CN102662950B - Active advertisement generation method based on Internet picture - Google Patents
Active advertisement generation method based on Internet picture Download PDFInfo
- Publication number
- CN102662950B CN102662950B CN201210047729.5A CN201210047729A CN102662950B CN 102662950 B CN102662950 B CN 102662950B CN 201210047729 A CN201210047729 A CN 201210047729A CN 102662950 B CN102662950 B CN 102662950B
- Authority
- CN
- China
- Prior art keywords
- picture
- advertisement
- database
- advertising
- image
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Abstract
The invention discloses an active advertisement generation method based on Internet pictures, belonging to the fields of image identification and computer software system integration. The generation method comprises a first step of setting a picture database, an advertisement database, an image feature database, and a picture-advertisement correspondence database, and sending to a background server feature codes, webpage URL information and picture URL information exclusive to the website when a webpage is loaded by a user browser; and a second step that the server process uses hash algorithm to judge whether the picture exists in the picture database according to the picture URL information, sending a corresponding advertisement to the browser if the picture exists in the picture database; and taking a different step to generate an advertisement for the picture if not. The generation method introduces static pictures into interactive content and makes full use of pictures already existed in various big websites, which makes social shopping based on rich media network advertisements and socialization be integrated with e-commerce with personalized recommendation, and makes advertisements extend from on-line stores to pictures of other websites.
Description
Technical field
The present invention relates to a kind of generation method of the active advertisement based on internet picture, belong to field of image recognition and computer software integration field.
Background technology
With respect to word, picture can provide more directly perceived and vivid content.At present on internet a large amount of picture mainly as the supporting explanation of word, and amateurish advertisement.And special display advertising is mainly used in showing the product of certain brand, location advertising is less, and function is comparatively single.
In a width picture, the resource that can be used for advertising display is a lot, and for example model's clothing, cap, hair fastener, sunglasses, earrings, necklace, brooch, ring, knapsack can be served as the content of advertisement.Numerous ad contents are integrated among a pictures, can greatly meet the consumption demand of different user.
Active advertisement refers to that user initiatively participates in, but not banner (banner), pop-up ad and the Interstitial Ad etc. that insert are forced in website, and active advertisement can greatly improve the conversion ratio of the web advertisement, and the user who improves advertisement experiences.
The object of the invention is to existing picture in each large website, forum and blog to become the interactively dynamic rich media with advertisement, user can browse wherein interested content, and the internet behavior of the generative process of advertisement on user and networking speed are without any impact.
Summary of the invention
Technical matters to be solved by this invention is that static images is introduced to interactive content for user's active operation, advertiser's advertisement is extended to from their on-line shop among the photo and video of other websites, without amendment procedure site and on-line shop's picture database content, provide a kind of generation method of the active advertisement based on internet picture simultaneously.
The present invention is for solving the problems of the technologies described above by the following technical solutions:
A generation method for active advertisement based on internet picture, comprises that step is as follows:
Steps A, active advertisement background data base is set in background server, described active advertisement background data base comprises picture database, advertising database, characteristics of image storehouse, display advertising corresponding relation storehouse, in described picture database, comprise the picture material of picture, in described advertising database, comprise various advertising messages; The specific features that described characteristics of image storehouse comprises for document image content, the corresponding relation of one-to-many between described display advertising corresponding relation storehouse recordable picture and advertisement;
Step B, the web page template of amendment website: in the condition code of exclusive this website of the body of web page template partial insertion, make the webpage of with good grounds this web page template generation of institute all comprise this condition code, this condition code records in background server;
Step C, in the time that the webpage of processing through step B is loaded by user browser, according to the information extraction webpage url information of browser address bar, extract picture url information by reading web page files source code, then the condition code of this website, webpage url information and picture url information are sent to background server;
Step D, server processes receive after the information that browser sends, and first judge whether the condition code of this website of recording in site information in webpage url information and background server mates, if do not mated, exit; If coupling, according to picture url information, use hash algorithm to judge whether this picture exists in picture database: as existed, according to the picture of display advertising corresponding relation storehouse record and the corresponding relation of advertisement, from advertising database, extract corresponding transmitting advertisement information to user browser, otherwise enter step e;
Step e, downloads picture file according to picture url information from Internet, and is saved in picture database;
Step F, carries out image recognition to the picture file content of obtaining based on GIST global characteristics and SIFT local feature, and sets up position sensing Hash, and this picture and other pictures that are kept at picture database are carried out to quick duplicate detection; If exist the image content of repetition, copy the image feature information of this picture, jump to step H; Otherwise, enter next step;
Step G, is transported to Image Classifier by picture file circulation, extracts all image feature informations of this picture; Then the image feature information extracting is saved to characteristics of image storehouse;
Step H, carries out full-text search according to image feature information to advertising database, completes mating of image feature information and advertising message, and matching result is saved to display advertising corresponding relation storehouse.
