quality image identification model of training and in conjunction with the identification feature, sieves multiple described candidate video frame images
Choosing, obtains video image Candidate Set, wherein the identification feature includes at least one spy for measuring video frame images quality
Sign；
Duplicate removal unit, for being gone to the video frame images in the video image Candidate Set according to image similarity
Weight, obtains the video image collection.
Optionally, the news acquisition module 420 includes:
Attribute information picking unit, for grabbing attribute information relevant to target video from internet, wherein described
Attribute information includes at least movie and television play title, Dao Yanming, Yan Yuanming, performer's role and plot；
News recognition unit, for the name of the performer in conjunction with involved in the video image, according to the attribute information and newly
News in Wen Ku is compared, and during comparison, according to the hits of attribute information type, the attribute of all types
The hit frequency and news number of words threshold value of information, identify news relevant to target video and its video image.
Optionally, the article generation module 430 includes:
Video atlas selection unit, it is raw for concentrating the video image for randomly selecting preset quantity from the video image
At video atlas；
Knowledge information extraction unit is regarded for the identification feature according to every image in the video atlas and with target
Frequently relevant attribute information obtains video relevant to the identification feature and attribute information from the knowledge mapping information bank
Knowledge information；
Summary info generation unit is regarded for the identification feature according to every image in the video atlas and with target
Frequently relevant attribute information, and in conjunction with the video knowledge information, from generated in an at least news in video atlas
The relevant summary info of every image；
Article assembled unit is obtained for every image in video atlas to be combined with corresponding summary info
Article relevant to target video.
It is raw that article provided by any embodiment of the invention can be performed in article generating means provided by the embodiment of the present invention
At method, have the corresponding functional module of execution method and beneficial effect.The not technical detail of detailed description in the present embodiment,
It can be found in a kind of article generation method that any embodiment of that present invention provides.
Example IV
Referring to Fig. 5, a kind of server 500 is present embodiments provided comprising: one or more processors 520；Storage dress
510 are set, for storing one or more programs, when one or more of programs are held by one or more of processors 520
Row, so that one or more of processors 520 realize a kind of article generation method provided by the embodiment of the present invention, comprising:
Video image collection is obtained from target video；
Identify an at least news relevant to target video and its video image；
In conjunction with the video knowledge information and an at least news in the knowledge mapping information bank pre-established, generate with
The relevant summary info of every video image, and generated and the target video phase according to the summary info and video image collection
The article of pass.
Certainly, it will be understood by those skilled in the art that processor 520 can also realize that any embodiment of that present invention is provided
A kind of article generation method technical solution.
The server 500 that Fig. 5 is shown is only an example, should not function and use scope band to the embodiment of the present invention
Carry out any restrictions.
As shown in figure 5, server 500 is showed in the form of universal computing device.The component of server 500 may include but
Be not limited to: one or more processor 520, storage device 510 connect different system components (including storage device 510 and place
Manage device 520) bus 550.
Bus 550 indicates one of a few class bus structures or a variety of, including memory bus or Memory Controller,
Peripheral bus, graphics acceleration port, processor or the local bus using any bus structures in a variety of bus structures.It lifts
For example, these architectures include but is not limited to industry standard architecture (ISA) bus, microchannel architecture (MAC)
Bus, enhanced isa bus, Video Electronics Standards Association (VESA) local bus and peripheral component interconnection (PCI) bus.
Server 500 typically comprises a variety of computer system readable media.These media can be it is any being capable of bedding and clothing
The usable medium that business device 500 accesses, including volatile and non-volatile media, moveable and immovable medium.
Storage device 510 may include the computer system readable media of form of volatile memory, such as arbitrary access
Memory (RAM) 511 and/or cache memory 512.Server 500 may further include other removable/not removable
Dynamic, volatile/non-volatile computer system storage medium.Only as an example, storage system 513 can be used for read and write can not
Mobile, non-volatile magnetic media (Fig. 5 do not show, commonly referred to as " hard disk drive ").Although being not shown in Fig. 5, Ke Yiti
For the disc driver for being read and write to removable non-volatile magnetic disk (such as " floppy disk "), and to moving non-volatile light
The CD drive of disk (such as CD-ROM, DVD-ROM or other optical mediums) read-write.In these cases, each driver
It can be connected by one or more data media interfaces with bus 550.Storage device 510 may include at least one program
Product, the program product have one group of (for example, at least one) program module, these program modules are configured to perform the present invention
The function of each embodiment.
