WO2019076294A1 - 推广信息投放方法、装置、介质及设备 - Google Patents

推广信息投放方法、装置、介质及设备 Download PDF

Info

Publication number
WO2019076294A1
WO2019076294A1 PCT/CN2018/110407 CN2018110407W WO2019076294A1 WO 2019076294 A1 WO2019076294 A1 WO 2019076294A1 CN 2018110407 W CN2018110407 W CN 2018110407W WO 2019076294 A1 WO2019076294 A1 WO 2019076294A1
Authority
WO
WIPO (PCT)
Prior art keywords
page
display object
similarity
keyword
image
Prior art date
Application number
PCT/CN2018/110407
Other languages
English (en)
French (fr)
Inventor
向卓林
刘肖琛
冯喆
苏麒匀
林世飞
赵艳
黄磊
Original Assignee
腾讯科技(深圳)有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 腾讯科技(深圳)有限公司 filed Critical 腾讯科技(深圳)有限公司
Publication of WO2019076294A1 publication Critical patent/WO2019076294A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06QINFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00Commerce
    • G06Q30/02Marketing; Price estimation or determination; Fundraising
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F16/00Information retrieval; Database structures therefor; File system structures therefor
    • G06F16/90Details of database functions independent of the retrieved data types
    • G06F16/95Retrieval from the web
    • G06F16/953Querying, e.g. by the use of web search engines
    • G06F16/9535Search customisation based on user profiles and personalisation

