CN102591868A - System and method for automatic generation of photograph guide - Google Patents
System and method for automatic generation of photograph guide Download PDFInfo
- Publication number
- CN102591868A CN102591868A CN2011100038132A CN201110003813A CN102591868A CN 102591868 A CN102591868 A CN 102591868A CN 2011100038132 A CN2011100038132 A CN 2011100038132A CN 201110003813 A CN201110003813 A CN 201110003813A CN 102591868 A CN102591868 A CN 102591868A
- Authority
- CN
- China
- Prior art keywords
- photo
- user
- guide
- request
- representative
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Abstract
The invention provides a system and a method for automatic generation of a photograph guide. The method includes: determining requirements of users according to environment and positions where the users are located and/or input of the users; obtaining relevant photos from a photo data bank according to the requirements of the users; selecting representative photos from the obtained relevant photos according to visual characteristics and context information of the photos; extracting shooting parameters of the representative photos, such as camera factories, camera models, exposure value, sensitivity, aperture sizes, focal distance, shutter speed and shooting data and time; and generating shooting guide to be presented to the users according to the representative photos and the shooting parameters of the representative photos so as to be favorable for the users to shoot professional photography works.
Description
Technical field
The present invention relates to be used to take pictures the guide system and method for generation automatically.
Background technology
Along with the development of electronic product and information storage technology, digital camera/video camera is widely used by domestic consumer by feat of practical, convenient, cheap and friendly operation interface.In order to make the especially amateur photography of users user take outstanding shooting works, each camera businessman also releases built-in separately parameter software is set automatically in the propaganda camera, and for example the auto-closing or the like of turning off the light is dodged in focusing automatically.These built-in functions provide operation interface more easily for the user.The present invention has also provided the system and method that a kind of user of being generates the guide of taking pictures automatically, comprises the automatic recommendation of taking visual angle and various parameters, greatly improves user experience.
Investigation through to the correlation technique document finds that existing systems or method can only be controlled a parameter.For example, patented claim No.US6195509B1 has invented a device of controlling exposure through the detection back-lighting.In order to detect back-lighting; This invention has comprised a focus detection and a photometering equipment; The focus detection distance of coming estimating target wherein through receiving target light, and photometering equipment comes assessment objective brightness through receiving target light in a plurality of photometerings zone.Yet well-known, outstanding photographic work need be according to the characteristics of photographic subjects, and combine various acquisition parameters and take the visual angle and could accomplish, and therefore, the method for the single acquisition parameters of existing control is difficult to obtain gratifying effect.
As another important step of the present invention, cluster is that one group of observed value is assigned in the different subsets (class), makes that the observed value in the same class is more similar.Kmean clustering algorithm and use very extensive now based on the clustering algorithm of Kmean; For example Kmean enhancement algorithms [patented claim No.US7590642B2] is conceived to " putting down " data, and just all observed values all use the proper vector of a regular length to represent.Yet the most of data in the reality all have the structure of more complicated, have comprised the characteristic of multiple modalities, are difficult to express with the proper vector of same length.For example, the network picture not only has visual signature, and usually also has the information of photo description, mark, spectators' comment, number of visits, comment number of times or the like each side; Document files also is not only a literal, also comprises author's introduction, issuing date and place, text categories, organizational form or the like.Under this situation, traditional if we utilize such as the Kmean clustering method, just need be with the characteristic of different modalities respectively at separately space clustering, and then adopt some extra rules that the result of a plurality of space clusterings is made up.This method has been ignored each mode internal relation each other, and this relation is one of important information for cluster.For example visual signature and textual description are to complement each other in the network picture, complement each other.
Summary of the invention
Because the problems referred to above; The invention provides a kind of system and method that is used for generating automatically the guide of taking pictures; It can recommend to take visual angle and acquisition parameters to the user according to user's different scenes and demand, and generates the shooting guide; Thereby help the user to shoot the photographic work of specialty, strengthen user experience.
According to an aspect of the present invention; The method that when user's request comprises or do not comprise geography information, provides a kind of guide of taking pictures to generate automatically; Comprise: confirm user's demand, according to user's different demands, system will obtain relevant photo from the photo data storehouse of outside.And relevant number of pictures is often huge, if the information that all photos is relevant with photo is presented to the user, user plane is difficult to get access to useful information on the contrary to the information of numerous and complicated.Thereby from these relevant photos, select representative typical photo significant for the present invention; Here will take a kind of clustering algorithm of extensive multi-modal data to address this problem; Cluster centre is named a person for a particular job and is counted as representational photo, the prominent example of just taking.Extract the acquisition parameters that representative photo is associated at last, present to the user with a kind of suitable mode in conjunction with described shooting example.