Program/utility 514 with one group of (at least one) program module 515 can store in such as storage dress
It sets in 510, such program module 515 includes but is not limited to operating system, one or more application program, other program moulds
It may include the realization of network environment in block and program data, each of these examples or certain combination.Program module
515 usually execute function and/or method in any embodiment described in the invention.
Server 500 can also be with one or more external equipments 560 (such as keyboard, sensing equipment, display 570 etc.)
Communication, can also be enabled a user to one or more equipment interact with the server 500 communicate, and/or with make the clothes
Any equipment (such as network interface card, modem etc.) that business device 500 can be communicated with one or more of the other calculating equipment
Communication.This communication can be carried out by input/output (I/O) interface 530.Also, server 500 can also be suitable by network
Orchestration 540 and one or more network (such as local area network (LAN), wide area network (WAN) and/or public network, such as because of spy
Net) communication.As shown in figure 5, network adapter 540 is communicated by bus 550 with other modules of server 500.It should be understood that
Although not shown in the drawings, other hardware and/or software module can be used in conjunction with server 500, including but not limited to: micro- generation
Code, device driver, redundant processing unit, external disk drive array, RAID system, tape drive and data backup are deposited
Storage system etc..
The program that processor 520 is stored in storage device 510 by operation, thereby executing various function application and number
According to processing, such as realize a kind of article generation method provided by the embodiment of the present invention.
Embodiment five
The embodiment of the present invention five provides a kind of storage medium comprising computer executable instructions, and the computer is executable
Instruction is used to execute a kind of article generation method when being executed by computer processor, this method comprises:
Video image collection is obtained from target video；
Identify an at least news relevant to target video and its video image；
In conjunction with the video knowledge information and an at least news in the knowledge mapping information bank pre-established, generate with
The relevant summary info of every video image, and generated and the target video phase according to the summary info and video image collection
The article of pass.
Certainly, a kind of storage medium comprising computer executable instructions, computer provided by the embodiment of the present invention
The method operation that executable instruction is not limited to the described above, can also be performed a kind of article provided by any embodiment of the invention
Relevant operation in generation method.
The computer storage medium of the embodiment of the present invention, can be using any of one or more computer-readable media
Combination.Computer-readable medium can be computer-readable signal media or computer readable storage medium.It is computer-readable
Storage medium for example may be-but not limited to-the system of electricity, magnetic, optical, electromagnetic, infrared ray or semiconductor, device or
Device, or any above combination.The more specific example (non exhaustive list) of computer readable storage medium includes: tool
There are electrical connection, the portable computer diskette, hard disk, random access memory (RAM), read-only memory of one or more conducting wires
(ROM), erasable programmable read only memory (EPROM or flash memory), optical fiber, portable compact disc read-only memory (CD-
ROM), light storage device, magnetic memory device or above-mentioned any appropriate combination.In this document, computer-readable storage
Medium can be any tangible medium for including or store program, which can be commanded execution system, device or device
Using or it is in connection.
Computer-readable signal media may include in a base band or as carrier wave a part propagate data-signal,
Wherein carry computer-readable program code.The data-signal of this propagation can take various forms, including but unlimited
In electromagnetic signal, optical signal or above-mentioned any appropriate combination.Computer-readable signal media can also be that computer can
Any computer-readable medium other than storage medium is read, which can send, propagates or transmit and be used for
By the use of instruction execution system, device or device or program in connection.
The program code for including on computer-readable medium can transmit with any suitable medium, including --- but it is unlimited
In wireless, electric wire, optical cable, RF etc. or above-mentioned any appropriate combination.
The computer for executing operation of the present invention can be write with one or more programming languages or combinations thereof
Program code, described program design language include object oriented program language-such as Java, Smalltalk, C++,
It further include conventional procedural programming language-such as " C " language or similar programming language.Program code can be with
It fully executes, partly execute on the user computer on the user computer, being executed as an independent software package, portion
Divide and partially executes or executed on a remote computer or server completely on the remote computer on the user computer.?
Be related in the situation of remote computer, remote computer can pass through the network of any kind --- including local area network (LAN) or
Wide area network (WAN)-be connected to subscriber computer, or, it may be connected to outer computer (such as mentioned using Internet service
It is connected for quotient by internet).
Note that the above is only a better embodiment of the present invention and the applied technical principle.It will be appreciated by those skilled in the art that
The invention is not limited to the specific embodiments described herein, be able to carry out for a person skilled in the art it is various it is apparent variation,
It readjusts and substitutes without departing from protection scope of the present invention.Therefore, although being carried out by above embodiments to the present invention
It is described in further detail, but the present invention is not limited to the above embodiments only, without departing from the inventive concept, also
It may include more other equivalent embodiments, and the scope of the invention is determined by the scope of the appended claims.