Definitions

  • the present invention relates to the field of promotion information recommendation technology, and in particular, to a promotion information delivery method, device, medium and device.
  • the commonly used promotion information delivery method is: displaying the promotion information corresponding to the products of the same category that the user has recently viewed in the page selected by the user, or displaying the higher bid in the webpage selected by the user. Promote information.
  • the embodiment of the invention provides a promotion information delivery method, including:
  • the promotion information is delivered to the user.
  • the embodiment of the invention further provides a promotion information delivery device, comprising:
  • An extracting module configured to extract a page image identifier of the specified page in response to an operation of viewing a specified page
  • a determining module configured to search for display object information corresponding to the page image identifier of the specified page from the corresponding relationship between the set page image identifier and the display object information; wherein, the display object image in the display object information corresponding to the page image identifier is The page image corresponding to the page identifier is matched; according to the correspondence between the pre-stored display object information and the promotion information, the promotion information corresponding to the display object information corresponding to the page image identifier of the specified page is obtained from the promotion information database;
  • the delivery module is configured to perform promotion information on the user according to the obtained promotion information.
  • An embodiment of the present invention further provides a promotion information delivery apparatus, including a processor; a memory for storing processor executable instructions; wherein the processor is configured to:
  • the promotion information is delivered to the user.
  • the embodiment of the present invention further provides a non-volatile computer storage medium, where the computer storage medium stores an executable program, and the executable program is executed by the processor to implement the step of the promotion information delivery method.
  • the embodiment of the present invention further provides a computing device, including a memory, a processor, and a computer program stored in the memory, where the processor implements the step of the promotion information delivery method when the program is executed.
  • FIG. 1 is a schematic flowchart of a method for placing a promotion information according to an embodiment of the present invention
  • FIG. 1B is a schematic flowchart of a method for placing a promotion information according to an embodiment of the present invention
  • FIG. 2 is a schematic flowchart of a method for determining promotion information corresponding to a page picture identifier of a specified page according to an embodiment of the present disclosure
  • FIG. 3 is a schematic flowchart of a method for setting a correspondence between a page picture identifier and display object information according to an embodiment of the present disclosure
  • FIG. 4 is a schematic flowchart of a method for setting a correspondence between a page keyword and display object information according to an embodiment of the present invention
  • FIG. 5 is a schematic flowchart of a method for calculating a similarity between a page image and a display object image according to an embodiment of the present disclosure
  • FIG. 6 is a schematic flowchart of a method for determining a local invariant feature vector of a page picture according to an embodiment of the present disclosure
  • FIG. 7 is a schematic diagram of an application scenario according to an embodiment of the present invention.
  • FIG. 8 is a schematic structural diagram of a promotion information placing apparatus according to an embodiment of the present invention.
  • FIG. 9 is a schematic structural diagram of hardware of a computing device according to an embodiment of the present invention.
  • the inventor finds that the promotion information based on the commonly used promotion information delivery method is not related to the content of the page currently viewed by the user. At this time, the user ignores the displayed on the page viewed by the user. The probability of promoting information is relatively large. Therefore, the commonly used method of promoting information dissemination has a problem that the click rate of the promotion information is low.
  • an embodiment of the present invention provides a promotion information delivery method, as shown in FIG. 1A, including:
  • Step 101A Extracting a page picture identifier of the specified page in response to an operation of viewing the specified page.
  • Step 102A Searching for display object information corresponding to the page image identifier of the specified page from the corresponding relationship between the set page image identifier and the display object information; wherein, the display object image in the display object information corresponding to the page image identifier and the page The picture image corresponding to the picture identifier matches.
  • step 103A the promotion information corresponding to the display object information corresponding to the page picture identifier of the specified page is obtained from the promotion information base according to the correspondence between the display object information and the promotion information stored in advance.
  • Step 104A Perform the promotion information delivery to the user according to the obtained promotion information.
  • the embodiment of the present invention further provides a promotion information delivery method, as shown in FIG. 1B, including:
  • Step 101B In response to an operation of viewing a specified page, extracting a page picture identifier of the specified page.
  • the specified page may be a page in the page recommendation list, and the body content of the specified page includes at least one image, and the at least one image is used as the page image of the specified page, and each page image corresponds to a page image identifier.
  • the page picture identifier may be composed of a page identifier and a picture identifier.
  • a specified image is selected from the plurality of images as the page image of the specified page, and when the specified page includes only one image, the image is used as the image. Specifies the page image of the page.
  • Step 102B Determine promotion information corresponding to the page picture identifier of the specified page.
  • the correspondence between the page image identifier and the promotion information of the page may be preset, and the promotion information corresponding to the page image identifier of the specified page may be searched from the correspondence relationship, and the page image identifier corresponding to the specified page may also be determined according to other manners.
  • Promotion information not limited here. Among them, the promotion information may specifically be information such as advertisements that can achieve the purpose of publicity.
  • the page picture identifier of the page may correspond to only one promotion information, and may also correspond to multiple promotion information. That is, the promotion information corresponding to the determined page picture identifier of the specified page may be one or more.
  • Step 103B Perform promotion information on the user according to the determined promotion information.
  • the promotion information is provided for the user, and further, when the determined promotion information is multiple, the promotion information may be sorted according to the promotion information sorting algorithm, and the determined promotion information is sorted, and the sorting is selected. The top set of promotion information is placed on the user.
  • the content selected by the user to view the content is generally the content that is of interest to the user, and the promotion information corresponding to the page image identifier of the specified page is determined, and the determined promotion information matches the image in the specified page, so that the promotion information is delivered to the user.
  • the content of the page that the user currently selects to view is related. At this time, the promotion information that is delivered to the user is more likely to be the promotion information that the user is interested in, and the click rate of the promotion information displayed on the page may be improved to some extent.
  • the embodiment of the present invention is applicable to a scenario in which a user can provide a viewable page including a picture.
  • the embodiment of the present invention can be applied to a news APP (application), where the specified page is selected by the user. Viewing the news details page, the page recommendation list is the recommended news displayed in the news recommendation interface in the news app.
  • a news APP application
  • the page recommendation list is the recommended news displayed in the news recommendation interface in the news app.
  • the promotion information corresponding to the page picture identifier of the specified page is determined:
  • Step 201 Searching for display object information corresponding to the page image identifier of the specified page from the corresponding relationship between the set page image identifier and the display object information; wherein, the display object image in the display object information corresponding to the page image identifier and the page The picture image corresponding to the picture identifier matches.
  • the corresponding relationship between the page image identifier and the display object information in the page with the page image in the page recommendation list is set and stored in advance, wherein the display object information may include the display object image, and the display object corresponding to the page image identifier.
  • the display object image in the information matches the page image corresponding to the page image identifier of the page.
  • the display object may be a commodity or a non-commodity, and is not limited herein. When the display object is a commodity, the display object information is product information, and when the display object is a non-product, the display object information is non-product information.
  • Step 202 If the display object information corresponding to the page image identifier of the specified page is found, the page image identifier corresponding to the specified page is obtained from the promotion information database according to the correspondence between the pre-stored display object information and the promotion information.
  • the promotion information corresponding to the display object information is obtained from the promotion information database according to the correspondence between the pre-stored display object information and the promotion information.
  • the display object image in the display object information corresponding to the page image identifier of the specified page matches the page image corresponding to the page image identifier of the specified page.
  • the promotion information corresponding to the display object information corresponding to the page image identifier is obtained from the promotion information database, wherein the promotion information corresponding to the display object information is based on the display Promotion information generated by object information.
  • Step 203 Determine the obtained promotion information as the promotion information corresponding to the page picture identifier of the specified page.
  • the content of the page that the user selects to view is generally the content that is of interest to the user.
  • the correspondence between the display object information and the page image identifier is set in advance, and the corresponding image is found according to the page image identifier in the specified page viewed by the user. Displaying the object information, and obtaining the promotion information corresponding to the found display object information from the promotion information database, and the found promotion information matches the picture in the specified page, so that the promotion information delivered to the user and the page currently selected by the user are viewed.
  • the content of the content is related. At this time, the promotion information delivered to the user is more likely to be the promotion information that the user is interested in, and the click rate of the promotion information displayed on the page may be improved to some extent.
  • step 203 may be performed in the following manner, the determined promotion information is inserted at a specified location of the specified page; and the specified page after the promotion information is inserted is returned to the user.
  • the designated location may be a lower area of the page image corresponding to the display object information corresponding to the promotion information, and the designated location may also be other areas in the specified page, where not Make a limit.
  • the promotion information delivery method provided by the embodiment of the present invention further includes: extracting a page keyword of the specified page.
  • the page keyword is a keyword extracted according to the text content corresponding to the specified page.
  • the promotion information delivery method provided by the embodiment of the present invention further includes:
  • the display object information corresponding to the page image identifier of the specified page is searched for from the corresponding relationship between the set page keyword and the display object information, where The display object keyword in the display object information corresponding to the page keyword matches the page keyword; if the display object information corresponding to the page keyword of the specified page is found, according to the pre-stored display object information and the promotion information Corresponding relationship, obtaining the promotion information corresponding to the display object information corresponding to the page keyword of the specified page from the promotion information base; and performing promotion information on the user according to the obtained promotion information.
  • the corresponding relationship between the page keyword of each page and the display object information in the page recommendation list is set and stored in advance, wherein the display object information may further include a display object keyword, and the display object information corresponding to the page keyword
  • the display object keyword matches the page keyword, and the display object keyword can be extracted from the display object title.
  • the display object keyword may also be a person name, an item name, and the like in the display object information.
  • the specific implementation process of the promotion information corresponding to the display object information corresponding to the page keyword of the specified page is obtained from the promotion information database, and according to the acquisition
  • the specific implementation process of the promotion information of the user refer to the above explanation, which will not be described here.
  • the content of the page that the user selects to view is generally the content that is of interest to the user.
  • the mapping relationship between the display object information and the page keyword to which the display object keyword matches the page keyword is set in advance, and is viewed according to the user.
  • the page keyword in the specified page finds the corresponding display object information, and obtains the promotion information corresponding to the found display object information from the promotion information database, and the found promotion information matches the keyword in the specified page.
  • the promotion information that is delivered to the user is related to the content of the page that the user currently selects to view, and the promotion information that is delivered to the user is more likely to be the promotion information that the user is interested in, and the promotion information displayed on the page may be improved to some extent. Clickthrough rate.
  • the correspondence between the page picture identifier and the display object information is set in the manner provided by FIG. 3:
  • Step 301 For each page, calculate the similarity between the page image of the page and the display object image in each of the display object information stored in advance.
  • the similarity between the page image in the page and the display object image in each display object information is determined.
  • the similarity calculation method between the page image and the display object image can be calculated according to the existing calculation method or the calculation method mentioned in the following.
  • Step 302 Display a display object picture whose similarity with the page picture of the page exceeds the first similarity threshold as a display object picture that matches the page picture of the page.
  • the display object image whose similarity with the page image exceeds the first similarity threshold is counted, and the statistical display object image is used as the display object image that matches the page image.
  • the first similarity threshold may be set according to an actual application scenario, which is not limited herein.
  • the similarity between the page image and the display object image exceeds the first similarity threshold, it may be stated that the page image includes content similar to the display object in the display object image.
  • Step 303 Determine display object information to which the display object picture that matches the page picture of the page belongs.
  • Step 304 The correspondence between the page image identifier corresponding to the page image of the page and the determined display object information is saved.
  • the correspondence between the page picture identifier and the display object information of each page can be obtained.
  • the correspondence between the page keyword and the display object information is set in the manner provided by FIG. 4:
  • Step 401 For each page, calculate the similarity between the page keyword of the page and the display object keyword in each of the display object information stored in advance.
  • the similarity between the page keyword in the page and the display object keyword in each display object information is determined.
  • the calculation method of the similarity between the page keyword and the display object keyword can be calculated according to the existing calculation method, which is not limited herein.
  • Step 402 Display object keywords whose similarity with the page keyword of the page exceeds the second similarity threshold as a display object keyword that matches the page keyword of the page.
  • the display object keyword whose similarity with the page keyword exceeds the second similarity threshold is counted, and the counted display object keyword is used as the display object keyword that matches the page keyword.