Further, the generation method of the active advertisement based on internet picture of the present invention, before step H carries out, also comprises according to current advertising theme and ad content, sets the step of advertising message scope in advertising database; In step H, adopt image feature information according to the advertising message scope of setting, advertising database is carried out to full-text search, complete coupling.
Further, the generation method of the active advertisement based on internet picture of the present invention, is also included in the step of recording user behavior in background server; After the advertising message matching with picture is loaded by user browser, record the number of times that this advertising message is loaded, clicked number of times.
Further, the generation method of the active advertisement based on internet picture of the present invention, the advertising message described in steps A in advertising database is provided by the website using background data base of commodity provider.
The technical solution used in the present invention has following effect:
The present invention introduces interactive content static images, and user only need to move to mouse in picture can present advertisement, and click advertisement just can browse to the particular content of advertisement.
The present invention makes Rich Media's web advertisement and society turn to ecommerce (especially B2C) weave in of basic social shopping and personalized recommendation, allows advertiser's advertisement extend among the picture of other websites from their on-line shop.
Brief description of the drawings
Fig. 1 is system flowchart of the present invention.
Fig. 2 is user's operation chart of the present invention.
Embodiment
Below in conjunction with accompanying drawing, technical scheme of the present invention is described in further detail:
As shown in Figure 1, overall flow of the present invention comprises that website preparation, user's operation, setting advertisement scope, advertisement pushing or advertisement generate four key links, and wherein advertisement generation is core link of the present invention.Advertisement generate comprised that picture obtains, image recognition, generation match advertisements, user behavior record, five key steps of manual examination and verification.Be elaborated respectively below in conjunction with accompanying drawing:
Prepare website:
Website master is without the program code of amendment website, only need to revise the web page template of website, at the code of the body of template partial insertion customization, in custom code, comprised website master's condition code (the main condition code in different websites is not identical), the webpage that generates of with good grounds template will automatically comprise custom code.In the time that webpage is loaded by user browser, custom code according to the information extraction webpage url of browser address bar, read web page files source code and extract picture url(undersized picture by disallowable), then website master's condition code, webpage url and picture url are sent to server.Server processes will be according to these Information Authentication website owner identifications, carry out picture searching or picture download, generating advertisement or advertisement.
User's operation:
User's operation and normal upper net operation are without any difference, and user only needs open any browser and inputs network address, and advertisement will be loaded on picture automatically.After advertisement is loaded, mouse is moved to picture by user, and many parts of advertisement meetings are appeared in one's mind automatically, clicks advertisement and can see and browse detailed content, and mouse is removed rear advertisement meeting and automatically disappeared.
Set advertisement scope:
Can set advertisement scope according to website master's requirement, if reduce the scope of coupling, the speed of generating advertisement can be accelerated, and number of ads can reduce, and the purpose of advertisement and specific aim can be stronger.
