CN102262638A - Method and device for indexing picture - Google Patents
Method and device for indexing picture Download PDFInfo
- Publication number
- CN102262638A CN102262638A CN2010101925270A CN201010192527A CN102262638A CN 102262638 A CN102262638 A CN 102262638A CN 2010101925270 A CN2010101925270 A CN 2010101925270A CN 201010192527 A CN201010192527 A CN 201010192527A CN 102262638 A CN102262638 A CN 102262638A
- Authority
- CN
- China
- Prior art keywords
- picture
- index
- family
- partner
- eigenwert
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Abstract
The embodiment of the invention discloses a method for indexing a picture. The method comprises the following steps of: obtaining a picture to be indexed; judging whether a characteristic value of the picture to be indexed is similar to the characteristic value of at least one indexed picture group, wherein the indexed picture group corresponds to at least one tag; and under the condition that the characteristic values of the picture to be indexed and the indexed picture group are similar, recommending the tag which corresponds to the indexed picture group to a user. According to the embodiment of the invention, more automation can be provided for the user when the user indexes the picture.
Description
Technical field
The present invention relates to Internet technology.More particularly, the method and apparatus that relates to the picture index.
Background technology
The picture index refers to other information beyond the additional image information of picture, and Word message for example is with the label as described picture.Now, along with popularizing of digital camera and internet, a large amount of pictures are taken, and share by the internet.After picture is by index, can retrieves faster and betterly and analyze picture.Therefore the picture index is just becoming the critical function of desktop or online pictures management instrument.
Simple picture index function realizes by manual index interface is provided to the user.The user associates picture and at least one label by described manual index interface.For some special application, may need the specific part in the picture is carried out index.For example, be used for social application need the part that picture contains face's object is carried out index, the application need that is used for automobile carries out index to the part that picture contains automotive subjects, is used for astronomical application need the part that picture contains the celestial body object is carried out index or the like.At these special applications, can further provide manual selection tool so that the user chooses described specific part as a new picture to the user, utilize described manual index interface that described new picture is carried out index then.This artificial mode fully is applied in current desktop picture management software, online pictures management website or the community network website very at large.
A kind of improvement as to this pure manual mode can provide automatic selection tool.For example can utilize face detection automatically extraction to be comprised the part of face's object as described new picture.In addition, can also provide to a certain degree automatic indexing.For example, be generally used for the occasion of community network application at the index of face's object, thereby can recommend the user to user's contact person as candidate's label.However, still there is the demand that the robotization of higher degree is provided to the user.
Summary of the invention
In view of this, the embodiment of the invention provides the method and apparatus of picture index, thereby the robotization of higher degree is provided to the user in picture index process.
The method of carrying out the picture index according to the embodiment of the invention comprises: obtain to treat the index picture; Whether the eigenwert of index picture family is similar at least one to judge the described eigenwert for the treatment of the index picture, and the wherein said picture of index family is corresponding at least one label; Treat under the index picture situation similar described, the label of the described picture of index family correspondence is recommended the user to the eigenwert of the described picture of index family.
The device that carries out the picture index according to the embodiment of the invention comprises: obtain module, be configured to obtain to treat the index picture; Judge module, whether the eigenwert of index picture family is similar at least one to be configured to judge the described eigenwert for the treatment of the index picture, and the wherein said picture of index family is corresponding at least one label; Recommending module is configured to treat under the index picture situation similar to the eigenwert of the described picture of index family described, and the label of the described picture of index family correspondence is recommended the user.
Employing can automatically will be recommended the user with the label of the picture of the index family that treats the index picture analogies, thereby realize automatic indexing to a certain extent according to the technical scheme of the embodiment of the invention.Particularly,, also comprise user's partner's the picture of index family, just can strengthen the chance that finds and treat the picture of the index family of index picture analogies if the described picture of index family not only comprises user's oneself the picture of index family.
Description of drawings
Fig. 1 is the process flow diagram according to the method for carrying out the picture index of the embodiment of the invention.