  • the second similarity threshold may be set according to an actual application scenario, which is not limited herein.
  • Step 403 Determine display object information to which the display object keyword that matches the page keyword of the page belongs.
  • Step 404 Save the correspondence between the page keyword of the page and the determined display object information.
  • the correspondence between the page keyword of each page and the display object information can be obtained.
  • the similarity between the page image and the display object image is calculated in the manner provided by FIG. 5:
  • Step 501 Determine a cosine similarity between the texture feature vector of the page image and the texture feature vector of the display object image as the first similarity.
  • Step 502 Determine a cosine similarity between the color feature vector of the page image and the color feature vector of the display object image as the second similarity.
  • Step 503 Determine a cosine similarity between the local invariant feature vector of the page image and the local invariant feature vector of the display object image as the third similarity.
  • Step 504 Calculate the similarity between the page picture and the display object picture according to the first similarity, the second similarity, and the third similarity.
  • steps 501, 502, and 503 are not limited.
  • a maximum value of the first similarity, the second similarity, and the third similarity may be selected as the similarity between the page image and the display object image, or from the first similarity, the second similarity, and the third Two values are randomly selected from the similarity, and the selected two values are averaged as the similarity between the page image and the display object image, or the mean values of the first similarity, the second similarity, and the third similarity are calculated.
  • the mean value is used as the similarity between the page image and the display object image, and is not limited herein. The similarity of any two pictures can be obtained by using the embodiment of the present invention.
  • the weighted summation results of the first similarity, the second similarity, and the third similarity are calculated; the obtained weighted summation result is used as the similarity between the page image and the display object image.
  • different weights are set for the texture feature, the color feature, and the local invariant feature, and the weighted summation results of the first similarity, the second similarity, and the third similarity are calculated as the page image and the display object image. Similarity.
  • the weight corresponding to the texture feature is a
  • the weight corresponding to the color feature is b
  • the weight corresponding to the local invariant feature is c
  • the first similarity is X
  • the second similarity is Y
  • the third similarity is The degree
  • the above weighted summation result H aX + bY + cZ.
  • the local invariant feature vector of the page image may be determined according to the content provided in FIG. 6:
  • Step 601 Extract a visual vocabulary vector from the page image and the stored display object images by using a sift (Scale-invariant feature transform) algorithm, and store the visual vocabulary vector.
  • a sift Scale-invariant feature transform
  • Step 602 The K-Means (ie, K-means) clustering algorithm is used to divide the visual vocabulary vector in the visual vocabulary set into N clusters.
  • Step 603 determining a central visual vocabulary vector of each of the obtained N clusters.
  • Step 604 Determine, for each visual vocabulary vector in the page picture, a central visual vocabulary vector with the smallest distance between the N central visual vocabulary vectors and the visual vocabulary vector, and determine the cluster to which the central visual vocabulary vector belongs. Determines the cluster to which the vocabulary vector belongs in the page image.
  • Step 605 Count the number of visual vocabulary vectors belonging to each cluster in the page picture.
  • Step 606 the statistically obtained numbers are combined into an N-dimensional vector as a local invariant feature vector of the page image.
  • the obtained 3-dimensional vector is [3, 6, 9], where 3 is the number of visual vocabulary vectors belonging to the first cluster in the page picture. 6 is the number of visual vocabulary vectors belonging to the second cluster in the page picture, and 9 is the number of visual vocabulary vectors belonging to the third cluster in the page picture.
  • the process of determining the local invariant feature vector of the display object picture is similar to the process of determining the local invariant feature vector of the page picture (ie, replacing the page picture in step 604 and step 605 with the display object picture), here Without further elaboration, the obtained locally invariant feature vector of the display object image is also an N-dimensional vector.
  • the color feature vector of any picture may be determined as follows:
  • each primary color can take 256 values, and divide 0 to 255 into four partitions: 0 to 63 are the 0th partition, 64 to 127 are the first partition, 128 to 191 are the second partition, and 192 to 255 are the third partition.
  • R (red) G (green) B (blue) three primary colors respectively correspond to four partitions, and 64 types of partition combinations are obtained, and for each pixel in the picture, the R value, the G value, and the B value of the pixel are determined.
  • the R value, the G value, and the partition to which the B value belongs are determined respectively; the R value, the G value, and the identifier of the partition to which the B value belongs are used as the partition combination corresponding to the pixel; the statistics belong to each partition combination.
  • the number of pixels; the number of pixels belonging to each partition combination in the picture constitutes a 64-dimensional vector as the color feature vector of the picture. Among them, in the partition combination, the first number represents the partition to which R belongs, the second number represents the partition to which G belongs, and the third number represents the partition to which B belongs.
  • the group combination to which the pixel belongs is 001.
  • the texture feature vector of any picture may be determined as follows:
  • the edge pixel of the value is recorded as 1, and the value of the pixel other than the central pixel in the neighborhood is counted to obtain a Q-1 binary string corresponding to the pixel in the image; the pixel corresponding to the pixel in the picture
  • the Q-1 bit binary string is converted to a decimal value, and the decimal value corresponding to the pixel in the picture is obtained; according to the preset histogram statistical method, the histogram value corresponding to each pixel point in the picture is subjected to histogram statistics to determine the histogram.
  • the number of pixels in the picture included in each decimal value range corresponding to the figure; the number of pixels in the picture included in each decimal value range corresponding to the histogram is combined into a vector, as the figure Texture feature vector, wherein the texture feature vector dimension equal to the number of decimal numerical value range.
  • the value of Q is 3, and may be other values, which are not limited herein.
  • the promotion information delivery method provided by the embodiment of the present invention is described in detail below in conjunction with the actual application scenario.
  • a schematic diagram of an application scenario of a promotion information delivery method includes a user terminal 701 and a server 702, wherein a user application 701 is installed with a news application, and the user 70 can operate the user terminal.
  • the server 702 is installed with a promotion information delivery system, and the promotion information delivery system includes an online module and an offline module.
  • the user terminal 701 and the server 702 can communicate through a communication network such as a local area network, a wide area network, or a mobile Internet;
  • the user terminal device 701 can be a portable device (for example, a mobile phone, a tablet, a notebook computer, etc.) or a personal computer (PC, Personal Computer), the news application can be used for browsing news, such as Tencent News, Daily Express.
  • a communication network such as a local area network, a wide area network, or a mobile Internet
  • the user terminal device 701 can be a portable device (for example, a mobile phone, a tablet, a notebook computer, etc.) or a personal computer (PC, Personal Computer), the news application can be used for browsing news, such as Tencent News, Daily Express.
  • PC Personal Computer
  • the advertiser when the promotion information is an advertisement and the display object is an item, the advertiser sends the product picture, the product title, and the item price attribute of the pre-promotional item to the offline module; the offline module sends the item according to the advertiser.
  • the product attribute generates the advertisement data in accordance with the advertisement format, and sends the advertisement data to the advertisement library of the online module for saving, wherein the product attribute includes the product image, the product title, and the product price; the offline module extracts the product for each product attribute.
  • the product information in the attribute wherein the product information includes the product image and the product keyword; the offline module extracts the page image identifier of the page and the page keyword for each page pre-recommended on the current day, and determines according to the embodiment of the present invention.
  • the corresponding relationship between the set page image identifier and the product information and the correspondence between the page keyword and the product information are sent to the corresponding relation database in the online module for storage.
  • the online module extracts the page picture identifier and the page keyword of the specified page in response to the operation of the user 70 to view the specified page sent by the news application in the user terminal 701; the correspondence between the determined page picture identifier and the product information Searching for the product information corresponding to the page image identifier of the specified page; wherein the product image in the product information corresponding to the page image identifier matches the page image corresponding to the page image identifier; if the page image of the specified page is found Identifying the corresponding product information, obtaining an advertisement corresponding to the product information corresponding to the page image identifier of the specified page from the advertisement library according to the corresponding relationship between the pre-stored product information and the advertisement; and performing the user on the obtained advertisement according to the obtained advertisement If the product information corresponding to the page image identifier of the specified page is not found, the product information corresponding to the page keyword of the specified page is searched for from the corresponding relationship between the determined page keyword and the product information, where The exhibition in the display object information corresponding to the page
  • the embodiment of the invention further provides a promotion information delivery device, as shown in FIG. 8, comprising:
  • the extracting module 801 is configured to extract a page picture identifier of the specified page in response to an operation of viewing the specified page;
  • a determining module 802 configured to determine promotion information corresponding to the page picture identifier of the specified page
  • the delivery module 803 is configured to perform promotion information on the user according to the determined promotion information.
  • the determining module 802 is configured to:
  • the display object corresponding to the page image identifier of the specified page is obtained from the promotion information database according to the correspondence between the pre-stored display object information and the promotion information. Promotion information corresponding to the information;
  • the obtained promotion information is determined as the promotion information corresponding to the page picture identifier of the specified page.
  • the promotion information delivery device provided by the embodiment of the present invention further includes: a search module 804, an acquisition module 805, where:
  • the extracting module 801 is further configured to extract a page keyword of the specified page
  • the searching module 804 is configured to: if the determining module 802 does not find the display object information corresponding to the page image identifier of the specified page, search for the specified information from the corresponding relationship between the set page keyword and the display object information.
  • the obtaining module 805 is further configured to: if the searching module 804 finds the display object information corresponding to the page keyword of the specified page, according to the correspondence between the pre-stored display object information and the promotion information, the obtaining information library And obtaining, in the obtaining, the promotion information corresponding to the display object information corresponding to the page keyword of the specified page;
  • the delivery module 803 is further configured to perform promotion information delivery on the user according to the promotion information acquired by the obtaining module 805.
  • the promotion information delivery device provided by the embodiment of the present invention further includes:
  • the first setting module 806 is configured to set a correspondence between the page picture identifier and the display object information in the following manner:
  • the promotion information delivery device provided by the embodiment of the present invention further includes:
  • the second setting module 807 is configured to determine a correspondence between the page keyword and the display object information in the following manner:
  • the similarity between the page image and the display object image is calculated as follows:
  • the similarity between the page picture and the display object picture is calculated according to the first similarity, the second similarity, and the third similarity.
  • the computing the similarity between the page image and the display object image according to the first similarity, the second similarity, and the third similarity including:
  • the obtained weighted summation result is used as the similarity between the page image and the display object image.
  • the embodiment of the present invention provides a non-volatile computer storage medium, where the computer storage medium stores an executable program, and the executable program is executed by the processor to implement the step of any promotion information delivery method provided by the foregoing embodiment.
  • the embodiment of the present invention further provides a computing device, including a memory, a processor, and a computer program stored in the memory.
  • a computing device including a memory, a processor, and a computer program stored in the memory.
  • the processor executes the program, the steps of the promotion information delivery method in the foregoing embodiments are implemented.
  • the embodiment of the present invention further provides a computing device for performing the promotion information delivery method in the foregoing embodiment.
  • FIG. 9 it is a schematic diagram of a hardware structure of the computing device according to the implementation of the present invention.
  • the computing device may include a memory 901, a processor 902, and a computer program stored on the memory, and when the processor executes the program, the steps of any of the above-mentioned embodiments are implemented.
  • the memory 901 may include a read only memory (ROM) and a random access memory (RAM), and provides the processor 902 with program instructions and data stored in the memory 901.
  • the computing device described in the embodiment of the present invention may further include an input device 903, an output device 904, and the like.
  • the input device 903 may include a keyboard, a mouse, a touch screen, etc.;
  • the output device 904 may include a display device such as a liquid crystal display (LCD), a cathode ray tube (CRT), a touch screen, or the like.
  • the memory 901, the processor 902, the input device 903, and the output device 904 may be connected by a bus or other means, as exemplified by a bus connection in FIG.
  • the processor 902 calls the program instructions stored in the memory 901 and executes the promotion information delivery method provided in the above embodiment according to the obtained program instructions.
  • the method, device, medium and device for promoting promotion information provided by the embodiments of the present invention have the following beneficial effects:
  • the content selected by the user to view the content is generally the content that is of interest to the user, and the promotion information corresponding to the page image identifier of the specified page is determined, and the determined promotion information matches the image in the specified page, so that the promotion information is delivered to the user.
  • the content of the page that the user currently selects to view is related. At this time, the promotion information that is delivered to the user is more likely to be the promotion information that the user is interested in, and the click rate of the promotion information displayed on the page may be improved to some extent.
  • modules of the promotional information delivery device are mentioned in the above detailed description, such division is merely exemplary and not mandatory. Indeed, in accordance with embodiments of the present invention, the features and functions of the two or more modules described above may be embodied in one module. Conversely, the features and functions of one of the modules described above may be further divided into multiple modules.
  • the computer program instructions can also be stored in a computer readable memory that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer readable memory produce an article of manufacture comprising the instruction device.
  • the apparatus implements the functions specified in one or more blocks of a flow or a flow and/or block diagram of the flowchart.
  • These computer program instructions can also be loaded onto a computer or other programmable data processing device such that a series of operational steps are performed on a computer or other programmable device to produce computer-implemented processing for execution on a computer or other programmable device.
  • the instructions provide steps for implementing the functions specified in one or more of the flow or in a block or blocks of a flow diagram.