Advertisement pushing or advertisement generate:
Steps A, picture obtain:
Server processes receive the information that custom code sends, and complete following function: judge according to webpage url and website master's condition code whether webpage derives from legitimate site master; In picture library, search for matched record according to the url of picture, find after matched record, extract match advertisements by search pictures advertisement corresponding relation storehouse and advertisement base and push to user browser; If do not search matched record in picture library, according to the url of picture direct download pictures from Internet, for image recognition process, image content is identified, picture is saved to picture database simultaneously.
1. server processes judge whether site information in webpage url and website master's condition code mate, and are not to derive from legitimate site master if do not mate explanation webpage, and all following steps all can not be carried out.If web page source is in legitimate site master, according to picture url, use hash algorithm to judge whether picture exists in picture library, from advertisement base, extract corresponding transmitting advertisement information to user browser according to the corresponding relation of the picture recording in display advertising corresponding relation storehouse and advertisement as existed.Otherwise carry out step 2.;
2. download picture file according to picture url from Internet, and picture is saved in to picture library;
According to picture url, if can search matched record in picture library, once accessed mistake of this webpage is described, these pictures have all generated match advertisements, server processes, by search pictures advertisement corresponding relation storehouse and advertisement base, extract match advertisements and push to user browser.Illustrate that this webpage is accessed first and do not search matched record, corresponding picture all need be downloaded from Internet, then generates through the following steps match advertisements, and in this process, because match advertisements not yet generates, browser can carrying advertisement information.
Step B, image recognition:
The effect of image recognition process is to extract image feature information and is saved to characteristics of image storehouse.Because picture may be reprinted by different websites, so exist the identical but picture that url is different of a large amount of contents on Internet.First image recognition process judges in picture library, whether to exist same image, if had, duplicating image characteristic information is also saved to characteristics of image storehouse, otherwise image is delivered to Image Classifier, and the characteristic information that extracts image is saved to characteristics of image storehouse.
3. the picture file content of obtaining is carried out to image recognition, based on GIST global characteristics and SIFT local feature, and set up position sensing Hash, image and other images that are kept at picture library are carried out to quick duplicate detection.If exist the picture material of repetition, duplicating image characteristic information be saved to characteristics of image storehouse, jumps to
, otherwise carry out step
;
image is transported to Image Classifier, and the characteristic information that extracts image is saved to characteristics of image storehouse;
5. image is transported to other Image Classifiers, repeating step
, until all Image Classifiers are all disposed;
Sorter can be classified and extract specific features information image.For example, for the feature of face, to Image Parallel extract Hash-like feature, be transported to face classification device, if there is face to exist in image, the specific features (position, color etc.) of the various organs on face (face, nose, eyes etc.) will be extracted so, will produce the advertisements such as corresponding lipstick, face cream, sunglasses according to these information search advertisement base.The characteristic type that different sorters can be processed is different, and for example pedestrian's sorter need to extract HOG feature, and sex, age and scene classification device need to extract GIST feature.In piece image, may comprise polytype feature, after all sorters are disposed, in image, all specific features are extracted and are saved to characteristics of image storehouse.
Step C, generation match advertisements:
The quantity of the match advertisements generating is relevant with website master's requirement with content, different website masters, or the advertisement difference that same website master promoted mainly in different periods (realizing by " setting advertisement scope "), the match advertisements that same image content loads at different times may be different on number of ads and ad content.But in store complete characteristic information always in characteristics of image storehouse, for repeat later image generating advertisement time saved the time.
the image feature information and the predefined advertisement scope that obtain according to step B are carried out full-text search to advertisement base, complete the coupling of mass data, and matching result is saved to display advertising corresponding relation storehouse;
cannot generate the picture of match advertisements, submit manual examination and verification to;
Step D, user behavior record:
User's behavior is recorded and can be analyzed the effect that advertisement produces, and the user behavior of record comprises the loading number of times of advertisement, number of times that mouse moves to picture and clicks the number of times of advertisement.