In addition; In said method, said geography information comprise based on the GPS GPS judge the residing geographical longitude and latitude of photographing device (cell phone that the GPS function for example arranged with camera etc.), the shooting address imported through input equipment based on the user or the text description of geographic name.
In addition, in said method, the said user's request that does not comprise geography information comprises based on the captured photo example that can describe the user's request scene of photographing device; Text description based on the photographic subjects of user input.
In addition, in said method, the photo data storehouse of said outside comprises the network image that obtains based on photo sharing website or online photographic search engine or based on the collection of photographs of the shared collection of private photograph album.
In addition; In said method; Said comparison film set is carried out multi-modal cluster and comprised: the visual signature that photo is set is a master mode; Text message and various contextual information (including but not limited to photo description, mark, spectators' comment, number of visits, comment number of times etc.) are auxilliary mode, and the relation between definite different modalities; Based on the auxilliary mode characteristic of photo, type of selection center, exploratory ground; In conjunction with the major-minor mode characteristic of photo, merge the high class center of similarity degree; Based on the similarity at each type center, grade subordinate relation, symbiosis information etc., will remain photo and be assigned in each type; Based on local density's height of master mode, adjust the class center of each type; Distribute a residue photo and adjustment type two processes in center to repeat, till satisfying the cluster condition; And obtain final cluster result, and the corresponding photo in type of returning center.
In addition, in said method, said acquisition parameters includes but not limited to camera manufacturer, camera model, exposure value, light sensitivity, aperture size, focal length, shutter speed, shooting date and time.
According to an aspect of the present invention; The system that when user's request comprises or do not comprise geography information, provides a kind of guide of taking pictures to generate automatically; Comprise: the demand receiver, be configured to input based on the residing environment of user, position and/or user, confirm user's request; The photo data storehouse; Be configured to store the database of photo data; The network image that for example obtains based on photo sharing website or image search engine perhaps based on the collection of photographs of the shared collection of private photograph album, is used for therefrom selecting high-quality photos with the take pictures resource of guide of generation; The guide maker of taking pictures is configured to based on described user's request, extracts the visual signature and the contextual information of photo, from the photo data storehouse, obtains representative photo and acquisition parameters thereof; And take the guide show stand, be configured to show high-quality photos and each item acquisition parameters thereof to the user.
In addition, in said system, the said guide maker of taking pictures comprises: the information retrieval device, be configured to based on described user's request, and from the photo data storehouse, obtain relevant photo; And the representative photo selector switch, be configured to visual signature and contextual information according to photo, from the said photo that obtains, select representative photo.
In addition, in said system, said information retrieval device comprises: the demand analysis device is configured to confirm the kind of user's request and it is carried out corresponding pre-service; Feature extractor, the every secondary photo that is configured to from the photo data storehouse extracts text feature and visual signature; The measuring similarity device, the degree of correlation that is configured to measure the every secondary photo in query contents and the photo data storehouse; And sorting unit, being configured to degree of correlation based on every secondary photo, all photos in the comparison film data bank sort.
In addition, in said system, said representative photo selector switch comprises: context information collector, be configured to collect and the related information of photo, and include but not limited to photo description, mark, spectators' comment, number of visits, comment number of times or the like; The clustering processing device is configured to contextual information and visual signature based on photo, the set of comparison film carry out multi-modal cluster; And type center extraction apparatus, be configured to based on cluster result the photo that the compute classes center is corresponding.
Description of drawings
Fig. 1 illustrates the exemplary configuration figure that does not have the photograph taking of geography information guide automatic creation system;
Fig. 2 is the exemplary configuration figure that the photograph taking guide automatic creation system of geography information is shown;
Fig. 3 is the block diagram that the exemplary configuration of photograph taking guide automatic creation system is shown;
Fig. 4 is the exemplary plot of the automatic shooting guide that generates;
Fig. 5 is the block diagram that is shown in further detail the exemplary configuration of photograph taking guide automatic creation system;
Fig. 6 is the process flow diagram that the photograph taking guide under the exemplary situation that is illustrated in geography information generates method automatically;
Fig. 7 is the process flow diagram that the exemplary photograph taking guide that is illustrated under the situation that does not have geography information generates method automatically;
Fig. 8 is the exemplary process flow diagram that the clustering algorithm of extensive multi-modal data is shown.