Landscapes

  • Engineering & Computer Science (AREA)
  • Business, Economics & Management (AREA)
  • Databases & Information Systems (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Accounting & Taxation (AREA)
  • Development Economics (AREA)
  • Finance (AREA)
  • Strategic Management (AREA)
  • General Physics & Mathematics (AREA)
  • Entrepreneurship & Innovation (AREA)
  • General Business, Economics & Management (AREA)
  • Marketing (AREA)
  • Economics (AREA)
  • Game Theory and Decision Science (AREA)
  • Data Mining & Analysis (AREA)
  • General Engineering & Computer Science (AREA)
  • Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

本发明实施例提供一种推广信息投放方法、装置、介质及设备,该方法包括:响应于查看指定页面的操作,提取所述指定页面的页面图片标识;从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配;根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息;根据获取的推广信息,对用户进行推广信息投放。

Description

推广信息投放方法、装置、介质及设备
本申请要求于2017年10月19日提交中国专利局、申请号为201710977151.6、发明名称为“一种推广信息投放方法、装置、介质及设备”的中国专利申请的优先权,其全部内容通过引用结合在本申请中。
技术领域
本申请涉及推广信息推荐技术领域,尤其涉及一种推广信息投放方法、装置、介质及设备。
背景
目前,常用的推广信息投放方法为:在用户选择查看的页面中展示与用户最近浏览过的商品属于同一类的商品所对应的推广信息,或者,在用户选择查看的网页中展示出价较高的推广信息。
技术内容
本发明实施例提供了一种推广信息投放方法,包括:
响应于查看指定页面的操作,提取所述指定页面的页面图片标识;
从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配;
根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息;
根据获取的推广信息,对用户进行所述推广信息投放。
本发明实施例还提供了一种推广信息投放装置,包括:
提取模块,用于响应于查看指定页面的操作,提取所述指定页面的页面图片标识;
确定模块,用于从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配;根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息;
投放模块,用于根据获取的推广信息,对用户进行推广信息投放。
本发明实施例还提供了一种推广信息投放装置,包括处理器;用于存储处理器可执行指令的存储器;其中,所述处理器被配置为:
响应于查看指定页面的操作,提取所述指定页面的页面图片标识;
从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配;
根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息;
根据获取的推广信息,对用户进行所述推广信息投放。
本发明实施例还提供了一种非易失性计算机存储介质,所述计算机存储介质存储有可执行程序,该可执行程序被处理器执行实现上述推广信息投放方法的步骤。
本发明实施例还提供了一种计算设备,包括存储器、处理器及存储 在存储器上的计算机程序,所述处理器执行所述程序时实现上述推广信息投放方法的步骤。
附图简要说明
通过参考附图阅读下文的详细描述,本发明示例性实施例的上述以及其他目的、特征和优点将变得易于理解。在附图中,以示例性而非限制性的方式示出了本发明的若干实施方式,其中:
图1A为本发明实施例提供的推广信息投放方法的流程示意图;
图1B为本发明实施例提供的推广信息投放方法的流程示意图;
图2为本发明实施例提供的确定指定页面的页面图片标识对应的推广信息的方法流程示意图;
图3为本发明实施例提供的设置页面图片标识与展示对象信息的对应关系的方法流程示意图;
图4为本发明实施例提供的设置页面关键字与展示对象信息的对应关系的方法流程示意图;
图5为本发明实施例提供的计算页面图片与展示对象图片的相似度的方法流程示意图;
图6为本发明实施例提供的确定页面图片的局部不变特征向量的方法流程示意图;
图7为本发明实施例的一应用场景示意图;
图8为本发明实施例提供的推广信息投放装置的结构示意图;
图9为本发明实施例提供的计算设备的硬件结构示意图。
实施方式
为使本发明的目的、技术方案和优点更加清楚,下面将结合附图对 本发明可能的实施方式作进一步描述。
发明人在实现本发明的过程中发现,基于常用的这种推广信息投放方法所投放的推广信息,与用户当前查看的页面的内容并不相关,此时,用户忽略其查看的页面中所展示的推广信息的概率较大,因此,常用的这种推广信息投放方法存在推广信息点击率低的问题。
基于此,本发明实施例提供一种推广信息投放方法,如图1A所示,包括:
步骤101A,响应于查看指定页面的操作,提取所述指定页面的页面图片标识。
步骤102A,从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配。
步骤103A,根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息。
步骤104A,根据获取的推广信息,对用户进行所述推广信息投放。
上述各个步骤的具体实现可以参见下述实例,此处不作赘述。
在一些实例中,本发明实施例还提供一种推广信息投放方法,如图1B所示,包括:
步骤101B,响应于查看指定页面的操作,提取所述指定页面的页面图片标识。
具体实施时,指定页面可以为页面推荐列表中的页面,指定页面的正文内容包括至少一幅图片,将该至少一幅图片作为该指定页面的页面图片,每幅页面图片对应一个页面图片标识,其中,页面图片标识可由 页面标识和图片标识组成。
在一些实例中,当指定页面包括多幅图片时,从该多幅图片中选取一幅指定图片作为该指定页面的页面图片,当该指定页面仅包括一幅图片时,将该幅图片作为该指定页面的页面图片。
步骤102B,确定所述指定页面的页面图片标识对应的推广信息。
具体实施时,可以预先设置页面的页面图片标识与推广信息的对应关系,从该对应关系中查找指定页面的页面图片标识对应的推广信息,也可按照其它方式确定指定页面的页面图片标识对应的推广信息,这里不做限定。其中,推广信息具体可以为广告等能够达到宣传目的的信息。
需要说明的是,页面的页面图片标识可能仅对应一个推广信息,也可能对应多个推广信息。即,确定的指定页面的页面图片标识对应的推广信息可能为一个也可能为多个。
步骤103B,根据确定的推广信息,对用户进行推广信息投放。
具体实施时,确定推广信息后,针对用户进行推广信息投放,进一步地,在确定出的推广信息为多个的情况下,可按照推广信息排序算法,将确定的推广信息进行排序,从中选择排序靠前的设定数目个推广信息,投放给所述用户。
用户选择查看的页面中的内容一般为用户感兴趣的内容,确定指定页面的页面图片标识对应的推广信息,确定出的推广信息与指定页面中的图片相匹配,使得向用户投放的推广信息与用户当前选择查看的页面的内容相关,此时,向用户所投放的推广信息为用户感兴趣的推广信息的可能性较大,可以在一定程度上提高页面中所展示的推广信息的点击率。
需要说明的是,本发明实施例可应用于向用户提供可查看的包括图片的页面的场景中,比如,本发明实施例可应用于新闻APP(应用程序) 中,其中,指定页面为用户选择查看的新闻详情页,页面推荐列表为新闻APP中的新闻推荐界面中展示的推荐新闻。
在一些实例中,按照图2提供的内容,确定所述指定页面的页面图片标识对应的推广信息:
步骤201,从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配。
具体实施时,预先设置并存储页面推荐列表中,带有页面图片的页面中的页面图片标识与展示对象信息的对应关系,其中,展示对象信息可以包括展示对象图片,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配。其中,展示对象可以为商品或者非商品,这里不做限定,当展示对象为商品时,展示对象信息为商品信息,当展示对象为非商品时,展示对象信息为非商品信息。
步骤202,若查找到所述指定页面的页面图片标识对应的展示对象信息,则根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息。
具体实施时,指定页面的页面图片标识对应的展示对象信息中的展示对象图片与该指定页面的页面图片标识对应的页面图片相匹配。此时,根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取页面图片标识对应的展示对象信息所对应的推广信息,其中,展示对象信息对应的推广信息为根据该展示对象信息生成的推广信息。
步骤203,将获取的推广信息确定为所述指定页面的页面图片标识对应的推广信息。
用户选择查看的页面中的内容一般为用户感兴趣的内容,本发明实施例,预先设置展示对象信息与页面图片标识的对应关系,根据用户查看的指定页面中的页面图片标识,查找出相应的展示对象信息,并从推广信息库获取与查找出的展示对象信息对应的推广信息,查找出的推广信息与指定页面中的图片相匹配,使得向用户投放的推广信息与用户当前选择查看的页面的内容相关,此时,向用户所投放的推广信息为用户感兴趣的推广信息的可能性较大,可以在一定程度上提高页面中所展示的推广信息的点击率。