Claims (14)
1. a kind of article generation method, which is characterized in that the described method includes:
Video image collection is obtained from target video；
Identify an at least news relevant to target video and its video image；
In conjunction with the video knowledge information and an at least news in the knowledge mapping information bank pre-established, generate and every
The relevant summary info of video image, and it is relevant to the target video according to the summary info and the generation of video image collection
Article.
2. the method according to claim 1, wherein from target video obtain video image collection before, institute
State method further include:
Popular video is excavated from Video Reservoir, using the hot topic video as the target video.
3. according to the method described in claim 2, it is characterized in that, described excavate popular video, packet from Video Reservoir
It includes:
It obtains in the Video Reservoir to the description information of each video, wherein the description information is used for from different dimensions pair
Video is introduced；
The selection feature of each video is grabbed from internet according to the description information, wherein the selection feature is for never
Video temperature is evaluated with angle；
The temperature total score that the selection feature is determined to each video by the method for weighted sum, excavates heat according to temperature total score
Door video.
4. the method according to claim 1, wherein described obtain video image collection from target video, comprising:
The video source of target video is parsed, it is random in seconds in all videos frame image analytically obtained to obtain
The video frame images of preset quantity obtain multiple candidate video frame images of target video；
The identification feature in every candidate video frame image is obtained, the high quality image identification model obtained using preparatory training is simultaneously tied
The identification feature is closed, multiple described candidate video frame images are screened, obtain video image Candidate Set, wherein described
Identification feature includes at least one feature for measuring video frame images quality；
Duplicate removal is carried out according to image similarity to the video frame images in the video image Candidate Set, obtains the video image
Collection.
5. according to the method described in claim 4, it is characterized in that, the identification is relevant to target video and its video image
An at least news, comprising:
Relevant to target video attribute information is grabbed from internet, wherein the attribute information including at least video name,
Direct name, Yan Yuanming, performer's role and plot；
The name of the performer in conjunction with involved in the video image, is compared according to the attribute information with the news in news library,
And during comparison, according to the hit frequency and news of the hits of attribute information type, the attribute information of all types
Number of words threshold value identifies news relevant to target video and its video image.
6. according to the method described in claim 5, it is characterized in that, in conjunction with the video in the knowledge mapping information bank pre-established
Knowledge information and an at least news generate summary info relevant to every video image, and believe according to the abstract
Breath and video image collection generate article relevant to the target video, comprising:
The video image for randomly selecting preset quantity is concentrated from the video image, generates video atlas；
According to the identification feature of every image in the video atlas and attribute information relevant to target video, know from described
Know and obtains video knowledge information relevant to the identification feature and attribute information in profile information library；
According to the identification feature of every image in the video atlas and attribute information relevant to target video, and combine institute
Video knowledge information is stated, summary info relevant to every image in video atlas is generated from an at least news；
Every image in video atlas is combined with corresponding summary info, obtains article relevant to target video.
7. a kind of article generating means, which is characterized in that described device includes:
Image collection module, for obtaining video image collection from target video；
News obtains module, for identification an at least news relevant to target video and its video image；
Article generation module, in conjunction with the video knowledge information and described at least one in the knowledge mapping information bank pre-established
News generates summary info relevant to every video image, and generate according to the summary info and video image collection and
The relevant article of the target video.
8. device according to claim 7, which is characterized in that described device further include:
Target video excavates module, is used in image collection module before obtaining video image collection in target video, from video
Popular video is excavated in resources bank, using the hot topic video as the target video.
9. device according to claim 8, which is characterized in that the target video excavates module and includes:
Description information acquiring unit, for obtaining in the Video Reservoir to the description information of each video, wherein the description
Information is for being introduced video from different dimensions；
Feature picking unit is selected, for grabbing the selection feature of each video from internet according to the description information, wherein
The selection feature for evaluating video temperature from different perspectives；
Temperature total score computing unit, for that the selection feature will be determined that the temperature of each video is total by the method for weighted sum
Point, popular video is excavated according to temperature total score.
10. device according to claim 7, which is characterized in that described image obtains module and includes:
Video source resolution unit is parsed for the video source to target video, all videos frame image analytically obtained
In in seconds at random obtain preset quantity video frame images, obtain multiple candidate video frame images of target video；
Video Candidate Set screening unit is obtained, for obtaining the identification feature in every candidate video frame image, utilizes preparatory training
Obtained high quality image identification model and in conjunction with the identification feature, screens multiple described candidate video frame images, obtains
To video image Candidate Set, wherein the identification feature includes at least one feature for measuring video frame images quality；
Duplicate removal unit is obtained for carrying out duplicate removal according to image similarity to the video frame images in the video image Candidate Set
To the video image collection.