Embodiment
Below will be described in detail with reference to the attached drawings embodiments of the invention.
Fig. 1 illustrates the exemplary configuration figure that does not have the photograph taking of geography information guide automatic creation system.As shown in Figure 1; A fixed equipment 101, for example a personal computer comprises a central processor CPU 1011, network service 1012, input equipment for example keyboard 1013 and display device 1014; Wherein network service 1012 is by input equipment 1013; Be responsible for receiving user's request through network 102 from data server 103, display device 1014 is responsible for showing shooting guide 303, i.e. excellent photograph example and relevant acquisition parameters to the user.
Fig. 2 is the exemplary configuration figure that the photograph taking guide automatic creation system of geography information is shown.As shown in Figure 2; The mobile radio communication apparatus 201 that the global position system GPS function is arranged; For example a digital camera or a cell phone comprise that a microprocessor 2011 is used for carrying out software, the digital camera 2012 of taking guide and is used for taking digital photos or video, GPS positioning system 2013 based on confirming that from the signal of gps satellite system 202 current location of mobile device 201, network service 2014 receive data by the wireless network that has base station 203 204.Wherein wireless network 204 can connect with the data network 103 such as Internet.Identical with equipment 101, equipment 201 equally also comprises a display device 2015 to be responsible for showing the shooting guide to the user.
Below will describe the exemplary configuration flow process of photograph taking guide automatic creation system 300 with reference to Fig. 3 in detail.
Photograph taking guide automatic creation system 300 comprises a message recipient 301, guide maker 302, takes guide 303 and photo data storehouse 304.
Taking guide maker 302 is responsible for generating shooting guides 303 according to the user's request and the photo data storehouse 304 of message recipient 301 receptions.Take guide 303 and generally comprise photo example 3031 and corresponding acquisition parameters 3032 more than one.This part function can be by CPU1011 by being carried in software executing or microprocessor 2011 on the PC101 by the software executing that is carried on the camera 201.
To sum up; The function of taking guide maker 302 is to let numerous other users (for example numerous netizens) recommend how to take to a certain individual consumer; Comprise providing photo example 3031 and providing corresponding acquisition parameters 3032, and the result is shown on display 1014 or 2014.Obviously, this system utilizes the user to share a typical application of technology among the web2.0, but shares difference with traditional " entity " (document, photo or video), and the present invention has realized sharing of " technology ".The configuration and the operation of photograph taking guide automatic creation system will be described with reference to figure 5 after a while in more detail.
Fig. 4 is the exemplary plot to the automatic shooting guide that generates of zone name " Gold Gate Bridge ", comprises keyword 3031 (1), 3031 (2), 3031 (3) and corresponding acquisition parameters 3032 (1), 3032 (2), 3032 (3).For example, the shooting guide of generation may be displayed on the Wireless Telecom Equipment 201 of mobile device 101 or GPS embedding.
The exemplary configuration flow process of photograph taking guide automatic creation system 300 below will be described with reference to Fig. 5 in more detail.
As shown in Figure 5; Comprise a key modules in the photograph taking guide automatic creation system 300; Promptly take guide automatic generator 302, this maker comprises information retrieval device 3021, representative sample selector switch 3022, photo visual information 3023, photograph taking parameter 3024 and the photo text message 3025 and the contextual information 3026 that from photo data storehouse 304, extract.Wherein, information retrieval device 3021 comprises demand analysis device 3021A, feature extractor 3021B, similarity assessment device 3021C and sorting unit 3021D again; Representative sample selector switch 3022 comprises context collection device 3022A, clustering processing device 3022B and cluster centre selector switch 3022C again.
As an example, the target of information retrieval device 3021 is the user's requests that receive according to from message recipient 301, from photo data storehouse 304, collects relevant picture through network 104 or wireless network 204.The target of demand analysis device 3021A is the kind of decision demand and according to different kinds demand is carried out pre-service.The present invention can comprise following at least four kinds of demand types: latitude and longitude coordinates (for example can from GPS positioning system 2013, obtain), demand exemplary plot (for example can obtain from digital camera 2012), address information and textual description (for example can obtain through input equipment 1013).For latitude and longitude coordinates, pre-service mainly is meant the unitized of form; For the demand exemplary plot, pre-service mainly is meant basic Flame Image Process, for example schemes image intensifying, size scaling or the like; For address information, pre-service mainly is meant the GPS location; For textual description, pre-service mainly is meant basic text-processing The Application of Technology, and for example stem extracts (stemming), language unity (language unification) and (queryexpansion) or the like expanded in inquiry.Select as another kind, message recipient 301 also can design packet contains the function of demand analysis device 3021A.