在一些实例中,可按照以下方式执行步骤203,将确定的推广信息插入指定页面的指定位置处;将插入推广信息后的指定页面返回给所述用户。进一步地,在图2提供的实施例的基础上,指定位置可以为与该推广信息对应的展示对象信息所对应的页面图片的下方区域,指定位置也可以为指定页面中的其它区域,这里不做限定。
在一些实例中,本发明实施例提供的推广信息投放方法,还包括:提取所述指定页面的页面关键字。其中,页面关键字为根据指定页面对应的文本内容提取的关键字。在此基础上,本发明实施例提供的推广信息投放方法进一步包括:
若未查找到所述指定页面的页面图片标识对应的展示对象信息,则从设置的页面关键字与展示对象信息的对应关系中查找所述指定页面的页面关键字对应的展示对象信息,其中,页面关键字对应的展示对象信息中的展示对象关键字与该页面关键字相匹配;若查找到所述指定页面的页面关键字对应的展示对象信息,则根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面关键字对应的展示对象信息对应的推广信息;根据获取的推广信息,对所述用户进行推广信息投放。
具体实施时,预先设置并存储页面推荐列表中,各个页面的页面关键字与展示对象信息的对应关系,其中,展示对象信息还可以包括展示对象关键字,页面关键字对应的展示对象信息中的展示对象关键字与该页面关键字相匹配,展示对象关键字可从展示对象标题中提取。在一些实例中,所述展示对象关键字还可以为展示对象信息中的人名、物品名等。本发明实施例中,根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面关键字对应的展示对象信息对应的推广信息的具体实施过程以及根据获取的推广信息,对所述用户进行推广信息投放的具体实施过程可参考上文的解释说明,这里不做赘述。
用户选择查看的页面中的内容一般为用户感兴趣的内容,本发明实施例,预先设置与页面关键字相匹配的展示对象关键字所属的展示对象信息与页面关键字的对应关系,根据用户查看的指定页面中的页面关键字,查找出相应的展示对象信息,并从推广信息库获取与查找出的展示对象信息对应的推广信息,查找出的推广信息与指定页面中的关键字相匹配,使得向用户投放的推广信息与用户当前选择查看的页面的内容相关,向用户投放的推广信息为用户感兴趣的推广信息的可能性较大,可以在一定程度上提高页面中所展示的推广信息的点击率。
在一些实例中,按照图3提供的方式设置页面图片标识与展示对象信息的对应关系:
步骤301,针对每个页面,分别计算该页面的页面图片与预先存储的各个展示对象信息中的展示对象图片的相似度。
具体实施时,针对页面推荐列表中的每个页面,确定该页面中的页面图片与各个展示对象信息中的展示对象图片的相似度。页面图片与展示对象图片的相似度计算方式可以按照现有的计算方式,也可以按照下 文中提到的计算方式。
步骤302,将与该页面的页面图片的相似度超过第一相似度阈值的展示对象图片,作为与该页面的页面图片相匹配的展示对象图片。
具体实施时,统计与页面图片的相似度超过第一相似度阈值的展示对象图片,并将统计出的展示对象图片作为与该页面图片相匹配的展示对象图片。其中,第一相似度阈值可根据实际应用场景进行设定,这里不做限定。
需要说明的是,若页面图片与展示对象图片的相似度超过第一相似度阈值,则可说明该页面图片中包含与该展示对象图片中的展示对象相似的内容。
步骤303,确定与该页面的页面图片相匹配的展示对象图片所属的展示对象信息。
步骤304,保存该页面的页面图片对应的页面图片标识与确定的展示对象信息的对应关系。
具体实施时,提取与页面的页面图片相匹配的展示对象图片所属的展示对象信息,并将提取的展示对象信息标记为第一类展示对象信息;保存该页面的页面图片标识与第一类展示对象信息中的展示对象信息的对应关系。
利用本发明实施例,可得到各个页面的页面图片标识与展示对象信息的对应关系。
在一些实例中,按照图4提供的方式设置页面关键字与展示对象信息的对应关系:
步骤401,针对每个页面,分别计算该页面的页面关键字与预先存储的各个展示对象信息中的展示对象关键字的相似度。
具体实施时,针对页面推荐列表中的每个页面,确定该页面中的页 面关键字与各个展示对象信息中的展示对象关键字的相似度。页面关键字与展示对象关键字的相似度计算方式可以按照现有的计算方式,这里不做限定。
步骤402,将与该页面的页面关键字的相似度超过第二相似度阈值的展示对象关键字,作为与该页面的页面关键字相匹配的展示对象关键字。
具体实施时,统计与页面关键字的相似度超过第二相似度阈值的展示对象关键字,并将统计出的展示对象关键字作为与该页面关键字相匹配的展示对象关键字。其中,第二相似度阈值可根据实际应用场景进行设定,这里不做限定。
步骤403,确定与该页面的页面关键字相匹配的展示对象关键字所属的展示对象信息。
步骤404,保存该页面的页面关键字与确定的展示对象信息的对应关系。
具体实施时,提取与页面的页面关键字相匹配的展示对象关键字所属的展示对象信息,并将提取的展示对象信息标记为第二类展示对象信息;保存该页面的页面关键字与第二类展示对象信息中的展示对象信息的对应关系。
利用本发明实施例,可得到各个页面的页面关键字与展示对象信息的对应关系。
在一些实例中,采用图5提供的方式计算页面图片与展示对象图片的相似度:
步骤501,确定页面图片的纹理特征向量与展示对象图片的纹理特征向量的余弦相似度,作为第一相似度。
步骤502,确定页面图片的颜色特征向量与展示对象图片的颜色特 征向量的余弦相似度,作为第二相似度。
步骤503,确定页面图片的局部不变特征向量与展示对象图片的局部不变特征向量的余弦相似度,作为第三相似度。
步骤504,根据第一相似度、第二相似度以及第三相似度,计算页面图片与展示对象图片的相似度。
其中,步骤501、步骤502以及步骤503的执行顺序的先后不做限定。
具体实施时,可选取第一相似度、第二相似度和第三相似度中的最大值,作为页面图片与展示对象图片的相似度,或者,从第一相似、第二相似度和第三相似度中任意选取两个数值,对选取的两个数值求平均值作为页面图片与展示对象图片的相似度,或者,计算第一相似度、第二相似度和第三相似度的均值,将该均值作为页面图片与展示对象图片的相似度,这里不做限定。利用本发明实施例可得到任意两幅图片的相似度。
在一些实例中,计算第一相似度、第二相似度以及第三相似度的加权求和结果;将得到的加权求和结果作为页面图片与展示对象图片的相似度。
具体实施时,针对纹理特征、颜色特征以及局部不变特征分别设置不同的权值,计算第一相似度、第二相似度以及第三相似度的加权求和结果,作为页面图片与展示对象图片的相似度。
举例说明,假设纹理特征对应的权值为a、颜色特征对应的权值为b、局部不变特征对应的权值为c,第一相似度为X、第二相似度为Y、第三相似度为Z,则上述加权求和结果H=aX+bY+cZ。
具体实施时,可按照图6提供的内容确定页面图片的局部不变特征向量:
步骤601,利用sift(Scale-invariant feature transform,尺度不变特征转换)算法从页面图片以及存储的各个展示对象图片中提取视觉词汇向量,并存入视觉词汇集合。
步骤602,利用K-Means(即K均值)聚类算法,将视觉词汇集合中的视觉词汇向量划分为N个簇。
步骤603,确定得到的N个簇中每个簇的中心视觉词汇向量。
步骤604,针对页面图片中的每个视觉词汇向量,确定N个中心视觉词汇向量中与该视觉词汇向量之间的距离最小的中心视觉词汇向量,并将确定的该中心视觉词汇向量所属的簇确定为该页面图片中该词汇向量所属的簇。
步骤605,统计页面图片中属于每个簇的视觉词汇向量的数目。
步骤606,将统计得到的各个数目组成一个N维向量,作为该页面图片的局部不变特征向量。
假设执行步骤602后,得到3个簇,执行步骤606后,得到的3维向量为[3,6,9],其中,3为该页面图片中属于第一簇的视觉词汇向量的个数,6为该页面图片中属于第二簇的视觉词汇向量的个数,9为该页面图片中属于第三簇的视觉词汇向量的个数。
具体实施时,确定展示对象图片的局部不变特征向量的过程与确定页面图片的局部不变特征向量的过程类似(即将步骤604和步骤605中的页面图片替换为展示对象图片即可),这里不做赘述,得到的展示对象图片的局部不变特征向量也是一个N维向量。
具体实施时,可按照以下方式确定任一图片的颜色特征向量:
假设每种原色可以取256个值,将0~255分成四个分区:0~63为第0分区,64~127为第1分区,128~191为第2分区,192~255为第3分区,其中,R(红)G(绿)B(蓝)三原色分别对应四个分区,得 到64种分区组合,针对图片中的每个像素点,确定该像素点的R值、G值和B值,分别确定R值、G值以及B值所属的分区;将该像素点的R值、G值以及B值所属的分区的标识作为该像素点对应的分区组合;统计该图片中属于各个分区组合的像素点的个数;该图片中属于各个分区组合的像素点的个数组成一个64维的向量,作为该图片的颜色特征向量。其中,分区组合中,第一个数字代表R所属的分区,第二数字代表G所属的分区,第三个数字代表B所属的分区。
以图片中的一个像素点为例,说明确定该像素点所属的分区组合的过程:
假设该像素点的R值为60,G值为25,B值为100,则,60和25均属于第0分区,100属于第1分区,因此,该像素点所属的分组组合为001。
具体实施时,可按照以下方式确定任一图片的纹理特征向量:
针对图片中的每个像素点,确定以该像素点为中心点的Q*Q的邻域,其中,Q可以为正整数;从该邻域内的第一个像素点开始,按照顺时针的轮选顺序,将该邻域内灰度值小于该邻域的中心像素点的灰度值的边缘像素点记为0,将该邻域内灰度值大于或等于该邻域的中心像素点的灰度值的边缘像素点记为1,统计该邻域内除中心像素点之外的像素点的值,得到该图片中该像素点对应的Q-1位二进制串;将该图片中该像素点对应的Q-1位二进制串转换为十进制数值,得到该图片中该像素点对应的十进制数值;按照预设直方图统计方法,对该图片中每个像素点对应的十进制数值进行直方图统计,确定直方图对应的每个十进制数值范围内所包含的该图片中的像素点个数;将直方图对应的每个十进制数值范围内所包含的该图片中的像素点个数组成一个向量,作为该图片的纹理特征向量,其中,该纹理特征向量的维度等于十进制数值范围 的个数。在一些实例中,Q的取值为3,还可以为其它数值,这里不做限定。
下面结合实际应用场景,对本发明实施例提供的推广信息投放方法进行详细说明。
如图7所示,为本发明实施例提供的推广信息投放方法的应用场景示意图,其中,包括用户终端701和服务器702,其中,用户终端701中安装有新闻应用程序,用户70可操作用户终端701中的新闻应用程序,服务器702中安装有推广信息投放系统,该推广信息投放系统包括线上模块和离线模块。其中,用户终端701和服务器702可以通过局域网、广域网或移动互联网等通信网络进行通信;用户终端设备701可以为便携设备(例如:手机、平板、笔记本电脑等),也可以为个人电脑(PC,Personal Computer),新闻应用程序可以为腾讯新闻、天天快报等用于浏览新闻的软件。需要注意的是,上述应用场景仅是为了便于理解本发明的精神和原理而示出,本发明的实施方式在此方面不受任何限制。相反,本发明的实施方式可以应用于适用的任何场景。
在一种可能的实际应用中,当推广信息为广告、展示对象为商品时,广告主将其预推广的商品的商品图片、商品标题以及商品价格属性发送至离线模块;离线模块根据广告主发送的商品属性生成符合广告格式的广告数据,并将广告数据发送至线上模块的广告库进行保存,其中,商品属性包括商品图片、商品标题以及商品价格;离线模块针对每个商品属性,提取该商品属性中的商品信息,其中,商品信息包括商品图片以及商品关键字;离线模块针对当日预推荐的每个页面,提取该页面的页面图片标识以及页面关键字,并按照本发明实施例提供的确定设置页面图片标识与商品信息的对应关系的方式以及设置确定页面关键字与商品信息的对应关系的方式,分别确定设置页面图片标识与商品信息的对 应关系以及页面关键字与商品信息的对应关系,并将确定设置的页面图片标识与商品信息的对应关系以及页面关键字与商品信息的对应关系发送至线上模块中的对应关系数据库进行保存。
线上模块响应于用户70通过用户终端701中的新闻应用程序发送的查看指定页面的操作,提取所述指定页面的页面图片标识和页面关键字;从确定的页面图片标识与商品信息的对应关系中查找所述指定页面的页面图片标识对应的商品信息;其中,页面图片标识对应的商品信息中的商品图片与该页面图片标识对应的页面图片相匹配;若查找到所述指定页面的页面图片标识对应的商品信息,则根据预先存储的商品信息与广告的对应关系,从广告库中获取所述指定页面的页面图片标识对应的商品信息对应的广告;根据获取的广告,对所述用户进行广告投放;若未查找到所述指定页面的页面图片标识对应的商品信息,则从确定的页面关键字与商品信息的对应关系中查找所述指定页面的页面关键字对应的商品信息,其中,页面关键字对应的展示对象信息中的展示对象关键字与该页面关键字相匹配;若查找到所述指定页面的页面关键字对应的商品信息,则根据预先存储的商品信息与广告的对应关系,从广告库中获取所述指定页面的页面关键字对应的商品信息对应的广告;根据获取的广告,对所述用户进行广告投放。
本发明实施例还提供一种推广信息投放装置,如图8所示,包括:
提取模块801,用于响应于查看指定页面的操作,提取所述指定页面的页面图片标识;
确定模块802,用于确定所述指定页面的页面图片标识对应的推广信息;
投放模块803,用于根据确定的推广信息,对用户进行推广信息投放。
在一些实例中,所述确定模块802用于:
从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配;
若查找到所述指定页面的页面图片标识对应的展示对象信息,则根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息;
将获取的推广信息确定为所述指定页面的页面图片标识对应的推广信息。
在一些实例中,本发明实施例提供的推广信息投放装置,还包括,查找模块804,获取模块805,其中:
所述提取模块801,还用于提取所述指定页面的页面关键字;
所述查找模块804,用于若所述确定模块802未查找到所述指定页面的页面图片标识对应的展示对象信息,则从设置的页面关键字与展示对象信息的对应关系中查找所述指定页面的页面关键字对应的展示对象信息,其中,页面关键字对应的展示对象信息中的展示对象关键字与该页面关键字相匹配;
所述获取模块805,还用于若所述查找模块804查找到所述指定页面的页面关键字对应的展示对象信息,则根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面关键字对应的展示对象信息对应的推广信息;
所述投放模块803,还用于根据所述获取模块805获取的推广信息,对所述用户进行推广信息投放。
在一些实例中,本发明实施例提供的推广信息投放装置,还包括:
第一设置模块806,用于按照以下方式设置页面图片标识与展示对象信息的对应关系:
针对每个页面,分别执行以下操作:
分别计算该页面的页面图片与预先存储的各个展示对象信息中的展示对象图片的相似度;
将与该页面的页面图片的相似度超过第一相似度阈值的展示对象图片,作为与该页面的页面图片相匹配的展示对象图片;
确定与该页面的页面图片相匹配的展示对象图片所属的展示对象信息;
保存该页面的页面图片对应的页面图片标识与确定的展示对象信息的对应关系。
在一些实例中,本发明实施例提供的推广信息投放装置,还包括:
第二设置模块807,用于按照以下方式确定页面关键字与展示对象信息的对应关系:
针对每个页面,分别执行以下操作:
分别计算该页面的页面关键字与预先存储的各个展示对象信息中的展示对象关键字的相似度;
将与该页面的页面关键字的相似度超过第二相似度阈值的展示对象关键字,作为与该页面的页面关键字相匹配的展示对象关键字;
确定与该页面的页面关键字相匹配的展示对象关键字所属的展示对象信息;
保存该页面的页面关键字与确定的展示对象信息的对应关系。
在一些实例中,采用如下方式计算页面图片与展示对象图片的相似度:
确定页面图片的纹理特征向量与展示对象图片的纹理特征向量的 余弦相似度,作为第一相似度;以及
确定页面图片的颜色特征向量与展示对象图片的颜色特征向量的余弦相似度,作为第二相似度;以及
确定页面图片的局部不变特征向量与展示对象图片的局部不变特征向量的余弦相似度,作为第三相似度;
根据第一相似度、第二相似度以及第三相似度,计算页面图片与展示对象图片的相似度。
在一些实例中,所述根据第一相似度、第二相似度以及第三相似度,计算页面图片与展示对象图片的相似度,包括:
计算第一相似度、第二相似度以及第三相似度的加权求和结果;
将得到的加权求和结果作为页面图片与展示对象图片的相似度。
本发明实施例提供一种非易失性计算机存储介质,所述计算机存储介质存储有可执行程序,该可执行程序被处理器执行实现上述实施例提供的任一推广信息投放方法的步骤。
本发明实施例还提供一种计算设备,包括存储器、处理器及存储在存储器上的计算机程序,所述处理器执行所述程序时实现上述实施例中的任一推广信息投放方法的步骤。
本发明实施例还提供一种计算设备,用于执行上述实施例中的推广信息投放方法,如图9所示,为本发明实施中所述的计算设备的硬件结构示意图,该计算设备具体可以为台式计算机、便携式计算机、智能手机、平板电脑等。具体地,该计算设备可以包括存储器901、处理器902及存储在存储器上的计算机程序,所述处理器执行所述程序时实现上述实施例中的任一推广信息投放方法的步骤。其中,存储器901可以包括只读存储器(ROM)和随机存取存储器(RAM),并向处理器902提供存储器901中存储的程序指令和数据。
进一步地,本发明实施例中所述的计算设备还可以包括输入装置903以及输出装置904等。输入装置903可以包括键盘、鼠标、触摸屏等;输出装置904可以包括显示设备,如液晶显示器(Liquid Crystal Display,LCD)、阴极射线管(Cathode Ray Tube,CRT),触摸屏等。存储器901,处理器902、输入装置903和输出装置904可以通过总线或者其他方式连接,图9中以通过总线连接为例。
处理器902调用存储器901存储的程序指令并按照获得的程序指令执行上述实施例提供的推广信息投放方法。
利用本发明实施例提供的推广信息投放方法、装置、介质及设备,具有以下有益效果:
用户选择查看的页面中的内容一般为用户感兴趣的内容,确定指定页面的页面图片标识对应的推广信息,确定出的推广信息与指定页面中的图片相匹配,使得向用户投放的推广信息与用户当前选择查看的页面的内容相关,此时,向用户所投放的推广信息为用户感兴趣的推广信息的可能性较大,可以在一定程度上提高页面中所展示的推广信息的点击率。
应当注意,尽管在上文详细描述中提及了推广信息投放装置的若干模块,但是这种划分仅仅是示例性的并非强制性的。实际上,根据本发明的实施方式,上文描述的两个或更多模块的特征和功能可以在一个模块中具体化。反之,上文描述的一个模块的特征和功能可以进一步划分为由多个模块来具体化。
本发明是参照根据本发明实施例的方法、设备(系统)、和计算机程序产品的流程图和/或方框图来描述的。应理解可由计算机程序指令实现流程图和/或方框图中的每一流程和/或方框、以及流程图和/或方框图中的流程和/或方框的结合。可提供这些计算机程序指令到通用 计算机、专用计算机、嵌入式处理机或其他可编程数据处理设备的处理器以产生一个机器,使得通过计算机或其他可编程数据处理设备的处理器执行的指令产生用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的装置。
这些计算机程序指令也可存储在能引导计算机或其他可编程数据处理设备以特定方式工作的计算机可读存储器中,使得存储在该计算机可读存储器中的指令产生包括指令装置的制造品,该指令装置实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能。
这些计算机程序指令也可装载到计算机或其他可编程数据处理设备上,使得在计算机或其他可编程设备上执行一系列操作步骤以产生计算机实现的处理,从而在计算机或其他可编程设备上执行的指令提供用于实现在流程图一个流程或多个流程和/或方框图一个方框或多个方框中指定的功能的步骤。
尽管已描述了本发明的优选实施例,但本领域内的技术人员一旦得知了基本创造性概念,则可对这些实施例作出另外的变更和修改。所以,所附权利要求意欲解释为包括优选实施例以及落入本发明范围的所有变更和修改。
显然,本领域的技术人员可以对本发明进行各种改动和变型而不脱离本发明的精神和范围。这样,倘若本发明的这些修改和变型属于本发明权利要求及其等同技术的范围之内,则本发明也意图包含这些改动和变型在内。

Claims (14)

  1. 一种推广信息投放方法,由计算设备执行,包括:
    响应于查看指定页面的操作,提取所述指定页面的页面图片标识;
    从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配;
    根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息;
    根据获取的推广信息,对用户进行所述推广信息投放。
  2. 根据权利要求1所述的方法,其中,还包括:
    提取所述指定页面的页面关键字;则
    若未查找到所述指定页面的页面图片标识对应的展示对象信息,则从设置的页面关键字与展示对象信息的对应关系中查找所述指定页面的页面关键字对应的展示对象信息,其中,页面关键字对应的展示对象信息中的展示对象关键字与该页面关键字相匹配;
    若查找到所述指定页面的页面关键字对应的展示对象信息,则根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面关键字对应的展示对象信息对应的推广信息。
  3. 根据权利要求1所述的方法,其中,按照以下方式设置页面图片标识与展示对象信息的对应关系:
    针对每个页面,分别执行以下操作:
    分别计算该页面的页面图片与预先存储的各个展示对象信息中的展示对象图片的相似度;
    将与该页面的页面图片的相似度超过第一相似度阈值的展示对象图片,作为与该页面的页面图片相匹配的展示对象图片;
    确定与该页面的页面图片相匹配的展示对象图片所属的展示对象信息;
    保存该页面的页面图片对应的页面图片标识与确定的展示对象信息的对应关系。
  4. 根据权利要求2所述的方法,其中,按照以下方式确定页面关键字与展示对象信息的对应关系:
    针对每个页面,分别执行以下操作:
    分别计算该页面的页面关键字与预先存储的各个展示对象信息中的展示对象关键字的相似度;
    将与该页面的页面关键字的相似度超过第二相似度阈值的展示对象关键字,作为与该页面的页面关键字相匹配的展示对象关键字;
    确定与该页面的页面关键字相匹配的展示对象关键字所属的展示对象信息;
    保存该页面的页面关键字与确定的展示对象信息的对应关系。
  5. 根据权利要求4所述的方法,其中,采用如下方式计算页面图片与展示对象图片的相似度:
    确定页面图片的纹理特征向量与展示对象图片的纹理特征向量的余弦相似度,作为第一相似度;以及
    确定页面图片的颜色特征向量与展示对象图片的颜色特征向量的余弦相似度,作为第二相似度;以及
    确定页面图片的局部不变特征向量与展示对象图片的局部不变特征向量的余弦相似度,作为第三相似度;
    根据第一相似度、第二相似度以及第三相似度,计算页面图片与展 示对象图片的相似度。
  6. 根据权利要求5所述的方法,其中,所述根据第一相似度、第二相似度以及第三相似度,计算页面图片与展示对象图片的相似度,包括:
    计算第一相似度、第二相似度以及第三相似度的加权求和结果;
    将得到的加权求和结果作为页面图片与展示对象图片的相似度。
  7. 根据权利要求1所述的方法,其中,所述根据获取的推广信息,对用户进行所述推广信息投放,包括:
    将所述推广信息插入所述指定页面的指定位置处;
    将插入推广信息后的指定页面返回给所述用户。
  8. 一种推广信息投放装置,包括:处理器;用于存储处理器可执行指令的存储器;其中,所述处理器被配置为:
    响应于查看指定页面的操作,提取所述指定页面的页面图片标识;
    从设置的页面图片标识与展示对象信息的对应关系中查找所述指定页面的页面图片标识对应的展示对象信息;其中,页面图片标识对应的展示对象信息中的展示对象图片与该页面图片标识对应的页面图片相匹配;
    根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面图片标识对应的展示对象信息对应的推广信息;
    根据获取的推广信息,对用户进行所述推广信息投放。
  9. 根据权利要求8所述的装置,其中,所述处理器进一步被配置为:
    提取所述指定页面的页面关键字;
    若未查找到所述指定页面的页面图片标识对应的展示对象信息,则从设置的页面关键字与展示对象信息的对应关系中查找所述指定页面的页面关键字对应的展示对象信息,其中,页面关键字对应的展示对象 信息中的展示对象关键字与该页面关键字相匹配;
    若查找到所述指定页面的页面关键字对应的展示对象信息,则根据预先存储的展示对象信息与推广信息的对应关系,从推广信息库中获取所述指定页面的页面关键字对应的展示对象信息对应的推广信息。
  10. 根据权利要求8所述的装置,其中,所述处理器进一步被配置为:
    按照以下方式设置页面图片标识与展示对象信息的对应关系:
    针对每个页面,分别执行以下操作:
    分别计算该页面的页面图片与预先存储的各个展示对象信息中的展示对象图片的相似度;
    将与该页面的页面图片的相似度超过第一相似度阈值的展示对象图片,作为与该页面的页面图片相匹配的展示对象图片;
    确定与该页面的页面图片相匹配的展示对象图片所属的展示对象信息;
    保存该页面的页面图片对应的页面图片标识与确定的展示对象信息的对应关系。
  11. 根据权利要求9所述的装置,其中,所述处理器进一步被配置为:
    按照以下方式确定页面关键字与展示对象信息的对应关系:
    针对每个页面,分别执行以下操作:
    分别计算该页面的页面关键字与预先存储的各个展示对象信息中的展示对象关键字的相似度;
    将与该页面的页面关键字的相似度超过第二相似度阈值的展示对象关键字,作为与该页面的页面关键字相匹配的展示对象关键字;
    确定与该页面的页面关键字相匹配的展示对象关键字所属的展示对象信息;
    保存该页面的页面关键字与确定的展示对象信息的对应关系。
  12. 根据权利要求10所述的装置,其中,所述处理器进一步被配置为采用如下方式计算页面图片与展示对象图片的相似度:
    确定页面图片的纹理特征向量与展示对象图片的纹理特征向量的余弦相似度,作为第一相似度;以及
    确定页面图片的颜色特征向量与展示对象图片的颜色特征向量的余弦相似度,作为第二相似度;以及
    确定页面图片的局部不变特征向量与展示对象图片的局部不变特征向量的余弦相似度,作为第三相似度;
    根据第一相似度、第二相似度以及第三相似度,计算页面图片与展示对象图片的相似度。
  13. 一种非易失性计算机存储介质,所述计算机存储介质存储有可执行程序,该可执行程序被处理器执行实现权利要求1-7任一所述方法的步骤。
  14. 一种计算设备,包括存储器、处理器及存储在存储器上的计算机程序,所述处理器执行所述程序时实现权利要求1-7任一所述方法的步骤。
PCT/CN2018/110407 2017-10-19 2018-10-16 推广信息投放方法、装置、介质及设备 WO2019076294A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710977151.6 2017-10-19
CN201710977151.6A CN110020123B (zh) 2017-10-19 2017-10-19 一种推广信息投放方法、装置、介质及设备

Publications (1)

Publication Number Publication Date
WO2019076294A1 true WO2019076294A1 (zh) 2019-04-25

Family

ID=66174305

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/110407 WO2019076294A1 (zh) 2017-10-19 2018-10-16 推广信息投放方法、装置、介质及设备

Country Status (2)

Country Link
CN (1) CN110020123B (zh)
WO (1) WO2019076294A1 (zh)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110489685A (zh) * 2019-08-19 2019-11-22 腾讯科技(武汉)有限公司 一种网页显示方法、系统和相关装置及存储介质
CN111953738B (zh) * 2020-07-20 2022-09-20 上海淇馥信息技术有限公司 信息投放方法及其装置、电子设备、存储介质
CN113822734B (zh) * 2021-01-18 2024-07-19 北京沃东天骏信息技术有限公司 用于生成信息的方法和装置

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008096759A (ja) * 2006-10-13 2008-04-24 Yafoo Japan Corp 画像データにマッチした広告を配信するための広告配信方法および広告配信装置
CN104156375A (zh) * 2013-05-16 2014-11-19 王方淇 一种基于众包的信息投放方法和系统
CN104217354A (zh) * 2014-09-02 2014-12-17 天津云辰科技有限公司 基于用户浏览网页内图片的广告投放方法
CN105378782A (zh) * 2013-07-11 2016-03-02 悟图索知株式会社 基于用户兴趣的产品信息推荐系统
CN105956878A (zh) * 2016-04-25 2016-09-21 广州出益信息科技有限公司 一种网络广告推送的方法及装置
CN106649799A (zh) * 2016-12-28 2017-05-10 北京金山安全软件有限公司 一种内容推广方法、装置及电子设备

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103425680B (zh) * 2012-05-18 2018-02-02 阿里巴巴集团控股有限公司 页面展示广告的选取方法和系统
CN106920117B (zh) * 2015-12-24 2020-06-30 北京奇虎科技有限公司 投放推广信息的方法及装置
CN106933816A (zh) * 2015-12-29 2017-07-07 北京大唐高鸿数据网络技术有限公司 基于全局特征和局部特征的跨镜头目标检索系统及方法

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2008096759A (ja) * 2006-10-13 2008-04-24 Yafoo Japan Corp 画像データにマッチした広告を配信するための広告配信方法および広告配信装置
CN104156375A (zh) * 2013-05-16 2014-11-19 王方淇 一种基于众包的信息投放方法和系统
CN105378782A (zh) * 2013-07-11 2016-03-02 悟图索知株式会社 基于用户兴趣的产品信息推荐系统
CN104217354A (zh) * 2014-09-02 2014-12-17 天津云辰科技有限公司 基于用户浏览网页内图片的广告投放方法
CN105956878A (zh) * 2016-04-25 2016-09-21 广州出益信息科技有限公司 一种网络广告推送的方法及装置
CN106649799A (zh) * 2016-12-28 2017-05-10 北京金山安全软件有限公司 一种内容推广方法、装置及电子设备

Also Published As

Publication number Publication date
CN110020123A (zh) 2019-07-16
CN110020123B (zh) 2023-05-12

Similar Documents

Publication Publication Date Title
US11282116B2 (en) Image quality assessment to merchandise an item
US20200175550A1 (en) Method for identifying advertisements for placement in multimedia content elements
US10528618B2 (en) Image processing and matching
US8583640B2 (en) Method and system of search result diversification
WO2018166288A1 (zh) 信息呈现方法和装置
US10360623B2 (en) Visually generated consumer product presentation
US10592855B2 (en) Method, system, and computer program product for planogram generation
CN105574089B (zh) 知识图谱的生成方法及装置、对象对比方法及装置
CN103377287A (zh) 一种投放物品信息的方法和装置
WO2019076294A1 (zh) 推广信息投放方法、装置、介质及设备
CN113283321A (zh) 用于自动内容识别的标记广告帧的方法和系统
WO2016201800A1 (zh) 信息获取方法、服务器、终端及数据库构建方法、装置
US11308262B2 (en) Systems and methods for converting static image online content to dynamic online content
US20160171548A1 (en) Method for identifying advertisements for placement in multimedia content elements
CN104185041A (zh) 视频交互广告的自动生成方法和系统
CN112100511B (zh) 一种偏好程度数据获得方法、装置以及电子设备
US10671660B2 (en) Contextual ranking of photos and text in search
CN110765296A (zh) 图像搜索方法、终端设备及存储介质
US20140200971A1 (en) System and method for matching informative content to a multimedia content element based on concept recognition of the multimedia content
CN115631495B (zh) 一种基于页面分析的spu获取方法及存储介质
CN106484710B (zh) 动态数据处理方法和装置、信息显示方法和装置
JP2018180610A (ja) 情報処理装置、情報配信方法、及び情報配信プログラム
CN115618018B (zh) 一种spu的知识图谱构建方法及存储介质
JP7190620B2 (ja) 情報処理装置、情報配信方法、及び情報配信プログラム
WO2024104286A1 (zh) 一种视频处理方法、装置、电子设备和存储介质

Legal Events

Date Code Title Description
NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18869134

Country of ref document: EP

Kind code of ref document: A1