after the viewed device of match advertisements loads, for all operations of picture, the number of times that comprises loading number of times, the number of times that mouse moves to picture of advertisement and click advertisement will be sent to user behavior record database;
the operation behavior of flow analysis process counting user;
when the operation of picture is exceeded to the threshold values of setting, submit manual examination and verification to;
Step e: manual examination and verification:
The object of manual examination and verification is to check that whether the corresponding ad content mating with picture is correct, and can revise.There are following two kinds of situations and will carry out manual examination and verification: the picture that cannot generate match advertisements; The operation of picture is exceeded to the threshold values of setting.
check picture library, characteristics of image storehouse, display advertising corresponding relation storehouse, advertisement base, the corresponding relation of amendment picture and advertisement is also saved to display advertising corresponding relation storehouse.
Detailed step by above flow process is described, and in the process that can find to generate in match advertisements, website master, without amendment procedure site, only need to revise webpage and generate template; The internet behavior of the generative process of advertisement on user and networking speed are without any impact.The advertisement generating is active advertisement, and whether user can select browse advertisements voluntarily, and the action of user's browse advertisements is very simple, only mouse need be moved to picture, and specific operation process as shown in Figure 2.
The present invention makes full use of existing picture in each large website, forum and blog, static images is introduced to interactive content, ecommerce (especially B2C) weave in that makes Rich Media's web advertisement and turn to basic social shopping and personalized recommendation with society, allows advertiser's advertisement extend among the picture of other websites from their on-line shop.
Claims (3)
1. a generation method for the active advertisement based on internet picture, is characterized in that, comprises that step is as follows:
Steps A, active advertisement background data base is set in background server, described active advertisement background data base comprises picture database, advertising database, characteristics of image storehouse, display advertising corresponding relation storehouse, in described picture database, comprise the picture material of picture, in described advertising database, comprise various advertising messages; The specific features that described characteristics of image storehouse comprises for document image content, the corresponding relation of one-to-many between described display advertising corresponding relation storehouse recordable picture and advertisement;
Step B, the web page template of amendment website: in the condition code of exclusive this website of the body of web page template partial insertion, make the webpage of with good grounds this web page template generation of institute all comprise this condition code, this condition code records in background server;
Step C, in the time that the webpage of processing through step B is loaded by user browser, according to the information extraction webpage url information of browser address bar, extract picture url information by reading web page files source code, then the condition code of this website, webpage url information and picture url information are sent to background server;
Step D, server processes receive after the information that browser sends, and first judge that whether site information in webpage url information and website master's condition code mate, if do not mated, exit; If coupling, according to picture url information, use hash algorithm to judge whether this picture exists in picture database: as existed, according to the picture of display advertising corresponding relation storehouse record and the corresponding relation of advertisement, from advertising database, extract corresponding transmitting advertisement information to user browser, otherwise enter step e;
Step e, downloads picture file according to picture url information from Internet, and is saved in picture database;
Step F, carries out image recognition to the picture file content of obtaining based on GIST global characteristics and SIFT local feature, and sets up position sensing Hash, and this picture and other pictures that are kept at picture database are carried out to quick duplicate detection; If exist the image content of repetition, copy the image feature information of this picture, jump to step H; Otherwise, enter next step;
Step G, is transported to Image Classifier by picture file circulation, extracts all image feature informations of this picture; Then the image feature information extracting is saved to characteristics of image storehouse; According to current advertising theme and ad content, in advertising database, set advertising message scope;
Step H, adopts image feature information according to the advertising message scope of setting, and advertising database is carried out to full-text search, completes mating of image feature information and advertising message, and matching result is saved to display advertising corresponding relation storehouse.
2. the generation method of the active advertisement based on internet picture according to claim 1, is characterized in that, is also included in the step of recording user behavior in background server; After the advertising message matching with picture is loaded by user browser, record the number of times that this advertising message is loaded, clicked number of times.