Feature extractor 3021B extracts the characteristic of photo in the photo data storehouse according to the different demands kind; For example for the demand that requirement of geographical information (latitude and longitude coordinates or address information) and textual description are arranged; Characteristic generally is around photo, to extract the text message 3025, for example TF-IDF characteristic etc.; For the demand that provides the demand exemplary plot, characteristic generally is from the visual information 3023 of photo, to extract, for example color and texture etc.The computing method of said TF-IDF characteristic, color and textural characteristics are known by those skilled in the art, are not described in detail in this.
Similarity assessment device 3021C is responsible in assesses user demand and the photo data storehouse matching degree of photo in 304.As an example, matching degree can use the inverse of cosine similarity and Euclidean distance to weigh.The computing method of said cosine similarity and Euclidean distance are known by those skilled in the art, are not described in detail in this.
Sorting unit 3021D is responsible for the result according to similarity assessment device 3021C, and the photo in the comparison film data bank 304 sorts according to similarity.The photo of N position will be collected and be used for further selecting typical representative photo to appear to the user before coming, and wherein N can select according to the practical application scene flexibly.
Representative sample selector switch 3022 is to be responsible for from the result of sorting unit 3021D, selecting typical representative photo, and this process need is taken all factors into consideration visual information 3023, text message on every side 3025 and the contextual information 3026 of photo.It is huge to consider that the number of pictures of from said sorting unit, exporting still is divided into, if the information that all photos is relevant with photo is presented to the user, user plane is difficult to get access to useful information on the contrary to the information of numerous and complicated.Thereby be necessary from these photos further to select representative typical photo; Here will take a kind of clustering algorithm of extensive multi-modal data to address this problem; Cluster centre is named a person for a particular job and is counted as representational photo, the prominent example of just taking.Specifically, representative sample selector switch 3022 comprises context collection device 3022A, clustering processing device 3022B and cluster centre selector switch 3022C.
Context collection device 3022A is responsible for from contextual information 3026, collecting the contextual information relevant with photo.Contextual information 3026 extensively exists in the picture resource of network environment, for example picture browsing number of times, comment number of times, picture URL or the like.In general, if picture receives a lot of users' concern, then selected is that the probability of example is bigger.
Clustering processing device 3022B is responsible for the photo based on contextual information of information retrieval device 3021 outputs is carried out cluster; If be difficult to realize effect preferably with general popular clustering algorithm in the present invention, following 2 reasons are arranged: the sample (the crowd's data in the internet) and characteristic kind (comprising text feature, visual signature and contextual feature or the like) quantity of 1) participating in cluster are various; 2) data of different types has different weights, different dimensions and different disposal routes.Fig. 8 will illustrate a kind of clustering method of large-scale multi-modal data, and mode also can be regarded as data type here.
Cluster centre selector switch 3022C is responsible for selecting the cluster centre of each type in the cluster result; Described cluster centre is meant a most representative sample in its residing type; In general; Cluster centre is often more similar with a lot of samples during it is residing type, and a common visual signature is often arranged.For a specific place, cluster centre often can provide the most comprehensively visual information.Thereby, recommend the most representative typical photo example to the user, the photo that cluster centre is corresponding will be hopeful to be selected most.After typical photo example was selected to accomplish, corresponding acquisition parameters 3024 will extract and submit to the user automatically.
Below will generate method automatically with reference to the photograph taking guide that Fig. 6 to 8 describes in detail according to the embodiment of the invention.
Fig. 6 is illustrated schematically in the process flow diagram that photograph taking guide under the situation of geography information generates method automatically.As an example, this flow process can be carried out by software by CPU1011 among the PC101 or by microprocessor in the digital camera 201 2011.
As shown in Figure 6, in step 601, geography information will be transfused to, and said geography information is meant place name, the address of artificial input or the latitude and longitude coordinates of being obtained by GPS equipment.For example GPS positioning system 2013 can receive geography information.