11. device according to claim 10, which is characterized in that the news obtains module and includes:
Attribute information picking unit, for grabbing attribute information relevant to target video from internet, wherein the attribute
Information includes at least movie and television play title, Dao Yanming, Yan Yuanming, performer's role and plot；
News recognition unit, for the name of the performer in conjunction with involved in the video image, according to the attribute information and news library
In news be compared, and during comparison, according to the hits of attribute information type, the attribute information of all types
The hit frequency and news number of words threshold value, news relevant to target video and its video image is identified.
12. device according to claim 11, which is characterized in that the article generation module includes:
Video atlas selection unit generates view for concentrating the video image for randomly selecting preset quantity from the video image
Frequency atlas；
Knowledge information extraction unit, for according to the identification feature of every image in the video atlas and with target video phase
The attribute information of pass obtains video knowledge relevant to the identification feature and attribute information from the knowledge mapping information bank
Information；
Summary info generation unit, for according to the identification feature of every image in the video atlas and with target video phase
The attribute information of pass, and in conjunction with the video knowledge information, generated from an at least news with it is every in video atlas
Open the relevant summary info of image；
Article assembled unit obtains and mesh for every image in video atlas to be combined with corresponding summary info
Mark the relevant article of video.
13. a kind of server, which is characterized in that the server includes:
One or more processors；
Storage device, for storing one or more programs,
When one or more of programs are executed by one or more of processors, so that one or more of processors are real
Now such as a kind of article generation method as claimed in any one of claims 1 to 6.
14. a kind of storage medium comprising computer executable instructions, the computer executable instructions are by computer disposal
For executing a kind of such as article generation method as claimed in any one of claims 1 to 6 when device executes.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811535231.7A CN109408672B (en) | 2018-12-14 | 2018-12-14 | Article generation method, article generation device, server and storage medium |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811535231.7A CN109408672B (en) | 2018-12-14 | 2018-12-14 | Article generation method, article generation device, server and storage medium |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109408672A true CN109408672A (en) | 2019-03-01 |
| CN109408672B CN109408672B (en) | 2020-09-29 |
Family
ID=65459369
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811535231.7A Active CN109408672B (en) | 2018-12-14 | 2018-12-14 | Article generation method, article generation device, server and storage medium |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109408672B (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110377891A (en) * | 2019-06-19 | 2019-10-25 | 北京百度网讯科技有限公司 | Generation method, device, equipment and the computer readable storage medium of event analysis article |
| CN113569065A (en) * | 2021-07-28 | 2021-10-29 | 北京达佳互联信息技术有限公司 | Topic list display method, device, system, equipment and storage medium |
| CN113660541A (en) * | 2021-07-16 | 2021-11-16 | 北京百度网讯科技有限公司 | News video abstract generation method and device |
| CN113795882A (en) * | 2019-09-27 | 2021-12-14 | 华为技术有限公司 | Emotion-based multimedia content summarization |
| CN114417808A (en) * | 2022-02-25 | 2022-04-29 | 北京百度网讯科技有限公司 | Article generation method and device, electronic equipment and storage medium |
| CN115334367A (en) * | 2022-07-11 | 2022-11-11 | 北京达佳互联信息技术有限公司 | Video summary information generation method, device, server and storage medium |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103945234A (en) * | 2014-03-27 | 2014-07-23 | 百度在线网络技术（北京）有限公司 | Video-related information providing method and device |
| US20170052964A1 (en) * | 2015-08-19 | 2017-02-23 | International Business Machines Corporation | Video clips generation system |
| CN106844322A (en) * | 2017-01-22 | 2017-06-13 | 百度在线网络技术（北京）有限公司 | Intelligent article generation method and device |
| CN106997387A (en) * | 2017-03-28 | 2017-08-01 | 中国科学院自动化研究所 | The multi-modal automaticabstracting matched based on text image |
| CN107239801A (en) * | 2017-06-28 | 2017-10-10 | 安徽大学 | Video attribute represents that learning method and video text describe automatic generation method |
| CN107316248A (en) * | 2017-06-26 | 2017-11-03 | 高岚 | A kind of system and method for increasing explanatory note for picture and generating blog article |