3. the generation method of the active advertisement based on internet picture according to claim 1, is characterized in that, the advertising message described in steps A in advertising database is provided by the website using background data base of commodity provider.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210047729.5A CN102662950B (en) | 2012-02-29 | 2012-02-29 | Active advertisement generation method based on Internet picture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201210047729.5A CN102662950B (en) | 2012-02-29 | 2012-02-29 | Active advertisement generation method based on Internet picture |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102662950A CN102662950A (en) | 2012-09-12 |
| CN102662950B true CN102662950B (en) | 2014-11-05 |
Family
ID=46772441
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201210047729.5A Active CN102662950B (en) | 2012-02-29 | 2012-02-29 | Active advertisement generation method based on Internet picture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102662950B (en) |
Families Citing this family (28)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103064874B (en) * | 2012-10-26 | 2016-08-03 | 北京奇虎科技有限公司 | Obtain the method for web page quality data and browser and server |
| CN103064873B (en) * | 2012-10-26 | 2017-03-08 | 北京奇虎科技有限公司 | A kind of web page quality data capture method and system |
| CN103856490A (en) * | 2012-11-28 | 2014-06-11 | 纽海信息技术（上海）有限公司 | Shopping website accessed by multiple network speed |
| CN103971244B (en) | 2013-01-30 | 2018-08-17 | 阿里巴巴集团控股有限公司 | A kind of publication of merchandise news and browsing method, apparatus and system |
| US9055097B1 (en) | 2013-03-15 | 2015-06-09 | Zerofox, Inc. | Social network scanning |
| US9674214B2 (en) | 2013-03-15 | 2017-06-06 | Zerofox, Inc. | Social network profile data removal |
| US9674212B2 (en) | 2013-03-15 | 2017-06-06 | Zerofox, Inc. | Social network data removal |
| US9027134B2 (en) | 2013-03-15 | 2015-05-05 | Zerofox, Inc. | Social threat scoring |
| US9191411B2 (en) * | 2013-03-15 | 2015-11-17 | Zerofox, Inc. | Protecting against suspect social entities |
| CN105874449B (en) * | 2013-11-08 | 2019-05-31 | 谷歌有限责任公司 | For extracting and generating the system and method for showing the image of content |
| US9544325B2 (en) | 2014-12-11 | 2017-01-10 | Zerofox, Inc. | Social network security monitoring |
| US10516567B2 (en) | 2015-07-10 | 2019-12-24 | Zerofox, Inc. | Identification of vulnerability to social phishing |
| CN105117463B (en) * | 2015-08-24 | 2019-08-06 | 北京旷视科技有限公司 | Information processing method and information processing unit |
| CN105095498A (en) * | 2015-08-24 | 2015-11-25 | 北京旷视科技有限公司 | Information processing method and device |
| CN105160561A (en) * | 2015-09-30 | 2015-12-16 | 北京中搜网络技术股份有限公司 | Situational picture advertising method and apparatus |
| CN109670127A (en) * | 2015-12-16 | 2019-04-23 | 北京奇虎科技有限公司 | Browser, the method and system for realizing pop-up reminding window |
| CN108021417B (en) * | 2016-11-01 | 2022-02-18 | 阿里巴巴集团控股有限公司 | Method and device for processing banner elements of pages |
| US11256812B2 (en) | 2017-01-31 | 2022-02-22 | Zerofox, Inc. | End user social network protection portal |
| US11394722B2 (en) | 2017-04-04 | 2022-07-19 | Zerofox, Inc. | Social media rule engine |
| US10868824B2 (en) | 2017-07-31 | 2020-12-15 | Zerofox, Inc. | Organizational social threat reporting |
| US11165801B2 (en) | 2017-08-15 | 2021-11-02 | Zerofox, Inc. | Social threat correlation |
| US11418527B2 (en) | 2017-08-22 | 2022-08-16 | ZeroFOX, Inc | Malicious social media account identification |
| US11403400B2 (en) | 2017-08-31 | 2022-08-02 | Zerofox, Inc. | Troll account detection |
| CN107909385B (en) * | 2017-10-16 | 2021-11-19 | 微梦创科网络科技（中国）有限公司 | Active advertisement creative realization method and system based on microblog user |
| US11134097B2 (en) | 2017-10-23 | 2021-09-28 | Zerofox, Inc. | Automated social account removal |