In step 602, a large amount of image datas obtain from outside photo resource.For example, information retrieval device 3021 is collected photo according to the geography information of input from photo data storehouse 304.Described photo resource can be the network image that photo sharing website or online image search engine obtain, or based on the collection of photographs of the shared collection of private photograph album.From these photos, one group of representative typical photo is selected in step 603, and in order to select these typical photos, the process need of cluster is implemented, and the method for concrete cluster is illustrated as an example in Fig. 8.For example, representative sample selector switch 3022 is selected representative sample by clustering processing device 3022B.
In step 604, the metadata relevant with said selecteed representational sample is extracted out, and generation shooting guide is presented to the user with the result in step 605.For example, taking guide 400 is to generate the instance of coming out to zone name " Gold Gate Bridge ".
Fig. 7 is illustrated schematically in the process flow diagram that photograph taking guide under the situation that does not have geography information generates method automatically.As an example, this flow process also can be carried out by software by CPU1011 among the PC101 or by microprocessor in the digital camera 201 2011.
As shown in Figure 7, in step 701, text key word or demand legend or both are transfused to simultaneously.Then step 702 is obtained a large amount of image datas according to text key word or demand legend from outside photo resource.If the input content only is that text key word is transfused to, the text-based image retrieval technology will be used, and described text-based image retrieval is meant the text message that utilizes image to be associated, and handles image retrieval problem according to the mode of text retrieval; If the input content only is the demand legend, the CBIR technology will be used, and described CBIR is meant according to Image Visual Feature, according to the distance ordering in feature space; If the existing text key word of input content also has the demand legend, adopt after the text-based image retrieval, can be optimized result for retrieval through the demand legend.Result for retrieval to last sorts, and the top n photo will be collected, and N is a positive integer here, can confirm according to the application scenarios of reality.
The processing of subsequent step is identical with the processing of the top described step of process flow diagram with reference to Fig. 6, is not described in detail in this.
Fig. 8 is the exemplary process flow diagram 603 that the clustering algorithm of extensive multi-modal data is shown, and mode is meant data type here.The target of this flow process is that fairly large photo is carried out cluster, and regards cluster centre as representative typical photo.
In step 6031, the data (sample) with multi-modal characteristic are transfused to, and for example the photo that obtains in information retrieval device 3021 is imported in the representative sample selector switch 3022, and clustering algorithm will be performed in representative sample selector switch 3022.
In step 6032, master mode and auxilliary mode will be set up according to application scenarios, and master mode is meant main data type here, and auxilliary mode is meant auxiliary data type.For example, the network picture can be characterized by visual signature 3023, simultaneously also can by around text message 3025 (for example picture description, mark, comment and URL etc.) and various contextual information 3026 (for example picture browsing number of times, number of reviews or the like) characterize.If clustering algorithm is an object with this network picture, visual signature is the most direct information often, thus be set to master mode, and other each category feature is traditionally arranged to be auxilliary mode.
In step 6033, inner relation and the relation between the different modalities of various mode established.For the relation of mode inside, the sample of participating in cluster is characterized as being an eigenvectors, and inner relation can be characterized as being similarity/distance, grade subordinate and coexistence relation or the like.For the relation between the different modalities, available already present association characterizes, if for example visual signature all is to be used for describing an identical width of cloth picture with text message, then this visual signature just exists related with text message.
In step 6034; The tentative definite cluster centre of characteristic according to auxilliary mode; Here the most representative point in the class under cluster centre can be regarded as, in most cases, cluster centre is directly replaced by all some mean value of coordinate in feature space in the affiliated class.According to different application, different mode characteristics can be used.Be example still with the network picture; Often intrinsic dimensionality is bigger as the visual signature of master mode, and operand is beyond affordability for large-scale data, and text mark information can be selected as important clue and select cluster centre; For example; If a width of cloth picture only has only a mark, and this mark very common (for example " animal ", " trees " etc.), then this width of cloth picture probably is the typical picture that comprises tag content; And the possibility that is selected as cluster centre point is very big, and this possibility is referred to as " exploratory score " in the present invention.Although it is thus clear that the method accuracy of this initial selected central point is not very high, have the high advantage of efficient, in subsequent step, can do further adjustment again to central point.If but direct direct cluster on visual signature, to such an extent as to calculated amount is difficult to be applied in the middle of the large-scale data too greatly.If μ
i(j) (i=1,2 ..., M; J=1,2 ..., N) being score on j the mode of i width of cloth image, the exploratory company that may be defined as each mode of then final i width of cloth image takes advantage of form:
The image of the correspondence that exploratory score is the highest is exploratory type of final center.