| CN108062364A (en) * | 2017-12-05 | 2018-05-22 | 优酷网络技术（北京）有限公司 | Information displaying method and device |
-
2018
- 2018-12-14 CN CN201811535231.7A patent/CN109408672B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103945234A (en) * | 2014-03-27 | 2014-07-23 | 百度在线网络技术（北京）有限公司 | Video-related information providing method and device |
| US20170052964A1 (en) * | 2015-08-19 | 2017-02-23 | International Business Machines Corporation | Video clips generation system |
| CN106844322A (en) * | 2017-01-22 | 2017-06-13 | 百度在线网络技术（北京）有限公司 | Intelligent article generation method and device |
| CN106997387A (en) * | 2017-03-28 | 2017-08-01 | 中国科学院自动化研究所 | The multi-modal automaticabstracting matched based on text image |
| CN107316248A (en) * | 2017-06-26 | 2017-11-03 | 高岚 | A kind of system and method for increasing explanatory note for picture and generating blog article |
| CN107239801A (en) * | 2017-06-28 | 2017-10-10 | 安徽大学 | Video attribute represents that learning method and video text describe automatic generation method |
| CN108062364A (en) * | 2017-12-05 | 2018-05-22 | 优酷网络技术（北京）有限公司 | Information displaying method and device |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110377891A (en) * | 2019-06-19 | 2019-10-25 | 北京百度网讯科技有限公司 | Generation method, device, equipment and the computer readable storage medium of event analysis article |
| CN113795882A (en) * | 2019-09-27 | 2021-12-14 | 华为技术有限公司 | Emotion-based multimedia content summarization |
| CN113660541A (en) * | 2021-07-16 | 2021-11-16 | 北京百度网讯科技有限公司 | News video abstract generation method and device |
| CN113660541B (en) * | 2021-07-16 | 2023-10-13 | 北京百度网讯科技有限公司 | Method and device for generating abstract of news video |
| CN113569065A (en) * | 2021-07-28 | 2021-10-29 | 北京达佳互联信息技术有限公司 | Topic list display method, device, system, equipment and storage medium |
| CN114417808A (en) * | 2022-02-25 | 2022-04-29 | 北京百度网讯科技有限公司 | Article generation method and device, electronic equipment and storage medium |
| CN114417808B (en) * | 2022-02-25 | 2023-04-07 | 北京百度网讯科技有限公司 | Article generation method and device, electronic equipment and storage medium |
| CN115334367A (en) * | 2022-07-11 | 2022-11-11 | 北京达佳互联信息技术有限公司 | Video summary information generation method, device, server and storage medium |
| CN115334367B (en) * | 2022-07-11 | 2023-10-17 | 北京达佳互联信息技术有限公司 | Method, device, server and storage medium for generating abstract information of video |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109408672B (en) | 2020-09-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US12094209B2 (en) | Video data processing method and apparatus, device, and medium | |
| CN109408672A (en) | A kind of article generation method, device, server and storage medium | |
| CN109145784B (en) | Method and apparatus for processing video | |
| CN108322788B (en) | Advertisement display method and device in live video | |
| CN109376603A (en) | A kind of video frequency identifying method, device, computer equipment and storage medium | |
| US10834379B2 (en) | 2D-to-3D video frame conversion | |
| US9626585B2 (en) | Composition modeling for photo retrieval through geometric image segmentation | |
| CN101877060A (en) | Messaging device and method and program | |
| CN102214304A (en) | Information processing apparatus, information processing method and program | |
| KR20120003447A (en) | Clustering videos by location | |
| CN104320670A (en) | Summary information extracting method and system for network video | |
| WO2023056835A1 (en) | Video cover generation method and apparatus, and electronic device and readable medium | |
| CN111754254A (en) | System and method for dynamically creating and inserting immersive promotional content in multimedia | |
| Yeh et al. | Relative features for photo quality assessment | |
| CN110519620A (en) | Recommend the method and television set of TV programme in television set | |
| CN115222858A (en) | Method and equipment for training animation reconstruction network and image reconstruction and video reconstruction thereof | |
| CN115379290A (en) | Video processing method, device, equipment and storage medium | |
| Husa et al. | HOST-ATS: automatic thumbnail selection with dashboard-controlled ML pipeline and dynamic user survey | |
| CN108833964A (en) | A kind of real-time successive frame Information Embedding identifying system | |
| Zhou et al. | Modeling perspective effects in photographic composition | |
| Godi et al. | Indirect match highlights detection with deep convolutional neural networks | |
| US20170061215A1 (en) | Clustering method using broadcast contents and broadcast related data and user terminal to perform the method | |
| WO2022072794A1 (en) | Prediction of nba talent and quality from non-professional tracking data | |
| CN112449249A (en) | Video stream processing method and device, electronic equipment and storage medium | |
| CN108600864A (en) | A kind of preview generation method and device |
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 |