Fig. 2 is the process flow diagram according to the method for carrying out the picture index of the embodiment of the invention.
Fig. 3 is the block diagram according to the device that carries out the picture index of the embodiment of the invention.
Embodiment
Embodiments of the invention are described with reference to the accompanying drawings.In the following description, many details have been set forth so that more fully understand the present invention.But those skilled in the art of the present technique understand easily, and realization of the present invention can not have some in these details, and the present invention is not limited to the specific embodiment introduced.On the contrary, can consider to implement the present invention, and no matter whether they relate to different embodiment with the combination in any of following feature and key element.Therefore, following aspect, feature, embodiment and advantage for illustrative purposes only usefulness and should not be counted as the key element or the qualification of claims, unless clearly propose in the claim.What also need to illustrate a bit is, for fear of having blured the present invention because of unnecessary details, only show in the accompanying drawings with according to closely-related apparatus structure of the solution of the present invention and/or treatment step, and omitted other details little with relation of the present invention.In addition, unless painstakingly use " directly " or " indirectly " to be limited, the connection in the present specification had both comprised direct connection, also comprised indirectly connecting.
For the robotization of higher degree is provided to the user, can introduce used label as candidate's label of recommending to the user in the process of picture index.In some following embodiment, adopt the picture that contains face to describe as example.As previously mentioned, the picture that contains face can be the part of another picture.Those skilled in the art can understand, and can be used for the occasion of any picture index according to the technical scheme of the embodiment of the invention.
With reference to Fig. 1, Fig. 1 is the process flow diagram according to the method for carrying out the picture index of the embodiment of the invention.
Can be by the method for carrying out picture index of index application implementation according to the embodiment of the invention.In certain embodiments, index is used and is run on the server.Described server provides to the client that connects by network can be by the document that runs on the browser resolves on the client, for example html document.The described document of described browser resolves, thus user interface generated, so that the user can specify the picture that needs index by described user interface.After the user indicated appointment to finish by any way, described client computer picture to be named sent to server.Like this, running on index on the server uses and just can obtain to treat the index picture by the network transmission.
In further embodiments, described index application is a desktop application.This is used and directly generates user interface, so that the user can specify the picture that needs index by described user interface.After the user indicated appointment to finish by any way, described index was used and is loaded the described index picture for the treatment of.
One typically comprises several pictures of index corresponding to same label in the index picture family.The picture of index that it will be understood by those skilled in the art that single width also can be regarded as and comprises the width of cloth picture of the index family of index picture, and this moment, the eigenwert of index picture family was exactly the described width of cloth eigenwert of index picture.For example, user's index one or more contain the picture of the face of John, one or more contain the picture of the face of Mike, one or more contain the picture of the face of Kate.These pictures that contain the face of John are formed an index picture family, and the picture that contains the face of Mike is formed another index picture family, and the picture that contains the face of Kate is formed another picture family.
Those skilled in the art can design the characteristic that various eigenwerts characterize picture, thus by more described eigenwert whether near judging whether two width of cloth pictures similar.According to one embodiment of the invention, the information that automatically produces in the time of can utilizing picture shooting, shooting time for example is as described eigenwert.According to a further embodiment of the invention, can set up the facial model of band parameter, calculate described model parameter respectively at the face's object in each picture then.If the distance of the model parameter vectors of the face's object in two width of cloth pictures, thinks then that the eigenwert of this two width of cloth picture is similar less than certain distance threshold.Another embodiment according to the present invention also can adopt the eigenwert as described picture such as color, profile, transform characteristics.It will be understood by those skilled in the art that above-mentioned various eigenwert can combine use.
In actual applications, with treat index picture and single width the index picture relatively may have bigger error, especially with image content as the situation of determining similarity under.This be because, some enchancement factor, for example the shooting condition of picture may cause the variation of characteristic of correspondence value.When the described picture of index family comprises several during the index picture, several that have comprised in the index picture family according to this index picture calculate the eigenwert of the described picture of index family.Owing to comprised several index pictures in the index picture family corresponding to same label, just comprise several index pictures corresponding to same target, therefore according to described several the eigenwert of the picture of the index family that calculates of index picture can remove described stochastic factor preferably, thereby improve and treat index picture accuracy relatively.