| CN109509013B (en) * | 2018-08-14 | 2023-08-15 | 广东长宝信息科技股份有限公司 | Method for pushing image information by mobile terminal equipment |
| CN109599039A (en) * | 2019-01-11 | 2019-04-09 | 张凯 | A kind of method and system playing e-advertising |
| CN112906696B (en) * | 2021-05-06 | 2021-08-13 | 北京惠朗时代科技有限公司 | English image region identification method and device |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1376996A (en) * | 2001-03-27 | 2002-10-30 | 国际商业机器公司 | Server and method for loading advertisement on web-pages, web-pages display device and method |
| US6654725B1 (en) * | 1998-11-09 | 2003-11-25 | Nec Corporation | System and method for providing customized advertising on the World Wide Web |
| CN101558421A (en) * | 2006-12-12 | 2009-10-14 | 雅虎公司 | Smart advertisement generating system |
-
2012
- 2012-02-29 CN CN201210047729.5A patent/CN102662950B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6654725B1 (en) * | 1998-11-09 | 2003-11-25 | Nec Corporation | System and method for providing customized advertising on the World Wide Web |
| CN1376996A (en) * | 2001-03-27 | 2002-10-30 | 国际商业机器公司 | Server and method for loading advertisement on web-pages, web-pages display device and method |
| CN101558421A (en) * | 2006-12-12 | 2009-10-14 | 雅虎公司 | Smart advertisement generating system |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102662950A (en) | 2012-09-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102662950B (en) | Active advertisement generation method based on Internet picture | |
| Zhang et al. | Object-level video advertising: an optimization framework | |
| KR102533972B1 (en) | Video Ingest Framework for Visual Search Platform | |
| US10769444B2 (en) | Object detection from visual search queries | |
| TWI631474B (en) | Method and device for product identification label and method for product navigation | |
| WO2019171128A1 (en) | In-media and with controls advertisement, ephemeral, actionable and multi page photo filters on photo, automated integration of external contents, automated feed scrolling, template based advertisement post and actions and reaction controls on recognized objects in photo or video | |
| CN108694210B (en) | Template generation method and device | |
| KR102102571B1 (en) | System and method for providing online shopping platform | |
| KR20180120571A (en) | System for social network service marketing using influencer | |
| CN102150178A (en) | System and method for using supplemental content items for search criteria for identifying other content items of interest | |
| US20200226651A1 (en) | Methods and systems for product discovery in user generated content | |
| US9449231B2 (en) | Computerized systems and methods for generating models for identifying thumbnail images to promote videos | |
| Ha et al. | Fashion conversation data on instagram | |
| US20230030560A1 (en) | Methods and systems for tagged image generation | |
| US20080091555A1 (en) | User generated style content | |
| CN110457596A (en) | A kind of resource recommendation processing method and processing device | |
| KR102102572B1 (en) | System and method for providing online shopping mall | |
| KR102062248B1 (en) | Method for advertising releated commercial image by analyzing online news article image | |
| KR20210041733A (en) | Method, apparatus and computer program for fashion item recommendation | |
| US11302048B2 (en) | Computerized system and method for automatically generating original memes for insertion into modified messages | |
| US20200226167A1 (en) | Methods and systems for dynamic content provisioning | |
| KR20210045652A (en) | An apparatus for providing virtual fashion experience contents service | |
| KR20210045663A (en) | A program for providing virtual fashion experience contents service | |
| KR20210045808A (en) | System for Interactive Content Service Platform based on Virtual Fashion Items | |
| KR20210045659A (en) | Program for virtual fashion item interface management |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |