CN101957825A - Method for searching image based on image and video content in webpage - Google Patents
Method for searching image based on image and video content in webpage Download PDFInfo
- Publication number
- CN101957825A CN101957825A CN2009101087887A CN200910108788A CN101957825A CN 101957825 A CN101957825 A CN 101957825A CN 2009101087887 A CN2009101087887 A CN 2009101087887A CN 200910108788 A CN200910108788 A CN 200910108788A CN 101957825 A CN101957825 A CN 101957825A
- Authority
- CN
- China
- Prior art keywords
- image
- target
- user
- searching
- search
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012545 processing Methods 0.000 claims description 18
- 238000010606 normalization Methods 0.000 claims description 13
- 238000007781 pre-processing Methods 0.000 claims description 5
- 239000013598 vector Substances 0.000 claims description 4
- 238000012986 modification Methods 0.000 claims description 2
- 230000004048 modification Effects 0.000 claims description 2
- 230000008569 process Effects 0.000 description 20
- 230000003068 static effect Effects 0.000 description 4
- 230000006399 behavior Effects 0.000 description 3
- 239000003795 chemical substances by application Substances 0.000 description 2
- 238000010276 construction Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000013473 artificial intelligence Methods 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000012790 confirmation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 230000006870 function Effects 0.000 description 1
- 230000002452 interceptive effect Effects 0.000 description 1
- 230000016507 interphase Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention relates to a method for searching an image based on an image and video content in a webpage, which comprises the following steps of: 1) receiving an image research request, 2) pre-treating a researched image, 3) normalizing the image research, 4) searching a target image, 5) setting a normalized true target and turning back to step two if user requires to adjust, or skipping to step six, 6) evaluating a user community and 7) listing a final researched targeted images. The method of the invention has better property, is more intelligent and is more practical.
Description
Technical Field
The invention relates to a method for searching information in the Internet, in particular to a method for searching images of information such as images and videos in the Internet.
Background
Along with the fact that the internet has become a common and important information source in daily life, many electronic businesses are built on the basis of search engines of internet information and are growing and maturing.
The general pattern is: people find some kind of things which are interested in themselves from massive information, then use search engines of general or vertical application types to find them, and then go to e-commerce websites to sweep, namely continue to use the in-site search engines provided by e-commerce websites to locate the specific things which are bought by themselves. The whole process is a process from coarse to fine in the thinking and behaviors of consumers, and the search engine plays a very key role. In this information-to-business process, the user's behavior can be understood as being split into three phases: first, there is extensive acquisition of information; then, the accurate positioning of the information is carried out; finally, it is a business implementation. In addition, various services come from different internet enterprises, in the conversion process, the instant shopping momentum of people tends to gradually move back along with the continuation of surfing time, the process is not too painful for commodities which can be described by precise language characters, for example, a certain model 3C product of a certain brand, such as a mobile phone, a computer and the like, can be purchased by a content provider of a webpage, and the characters describing the product can be linked with a search engine, so that the whole process is smooth, and the user behavior is more convenient.
Firstly, for commodities which are closely related to the purchase and the image display, such as products of fashion and clothes, the image cannot be taken for direct use due to the copyright problem of the picture reference; meanwhile, the user may be interested in the clothing worn by a certain model in a picture, which completely exceeds the processing capability of the existing text search engine and text-based image search engine (for labeling images, such as image search of Google). However, the nature of the image search based on the picture text identification content, which is being developed by the current search engine macrosmus, such as Google, Yahoo, Baidu, etc., is also text search.
Secondly, after the user searches the relevant information, how to further meet the business requirements of the user is also a great problem. The current search engine has a huge momentum, in order to solve the link from content to business, a vertical search shopping price comparison service based on text content is generally developed, namely, in order to solve the problems that the precision is too low and the search result is filled with too much useless information in the current search engine, but the challenge is still filled in finding the information needed by the user in the huge image library which is more intuitive for the user.
It can be seen that most of the image search engines of the so-called "image search" at present are text keyword searches, usually, images are collected from the internet, then some text information of the web pages where the images are located is extracted to build the label index for the images, and even more simply, only the titles of the web pages where the images are located are extracted as the image building index, and when a user submits a query to the search engine, only the similarity between the query and the pre-extracted information is considered and the result is returned.
In the real image search field, two layers are not considered at present.
Firstly, the method comprises the following steps: in the technical aspect of image recognition, due to the limitation of technical development, the current solutions mostly consider only color distribution and texture distribution of an image, but do not consider the role of more multidimensional rich information such as block features, shape features, contrast features and the like of the image content in image recognition, and use these elements to enrich and construct an image feature library.
Secondly, the method comprises the following steps: at the business application level of an image search engine, the current image submission and result ranking methods often adopt a simpler mode like outputting results of text search, the precision rate of image search is not ideal, and only the accuracy rate is concerned, the user is not considered to be interested in only part of the seen images, the copyright problem of the images is not considered, and what the user wants to obtain through an image search service is not considered from the perspective of the user requirements.
Moreover, the existing image search engines hardly consider the following two cases:
1) copyright of images because searching as an image inevitably processes the image in the background, but finding the searched image by a web robot and directly storing the image as a template is not feasible in commercial copyright regulations;
2) the focus of the images, which previously only considered the use of images, did not consider that the user may only be interested in parts of the images, rather than the entire images, such as two models, and the user may only be interested in clothing on a model, rather than searching with the entire images, and may only wish to search through parts of the clothing to find e-commerce information and merchants that provide similar goals.
Disclosure of Invention
The technical problem to be solved by the invention is to overcome the defects of the prior art, and provide a method for combining the ranking algorithm of the image search engine and the electronic commerce, which not only fully uses the existing text search engine and the mature technology of user community construction, but also considers some inherent characteristics of the image, so that the result of the image search has higher correlation, and the requirement of user query is better met.
The technical scheme adopted by the invention for solving the technical problems is to provide an image searching method based on image and video contents in a webpage, which comprises the following steps: firstly, receiving an image searching request; secondly, preprocessing a search image; thirdly, searching a normalized image; fourthly, searching a target image; if yes, setting a normalization true target, returning to the step two, and otherwise, continuing; sixthly, evaluating the user community; and seventhly, searching a list of final target images. Wherein,
the first step comprises the following steps: and intercepting a target image which is contained in the embedded image of the webpage content and is really interested by the user by adopting a webpage embedding and browser plug-in mode.
The second step comprises the following steps: and analyzing and processing the target image content uploaded by the client by the background server, namely constructing and storing sample information of the target image according to the color, texture, shape, block shape, contrast and vector characteristics of the target image.
The third step comprises: and carrying out normalization processing on the sample information of the target image to obtain a normalized true target.
The fourth step comprises the following steps: and taking the normalized true target as a comparison object, searching a directional electronic commerce website by using a network robot, and establishing a target image search list index from high to low according to the similarity.
The fifth step comprises the following steps: the opportunity is provided for the user to make active modifications to the normalized true target.
The sixth step comprises: and dynamically adjusting the network search result according to the content created by the user.
Compared with the prior art, the image searching method based on the image and video content in the webpage has better performance, is more intelligent and is more practical.
Drawings
Fig. 1 is a flowchart of an embodiment of an image searching method based on images and video contents in a web page according to the present invention.
Fig. 2 is a schematic system structure diagram of an embodiment of the image searching method based on the image and video content in the web page according to the present invention.
Detailed Description
The following detailed description is to be read with reference to the best mode embodiment as illustrated in the accompanying drawings.
The embodiment of the image searching method based on the image and video content in the webpage, as shown in fig. 1, includes: step 10: receiving an image search request; step 20: preprocessing a search image; step 30: searching a normalized image; step 40: searching a target image; step 50: if the user does not adjust, if so, setting a normalization true target, and returning to the step two, otherwise, continuing; step 60: evaluating a user community; and step 10: and searching the list by the final target image. Wherein,
original image input source: images from web site content, wherein the user submits the images in the following two input modes:
embedding a webpage: the method is characterized in that a plug-in is installed on a website, an icon button is added to an image on the webpage, a layer can be emerged when a user clicks the icon button, the original image can be displayed by the interactive animation effect through the layer, the user can select any part of the image for searching according to own wish frames, and then a search engine can give a result list of related targets.
The browser plug-in: the plug-in is installed on the browser, so that any image of any website can be selected by a user for searching, and then a related target list is obtained.
For copyright reasons, even if the partner has a photo, the photo cannot be shared for use, but the system can be grabbed by a network robot on the internet (oriented website alliance such as clothes and the like), and because the set of websites is limited, oriented image grabbing is feasible.
For a captured picture, an Image search like GOOGLE is to be accompanied by a link (legal), but must be in the form of a thumbnail and the picture link is to point to the source picture address.
The system needs to extract the characteristics of the image when developing visual search, so the picture (placed in an unstructured database) stored by the system is a format which is between the original image and the thumbnail, meets the requirements after being processed, and is pointed by the link of the source image, so that the picture can be legally stored and used.
Intercepting interested contents in an image embedded with webpage contents, which is the basis of all work and can be realized by adopting webpage embedding and browser plug-in modes; and extracting images from webpage contents, including conventional internet and wireless internet webpages, acquiring image contents in which a user is interested to be queried and searched, and uploading the image contents to a background server for processing.
The image content uploaded to the background server is subjected to computer analysis and storage processing, which is the basis for identifying images by later computer artificial intelligence; in order to more accurately identify the image without missing important image information, the image feature values that can be extracted are: color, texture, shape, block features, contrast, vector, etc. to obtain sample information for the target image, i.e., a library of features corresponding to a particular image. Because the image is identified by the computer, the computer processing of the image information to extract the key features of the image is particularly important, so that the objectivity and stability of the image information description can be ensured, and meanwhile, an image information base is established to facilitate the subsequent image normalization processing and fuzzy identification process. Obviously, in practical engineering implementation, in order to ensure the accuracy of information and not to omit important information in an image, the more features are, the better, and of course, the complexity of computer processing and the processing capability speed need to be balanced.
In the process, if the transmitted image does not find the true target image, the system directly enters the following steps to search for the target image; meanwhile, the image is automatically set as an agent and a real target, and a series of index tables are established, in short, the one-to-many relation between the agent and the real target image and a target image search result list obtained by searching the target image is established.
The method is characterized in that the normalization processing is carried out on the content of the selected image or part of the image, an image template which is completely consistent with or very similar to the requirement of a user, namely a 'true target', is found, and the step is the basis for searching completely identical or similar image content by subsequently continuing to use a computer network robot to a directional electronic commerce website. The normalization is performed because even the same photo image, the user may intercept the image content from different web pages of different websites, or the user may search for some interested target images of an image, such as a piece of clothing worn on a model, and so on, and therefore it is necessary to perform the normalization process on the image to obtain the meta-image of the "true target". This is a very important "many-to-one" image accurate recognition process, where many refer to the so-called image search requests caused by the fact that the image of the same target is derived from a plurality of different scenes, or in fact that different source images describe the image of the same target, or that "stealth" knowledge becomes "explicit" knowledge. The method is an indispensable link for ensuring the consistency of the content of the image search request received by the search engine or meeting the image requirements of the user, and is an indispensable step for accelerating the subsequent target image search, and the result list of the target image search is directly output without repeatedly searching the target image corresponding to one image with the same essence.
This is particularly important for those completely new search requests where there is no image of a real target in the system or where the proxy "real target" image was mentioned in the previous step, but to prevent the image search results from being too biased (the image features are almost the same from the image search perspective and are unavoidable), the system gives the user an opportunity to participate in a judgment, such as for the user to assist in setting the "real target" image for the image search.
This is a true target image search, and therefore the search results are output in a true sense to the user, and a "one-to-many" so-called "fuzzy" target search strategy is performed, and therefore the image feature value selection and the previous normalization process used here are not the same.
Target image searching and search result preprocessing, namely image searching, wherein the image searching is still performed, but because the image searching is based on the normalized true target, the image characteristic value used in the image searching is different from the image characteristic value used in the previous normalized image searching, and a target image searching list index is established in a system platform according to the image similarity from high to low; since the knowledge and decision making of things is a coarse-to-fine process, what the user first needs to be similar to and a result similar or close to the target image to assist in making the next decision when seeing the image of interest. The step is a one-to-many image fuzzy recognition processing process close to the requirement of a user, wherein a plurality refers to a plurality of similar or similar results; in order to describe the target image more accurately, the system also assists by the context keyword information of the webpage content where the image is located, so that the target image search result is more accurate.
In general, because the real target library is large enough, the search result can relatively accurately meet the user requirement through the normalized image search and the target image search; however, it is considered that if there is a new object, or if there is a significant difference between its image and the existing image of the real object, it can be considered as a new object, such as a black and white baseball cap photographed from above and a black and white baseball cap photographed from the side, even if there is a high probability that the two results will be output simultaneously in the target image search result, the system still gives the user an opportunity to self-adjust the search result, the user finds out the image which is considered to be most suitable for the actual situation by himself among the search results, such as a black and white baseball cap shot from the side, then, iterative search is carried out, the system automatically sets the image 'a black and white interphase baseball cap shot at the side' as a normalized 'true target' image, and a new normalization and target image search process is restarted to ensure the accuracy and uniqueness of the result.
The purpose of the method is to enable the search process and the search results to be intelligent and self-learning, and return the images to the user from high to low in similarity, so as to realize a ranking target list of the static and dynamic e-commerce information which is valuable to the user.
Considering the particularity of image search, the user is given the opportunity of correcting the search request once, and then the target image search list is output to the user for confirmation results or the 'true target' of the image search is reset according to the requirement; for example, a user first submits a black and white baseball cap image search, one of the search results may be a black and white ball, and in order to enhance the self-learning ability of the system and prevent the next search result from having a highly biased result, the system gives the user an opportunity to select to set a normalized true target black and white baseball cap and to re-search the target image, so that the search result is closer to the user's expectation; when someone submits a black and white baseball cap image search for the second time in the future, the system can obtain the final target image search result list very quickly.
In order to improve the search result to be more humanized and reflect the real life situation, the concept of UGC of the internet user community is borrowed and adopted, the network search result is further dynamically adjusted, so that the search result can be more real, and the corresponding e-commerce service provider link is recommended to the user to help implement the e-commerce. For example, if a black and white baseball cap is searched for, the user can select to delete the result from the search target list, so that the content does not appear in the image search result of the next black and white baseball cap, the search speed is greatly increased, and the user experience is better. The search result adjusted by the user can reflect the requirements of the user more truly, and when a user in the fashion industry generally searches relevant images such as fashion and the like, namely unconventional content which can be described by a plurality of keywords, the user in the fashion industry can not meet most of users in the community even if a unified search result based on similarity exists.
As shown in fig. 2, the system structure for implementing the image search method based on the image and video content in the web page of the present invention includes a hierarchical structure of a presentation layer, a service layer and a data layer, wherein the data layer may be further divided into three major parts: service data: including data on merchants, products, channels, businesses, etc. Dynamic index data: establishing an index of the search image and the target image search list through a target image search process, wherein the index can be referred by future search; meanwhile, the target image search list can be modified by a user so as to achieve the purpose of self-learning of the system. Still picture data: the image can be classified into a real target image and a search image in terms of types; the viewing mode can be divided into two modes of an original image and a thumbnail image.
Compared with the prior art, the image searching method based on the image and video content in the webpage has the advantages that:
through the multiple-to-one normalization processing of the images and the one-to-multiple target image searching, the static ranking and dynamic ranking parts are integrated, and UGC concepts participated by users are introduced, so that the searching results can reflect the actual electronic commerce situation; the copyright problem of the image and the focus problem of the image can be well solved.
By newly introducing and increasing the characteristics such as block characteristics, contrast, vectors and the like in the whole image content and comprehensively using the characteristics according to the actual situation, an image can be better described, and a practical and feasible basis is laid for accurately identifying the image later.
The problem that image searching is not as accurate as plain text searching is solved by combining dynamic state and static state, and meanwhile, a chance for correcting self-submitted requests is given to users; and further introduces the concept of UCG construction of a user community to more objectively return images to users from high to low in similarity.
In summary, the method of the present invention includes how to help the user to search more similar contents by using the interested picture or video contents as a search request when browsing the web page contents; simultaneously, the content fed back to the user is organized according to the image similarity and the relevance provided by the user, and is displayed in a target link list form of an e-commerce manufacturer, wherein the content contains static ranking and dynamic ranking concepts; and the popularity of an e-commerce service provider is comprehensively considered, UGC technology, namely user self-help dynamic adjustment of ranking, is integrated by means of the concept of user community evaluation, and finally e-commerce is realized according to a CPC or CPS mode. Compared with the existing method, the method has better performance, intelligence and practicability.
The present invention is not limited to the embodiments disclosed in the above preferred embodiments, and all embodiments obtained by simple deduction and replacement based on the above design idea belong to the implementation of the present invention.
Claims (7)
1. An image searching method based on image and video content in a webpage is characterized by comprising the following steps:
firstly, receiving an image searching request;
secondly, preprocessing a search image;
thirdly, searching a normalized image;
fourthly, searching a target image;
if yes, setting a normalization true target, returning to the step two, and otherwise, continuing;
sixthly, evaluating the user community; and the number of the first and second groups,
and seventhly, searching a list of the final target images.
2. The image searching method of claim 1, wherein the step one comprises: and intercepting a target image which is contained in the embedded image of the webpage content and is really interested by the user by adopting a webpage embedding and browser plug-in mode.
3. The image searching method of claim 2, wherein the second step comprises: and analyzing and processing the target image content uploaded by the client by the background server, namely constructing and storing sample information of the target image according to the color, texture, shape, block shape, contrast and vector characteristics of the target image.
4. The image searching method of claim 3, wherein the third step comprises: and carrying out normalization processing on the sample information of the target image to obtain a normalized true target.
5. The image searching method of claim 4, wherein the fourth step comprises: and taking the normalized true target as a comparison object, searching a directional electronic commerce website by using a network robot, and establishing a target image search list index from high to low according to the similarity.
6. The image searching method of claim 5, wherein the step five comprises: the opportunity is provided for the user to make active modifications to the normalized true target.
7. The image searching method of claim 6, wherein the sixth step comprises: and dynamically adjusting the network search result according to the content created by the user.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009101087887A CN101957825A (en) | 2009-07-17 | 2009-07-17 | Method for searching image based on image and video content in webpage |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2009101087887A CN101957825A (en) | 2009-07-17 | 2009-07-17 | Method for searching image based on image and video content in webpage |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101957825A true CN101957825A (en) | 2011-01-26 |
Family
ID=43485158
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2009101087887A Pending CN101957825A (en) | 2009-07-17 | 2009-07-17 | Method for searching image based on image and video content in webpage |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101957825A (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103258012A (en) * | 2013-04-16 | 2013-08-21 | 广东欧珀移动通信有限公司 | Method and device for acquiring picture information |
| CN103955543A (en) * | 2014-05-20 | 2014-07-30 | 电子科技大学 | Multimode-based clothing image retrieval method |
| CN105657445A (en) * | 2015-12-30 | 2016-06-08 | Tcl海外电子（惠州）有限公司 | TV shopping system and implementing method thereof |
| CN107729547A (en) * | 2017-11-01 | 2018-02-23 | 上海掌门科技有限公司 | Retrieve the method and apparatus of circle of friends message |
| CN107844238A (en) * | 2017-11-29 | 2018-03-27 | 佛山市因诺威特科技有限公司 | A kind of method and system for counting browsing device net page information |
| CN107993125A (en) * | 2017-11-29 | 2018-05-04 | 重庆猪八戒网络有限公司 | Creative design transaction hatching system and method based on Multi-stage refined |
| CN110955369A (en) * | 2019-11-19 | 2020-04-03 | 广东智媒云图科技股份有限公司 | Focus judgment method, device and equipment based on click position and storage medium |
| CN113609319A (en) * | 2021-07-26 | 2021-11-05 | 阿里巴巴（中国）有限公司 | Commodity searching method, device and equipment |
| CN117036203A (en) * | 2023-10-08 | 2023-11-10 | 杭州黑岩网络科技有限公司 | Intelligent drawing method and system |
-
2009
- 2009-07-17 CN CN2009101087887A patent/CN101957825A/en active Pending
Cited By (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103258012A (en) * | 2013-04-16 | 2013-08-21 | 广东欧珀移动通信有限公司 | Method and device for acquiring picture information |
| CN103955543A (en) * | 2014-05-20 | 2014-07-30 | 电子科技大学 | Multimode-based clothing image retrieval method |
| CN105657445A (en) * | 2015-12-30 | 2016-06-08 | Tcl海外电子（惠州）有限公司 | TV shopping system and implementing method thereof |
| CN107729547A (en) * | 2017-11-01 | 2018-02-23 | 上海掌门科技有限公司 | Retrieve the method and apparatus of circle of friends message |
| CN107844238A (en) * | 2017-11-29 | 2018-03-27 | 佛山市因诺威特科技有限公司 | A kind of method and system for counting browsing device net page information |
| CN107993125A (en) * | 2017-11-29 | 2018-05-04 | 重庆猪八戒网络有限公司 | Creative design transaction hatching system and method based on Multi-stage refined |
| CN110955369A (en) * | 2019-11-19 | 2020-04-03 | 广东智媒云图科技股份有限公司 | Focus judgment method, device and equipment based on click position and storage medium |
| CN113609319A (en) * | 2021-07-26 | 2021-11-05 | 阿里巴巴（中国）有限公司 | Commodity searching method, device and equipment |
| CN117036203A (en) * | 2023-10-08 | 2023-11-10 | 杭州黑岩网络科技有限公司 | Intelligent drawing method and system |
| CN117036203B (en) * | 2023-10-08 | 2024-01-26 | 杭州黑岩网络科技有限公司 | Intelligent drawing method and system |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11899681B2 (en) | Knowledge graph building method, electronic apparatus and non-transitory computer readable storage medium | |
| US10180979B2 (en) | System and method for generating suggestions by a search engine in response to search queries | |
| CN101957825A (en) | Method for searching image based on image and video content in webpage | |
| JP5264511B2 (en) | Collaborative structured tagging for item encyclopedias | |
| CN103473273B (en) | Information search method, device and server | |
| US10210179B2 (en) | Dynamic feature weighting | |
| US10719836B2 (en) | Methods and systems for enhancing web content based on a web search query | |
| US20110072015A1 (en) | Tagging content with metadata pre-filtered by context | |
| US9177341B2 (en) | Determining search relevance from user feedback | |
| CN103064903A (en) | Method and device for searching images | |
| US20170287041A1 (en) | Information processing apparatus, information processing method, and information processing program | |
| CN110264283B (en) | Popularization resource display method and device | |
| CN111475725A (en) | Method, apparatus, device, and computer-readable storage medium for searching for content | |
| EP3485394A1 (en) | Contextual based image search results | |
| US9613283B2 (en) | System and method for using an image to provide search results | |
| KR20070105722A (en) | Automatic generation method of the initial query set for image search on the mobile web | |
| US8903817B1 (en) | Determining search relevance from user feedback | |
| CN112115354B (en) | Information processing method, device, server and storage medium | |
| KR20150045236A (en) | Method and apparatus for managing scrap information | |
| CN110827063A (en) | Multi-strategy fused commodity recommendation method, device, terminal and storage medium | |
| CN113495985A (en) | Article recommendation method and device, short video recommendation method and device and server | |
| CN114564653A (en) | Information recommendation method and device, server and storage medium | |
| CN107784061B (en) | Method and system for determining image-based content style and machine-readable medium | |
| CN114662010B (en) | Explicit and implicit collaborative filtering recommendation method based on multitask learning | |
| TW201901493A (en) | Data search method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20110126 |