Suppose, one the picture of index in the index picture family all take and have label " Paris " owing to it in Paris, and treat index picture and these the shooting time of index picture differed five minutes, treat that so the index picture may also be to take in Paris, thus the user may with treat the index picture also index be " Paris ".Again for example, if a width of cloth the index picture because wherein personage is John and all have label " John ", and treat that the model parameter that face had in the index picture satisfies certain approximate condition with the model parameter that face had in the index picture, treat that so the personage in the index picture may also be John, thus the user may with treat the index picture also index for " John ".Again for example, comprise promptly that in eigenwert shooting time comprises the occasion of facial model parameter again, if angle from model parameter, treat index picture and two index picture family all satisfy described approximate condition, can further relatively treat index picture and these two the close degree of shooting time of index picture family.
According to reason described above, each picture of index that has comprised in the index picture family is many more, and treats that index picture accuracy relatively is just high more.In general, the index picture is user's self maintained, promptly only will treat index picture and user index picture family comparison in step 102.This may be unfavorable for forming index picture family, and the picture of the index family that perhaps causes forming comprises the fewer picture of index.Still be example with index at face's object.In the above-described embodiments, index picture family is not similar if treat index picture and any one user, and the user still needs to utilize manual index instrument that this picture is carried out index so.The user need carry out repeatedly manually index could form index picture family.When the user had begun that just picture carried out index, still needed user to carry out manual index owing to there is not user index picture family this moment.
Below with reference to the further improvement to step 102 of Fig. 2 description according to the embodiment of the invention.
Use for online index, for example the index that is provided by the community network website is used, and can directly obtain the set partner's list of user.Use for the desktop index, can from user's e-mail applications, obtain the partner of contact person as the user.Those skilled in the art can also implement the mode of any user of obtaining partner list.
Here suppose that described partner has also carried out indexing work to its picture of managing.According to one embodiment of present invention, can obtain partner's the picture of index itself and its pairing label, then each that calculates the partner from partner's the picture of index eigenwert of index picture family.If there are several index pictures corresponding to same label, then by these the index picture calculate the eigenwert of index picture family.As previously mentioned, the picture of index of single width also can be regarded as and comprises the width of cloth picture of the index family of index picture, and this moment, the eigenwert of index picture family was exactly the described width of cloth eigenwert of index picture.
According to a further embodiment of the invention, can directly obtain eigenwert itself and corresponding label, and not obtain partner's pairing picture of index picture family.This helps carrying out privacy control preferably.The partner may not wish the picture of its management itself is shared.Directly obtain eigenwert and corresponding label, both avoided image sharing itself, there is abundant information can be used for user's the index picture for the treatment of is carried out index again, because when treating that relatively index picture and index picture family be whether similar, be directly treat the eigenwert of index picture and the eigenwert of index picture family compare.
A kind of situation is, the user is the label that do not overlap of the pairing label of index picture family and partner's pairing label of index picture family.For example, the user has four index picture families, difference corresponding label " myself ", " John ", " Mike " and " Kate ", and the partner has two index picture families, respectively corresponding label " Jack " and " Mary ".Like this, the partner of the partner's of the described eigenwert for the treatment of the index picture and corresponding label " Jack " index picture family eigenwert and corresponding label " Mary " eigenwert of index picture family can be compared.
Another kind of situation is that certain user pairing label of index picture family has the label that overlaps with partner's pairing label of index picture family.For example, the user has four index picture families, difference corresponding label " myself ", " John ", " Mike " and " Kate ", and the partner has three index picture families, respectively corresponding label " Kate ", " Jack " and " Mary ".Like this, need certain means that the partner of user of corresponding label " Kate " index picture family and corresponding label " Kate " index picture family is merged, and then calculate the eigenwert of the figure of the index family group after this merging.Those skilled in the art can design a lot of methods, the eigenwert of the new index feature family that comes these two families of index feature of eigenvalue calculation according to two groups of index feature to merge to obtain.For example, can be weighted the eigenwert of summation to the eigenwert of index picture family according to the number of pictures that comprises in the index picture family with the family of index feature calculating described merging and obtain.Under the situation of the picture itself of index that has obtained the partner and its pairing label, can calculate the eigenwert that merges the family of index feature that obtains according to the picture itself in the picture itself in user's index feature family and the partner's index feature family.Under more than one situation of user's partner, may have the three or more picture of index families and have identical label.Still can use said method to handle to have the three or more picture of the index family of identical label and merge.
According to such method, can come to be user's recommended candidate label with reference to the indexing work that the partner had carried out.For example, the picture that contains Jack is crossed in suppose partner index, even has formed bigger partner corresponding to label " Jack " index picture family.When user's appointment treat that the index picture comprises Jack the time, even the picture that contains Jack is crossed in index never before the user, if system determines that this treats that the index picture to similar corresponding to the partner's of label " Jack " index picture family eigenwert, just can recommend the user with label " Jack ".
The present inventor is through discovering, in user's the picture of index, the picture that contains the user is maximum, thereby the picture of the index family with label " myself " or similar label may comprise the more picture of index, and the picture of the index family with label " John ", " Mike " or " Kate " may comprise the less picture of index, even includes only width of cloth index picture.Under this research conclusion, the indexing work that had carried out with reference to the partner come for user's recommended candidate label particularly useful.For the user,, for the picture of the people's face that comprises these partners, can be recommended label more accurately at least through step 201 to 203.For example, suppose that the user is Lisa to a partner's address, and one of this partner the label of index picture family correspondence be " myself ".When user's appointment treat that the index picture comprises Lisa the time, even the picture that contains Lisa is crossed in index never before the user, if system determines that this eigenwert corresponding to the picture of the index family of label " myself " for the treatment of index picture and this partner is similar, just label " Lisa " can be recommended the user.
The user determined a label as actual with treat label that the index picture is associated after, alternatively, can upgrade the eigenwert of described index picture family.If the determined label of user is not selected from described recommendation label, newly to set up a user for this picture so and treat index picture family, the eigenwert of this picture is exactly the eigenwert that this newly-established user treats index picture family.If the determined label of user is to select from the label of being recommended, an index picture family must be arranged so, or be user index picture family, or be partner index picture family, corresponding with determined label.Can treat that the existing picture in index picture and the described picture of the index family revises the eigenwert of the described picture of index family according to this.The method that can also merge the eigenwert of the new index feature family that obtains with reference to these two families of index feature of above-mentioned eigenvalue calculation according to two groups of index feature.It is pointed out that the user and picture carried out in the process of index that described partner index picture family should not be modified.In other words, should obtain partner pairing eigenwert of index picture family and label in read-only mode.So just can meet the requirement of privacy control.
In step 201, can filter partner's list according to certain standard.According to one embodiment of present invention, can select with the user between exchange number of times greater than the selected described partner's list of the partner who exchanges frequency threshold value.The interchange here comprises and mutual between user and the partner also comprises user capture partner's the content shared or the content shared of partner's calling party etc.Described interchange number of times can be the user with the partner between always exchange number of times, also can be in a period of time recently with the user between the recent number of times that exchanges.According to still a further embodiment, can select among user's the partner, in a period of time recently, carry out the selected described partner's list of partner that picture library upgrades.According to still another embodiment of the invention, can select among user's the partner the selected described partner's list of partner that number of pictures is more.
It is generally acknowledged that partner's list is filtered to be needed to calculate the condition of filtering institute's reference, for example exchanges number of times, and uses filtering rule, therefore can need more storage resources, computational resource etc.But the present inventor finds through experiment, partner's list is filtered the speed that can accelerate to handle on the whole.Owing to need dynamically obtain partner pairing eigenwert of index picture family and label according to partner's list, therefore need consumption of network resources and computational resource to transmit and handle label, parameter value even the picture of picture itself.But index picture family has bigger help to form for the user to have only some partners' picture library.For example, if the total interchange number of times between user and certain partner or to exchange number of times in the recent period bigger, in close relations between user and this partner is described, very possible user and this partner can be identical to identical object index label, especially when described when liking the object of user and this partner's common understanding.Again for example, after once index picture family had carried out the picture index with reference to certain partner's partner on the user, the user had formed corresponding to these partners' index picture family user index picture family.If this partner did not carry out the renewal of picture after this, can provide the possibility of more helps just smaller for the user so.Again for example, if partner's number of pictures is fewer, so the partner who forms from these pictures the number of pictures that comprises of index picture family is also fewer, thereby be difficult to carry out exactly eigenwert relatively and the label recommendation.
According to one embodiment of present invention, the user can be provided with partner's degree of depth, thus the indexing work that with reference to other people picture is carried out in the larger context.Partner's list that described partner's degree of depth indicates described partner to know not to be arranged in the user.For example,, then obtaining the partner when the pairing eigenwert of index picture family and label, only considering pairing eigenwert of partner and label on partner's list of user if partner's degree of depth is 1; If partner's degree of depth is 2, then further consider pairing eigenwert of partner and label on partner's list of described partner, by that analogy.
Fig. 3 is that this device comprises according to the block scheme of the system that is used for the picture index of the embodiment of the invention:
Obtain module, be configured to obtain to treat the index picture;
Judge module, whether the eigenwert of index picture family is similar at least one to be configured to judge the described eigenwert for the treatment of the index picture, and the wherein said picture of index family is corresponding at least one label;
Recommending module is configured to treat under the index picture situation similar to the eigenwert of the described picture of index family described, and the label of the described picture of index family correspondence is recommended the user.
Those having ordinary skill in the art will appreciate that can use a computer executable instruction and/or be included in the processor control routine of above-mentioned method and system realizes, for example on such as the mounting medium of disk, CD or DVD-ROM, such as the programmable memory of ROM (read-only memory) (firmware) or data carrier, provide such code such as optics or electronic signal carrier.Being used to of present embodiment handled the device of picture index and assembly thereof can be by such as VLSI (very large scale integrated circuit) or gate array, realize such as the semiconductor of logic chip, transistor etc. or such as the hardware circuit of the programmable hardware device of field programmable gate array, programmable logic device etc., also can use the software of carrying out by various types of processors to realize, also can by the combination of above-mentioned hardware circuit and software for example firmware realize.
Though illustrated and described some exemplary embodiments of the present invention, but it will be understood by those skilled in the art that, under the prerequisite that does not depart from principle of the present invention and spirit, can change these embodiments, scope of the present invention is limited by claims and equivalence transformation thereof.
Claims (12)
1. method of carrying out the picture index comprises:
The index picture is treated in acquisition;
Whether the eigenwert of index picture family is similar at least one to judge the described eigenwert for the treatment of the index picture, and the wherein said picture of index family is corresponding at least one label;
Treat under the index picture situation similar described, the label of the described picture of index family correspondence is recommended the user to the eigenwert of the described picture of index family.
2. method according to claim 1, wherein judge the described eigenwert for the treatment of the index picture and at least one eigenwert similar comprising whether of index picture family:
Obtain partner's list of user;
According to described partner's list, obtain partner pairing eigenwert of index picture family and label; And
Judge and describedly treat whether the eigenwert of index picture is similar at least one partner's index picture family eigenwert.
3. method according to claim 2 further comprises: described partner's list is filtered.
4. method according to claim 3, wherein one of at least described partner's list is filtered according to following:
With exchanging number of times, upgrade the number of times of picture, the number of pictures that the partner has in the recent period between the user.
5. method according to claim 2, wherein obtain the partner the pairing eigenwert of index picture family and label comprise:
Directly obtain partner pairing eigenwert of index picture family itself and label.
6. method according to claim 2, wherein obtain the partner the pairing eigenwert of index picture family and label comprise:
Obtain partner pairing eigenwert of index picture family and label in read-only mode.
7. device that carries out the picture index comprises:
Obtain module, be configured to obtain to treat the index picture;
Judge module, whether the eigenwert of index picture family is similar at least one to be configured to judge the described eigenwert for the treatment of the index picture, and the wherein said picture of index family is corresponding at least one label;
Recommending module is configured to treat under the index picture situation similar to the eigenwert of the described picture of index family described, and the label of the described picture of index family correspondence is recommended the user.
8. device according to claim 7, wherein said judge module comprises:
Be configured to obtain the device of partner's list of user;
Be configured to according to described partner's list, obtain partner's device of the pairing eigenwert of index picture family and label; And
Be configured to judge the described eigenwert for the treatment of the index picture and at least one partner's index picture family the whether similar device of eigenwert.
9. device according to claim 8 further comprises:
Be configured to device that described partner's list is filtered.
10. device according to claim 9, wherein filter described partner's list one of at least according to following:
With exchanging number of times, upgrade the number of times of picture, the number of pictures that the partner has in the recent period between the user.
11. device according to claim 8, wherein said be configured to obtain the partner the device of the pairing eigenwert of index picture family and label comprise:
Be configured to directly to obtain partner's device of the pairing eigenwert of index picture family itself and label.
12. device according to claim 8, wherein said be configured to obtain the partner the device of the pairing eigenwert of index picture family and label comprise:
Be configured to obtain partner's device of the pairing eigenwert of index picture family and label in read-only mode.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101925270A CN102262638A (en) | 2010-05-31 | 2010-05-31 | Method and device for indexing picture |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101925270A CN102262638A (en) | 2010-05-31 | 2010-05-31 | Method and device for indexing picture |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN102262638A true CN102262638A (en) | 2011-11-30 |
Family
ID=45009268
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010101925270A Pending CN102262638A (en) | 2010-05-31 | 2010-05-31 | Method and device for indexing picture |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102262638A (en) |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103530712A (en) * | 2012-07-05 | 2014-01-22 | 鸿富锦精密工业（深圳）有限公司 | Image sample establishing system and image sample establishing method |
| CN104239457A (en) * | 2014-09-02 | 2014-12-24 | 百度在线网络技术（北京）有限公司 | Search recommending method and device |
| CN104572735A (en) * | 2013-10-23 | 2015-04-29 | 华为技术有限公司 | Photo tagging word recommending method and device |
| CN106469199A (en) * | 2016-08-31 | 2017-03-01 | 福建天晴数码有限公司 | Method and system recommended by a kind of interactive mode picture |
| CN106815351A (en) * | 2017-01-18 | 2017-06-09 | 腾讯科技（深圳）有限公司 | A kind of label recommendation method and server |
| CN108287919A (en) * | 2018-02-13 | 2018-07-17 | 广东欧珀移动通信有限公司 | Access method, device, storage medium and the electronic equipment of web application |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101286174A (en) * | 2008-05-09 | 2008-10-15 | 北京大学 | File label intelligent prompt method |
| US20090171783A1 (en) * | 2008-01-02 | 2009-07-02 | Raju Ruta S | Method and system for managing digital photos |
| CN101510205A (en) * | 2009-03-16 | 2009-08-19 | 广州市有福数码科技有限公司 | Method for implementing automatically clustering photographs, apparatus and system |
-
2010
- 2010-05-31 CN CN2010101925270A patent/CN102262638A/en active Pending
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20090171783A1 (en) * | 2008-01-02 | 2009-07-02 | Raju Ruta S | Method and system for managing digital photos |
| CN101286174A (en) * | 2008-05-09 | 2008-10-15 | 北京大学 | File label intelligent prompt method |
| CN101510205A (en) * | 2009-03-16 | 2009-08-19 | 广州市有福数码科技有限公司 | Method for implementing automatically clustering photographs, apparatus and system |
Cited By (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103530712A (en) * | 2012-07-05 | 2014-01-22 | 鸿富锦精密工业（深圳）有限公司 | Image sample establishing system and image sample establishing method |
| CN103530712B (en) * | 2012-07-05 | 2016-12-21 | 鸿富锦精密工业（深圳）有限公司 | picture sample establishing system and method |
| CN104572735A (en) * | 2013-10-23 | 2015-04-29 | 华为技术有限公司 | Photo tagging word recommending method and device |
| CN104572735B (en) * | 2013-10-23 | 2018-02-23 | 华为技术有限公司 | A kind of picture mark words recommending method and device |
| CN104239457A (en) * | 2014-09-02 | 2014-12-24 | 百度在线网络技术（北京）有限公司 | Search recommending method and device |
| CN104239457B (en) * | 2014-09-02 | 2019-03-12 | 百度在线网络技术（北京）有限公司 | Search for recommended method and device |
| CN106469199A (en) * | 2016-08-31 | 2017-03-01 | 福建天晴数码有限公司 | Method and system recommended by a kind of interactive mode picture |
| CN106469199B (en) * | 2016-08-31 | 2019-04-30 | 福建天晴数码有限公司 | A kind of interactive mode picture recommendation method and system |
| CN106815351A (en) * | 2017-01-18 | 2017-06-09 | 腾讯科技（深圳）有限公司 | A kind of label recommendation method and server |
| CN108287919A (en) * | 2018-02-13 | 2018-07-17 | 广东欧珀移动通信有限公司 | Access method, device, storage medium and the electronic equipment of web application |
| CN108287919B (en) * | 2018-02-13 | 2020-05-12 | Oppo广东移动通信有限公司 | Webpage application access method and device, storage medium and electronic equipment |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US10552542B1 (en) | Techniques for graph based natural language processing | |
| CN107430531B (en) | Method and system for managing permissions to access mobile device resources | |
| CN102262638A (en) | Method and device for indexing picture | |
| US20100132049A1 (en) | Leveraging a social graph from a social network for social context in other systems | |
| CN109508879B (en) | Risk identification method, device and equipment | |
| US10163172B2 (en) | Calculating lists of events in activity streams | |
| EP2481005A2 (en) | Shared face training data | |
| TW202001743A (en) | Credit-guarantee-based service processing method, apparatus and device | |
| CN109993026B (en) | Training method and device for relative recognition network model | |
| Correa et al. | Whacky!-what anyone could know about you from twitter | |
| EP2635981A1 (en) | Image search | |
| CN108932525B (en) | Behavior prediction method and device | |
| Caminha et al. | A smart middleware to perform semantic discovery and trust evaluation for the Internet of Things | |
| CN111723280B (en) | Information processing method and device, storage medium and electronic equipment | |
| CN104572649B (en) | The processing method of the data of distributed memory system, apparatus and system | |
| CN111275445B (en) | Data processing method, device and equipment | |
| Fan et al. | Model-Heterogeneous Federated Learning for Internet of Things: Enabling Technologies and Future Directions | |
| CN103391239A (en) | Method and device for chatting in SNS (social networking services) | |
| CN113902303B (en) | Privacy model automatic recommendation system, algorithm, equipment and storage medium based on user satisfaction | |
| CN113343090B (en) | Method, apparatus, device, medium and product for pushing information | |
| CN111522993B (en) | Tag validity management method, device, network equipment and storage medium | |
| CN111930620B (en) | Application running environment data processing method and device | |
| CN112214387B (en) | Knowledge graph-based user operation behavior prediction method and device | |
| Zhang et al. | Coarse to fine: Weak feature boosting network for salient object detection | |
| US20200195602A1 (en) | Online system and method for receiving feedback regarding apparel options via an online community |
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: 20111130 |