In step 6035, if two or more exploratory type of center is more close in master mode or auxilliary mode space, then said exploratory type of center will be merged, and promptly removes a class center at random.Close program can be used various similarities/distance metric method, for example L1 distance, L2 distance and cosine similarity etc.Also can according to the practical application scenario definition special apart from balancing method, for example, mark " animal " and " dog " close on semantic level, and mark " animal " and " football " distance is just far away comparatively speaking.In some applications; The coexistence relation of mark also can be employed; For example, mark " buildings " and " open air " occur in the same secondary picture of being everlasting, and then exist the picture of mark " buildings " more approaching to a certain extent with the distance of the picture that has mark " open air ".
In step 6036, after tentative establishment cluster centre, other sample images will be distributed to each cluster centre to form class separately.According to the application scenarios of reality, the associate feature between various samples can be utilized.For example, sample and the distance of type center, compossibility of the mark of sample and the mark at type center or the like in the visual signature space.if
(i=1; 2 ... M; J=1,2 ... L; K=1,2 ... be i picture and j type the related score at k exploratory type of center S), then i picture should be assigned to G exploratory class center, can be drawn by following formula:
Here ω
jBe the weights of every type of association, can rule of thumb artificially confirm, also can obtain by machine learning method through one group of observed value.
In step 6037, all distribute to each cluster centre with after forming class separately at all samples, exploratory type of center will be adjusted.New class center is to be the corresponding point of high density at the master mode feature space.Adjusting type two steps in center in the step 6036 in distribution residue photo and the step 6037 can repeat, till satisfying the cluster condition.Here the cluster condition can be the number of times that repeats, and can confirm according to the practical application scene.After repeating repeatedly, final cluster result and cluster centre will be determined, and final cluster centre will be counted as representative photo 1031.Based on these representational photos, extract the metadata of photo and select useful acquisition parameters 1032 and present to the user and browse.Acquisition parameters can include but not limited to camera manufacturer, camera model, exposure value, light sensitivity, aperture size, focal length, shutter speed, shooting date and time or the like.
Those skilled in the art can conceive the whole bag of tricks selected presentation graphics and acquisition parameters are presented to the user.For example, can be according to representative fraction in quantity type of providing center of photo in each type, and the image that will have a high representative fraction is presented in the predetermined display part.Also can representative fraction be stored in the storer, or otherwise present to the user, as with order printing of representative fraction etc.
Should be noted in the discussion above that aforementioned a series of processing can pass through hardware or software executing.Passing through under the situation of the more aforementioned processing of software executing, the program that constitutes this software is installed from network or recording medium.
The method and system of presentation graphics that is used for seeking image collection according to the embodiment of the invention has been described with reference to the drawings above.Should be noted in the discussion above that the foregoing description only is exemplary, and also nonrestrictive.Those skilled in the art can carry out various modifications and replacement to the embodiment of the invention fully, and do not deviate from scope of the present invention.
Claims (10)
1. method that the guide that is used to take pictures generates automatically comprises:
Based on the residing environment of user, position and/or user's input, confirm user's request;
Based on described user's request, from the photo data storehouse, obtain relevant picture;
According to the visual signature and the contextual information of photo, from the said relevant picture of obtaining, select representative photo;
Extract each item acquisition parameters of said representational photo; And
Based on said representational photo and acquisition parameters thereof, generate the guide of taking pictures and present to the user.
2. the method for claim 1, wherein said definite user's request comprises:
Judge the residing latitude and longitude coordinates of photographing device based on global position system GPS; Or
Based on the shooting address of user's input or the text description of geographic name; Or
Based on the captured photo example that can describe the user's request scene of photographing device; Or
Text description based on the photographic subjects of user input.
3. the method for claim 1, wherein said photo data storehouse comprises:
The network image that obtains based on photo sharing website or online image search engine; Or
The collection of photographs of collecting based on sharing of private photograph album.
4. the representative photo of the method for claim 1, wherein said selection comprises:
Collect the contextual information of photo, include but not limited to photo description, mark, spectators' comment, number of visits, comment number of times;
Extract the visual signature of photo, include but not limited to color, texture, shape;
Based on the contextual information and the visual signature of photo, multi-modal cluster is carried out in the set of comparison film;
Based on cluster result, the photo that the compute classes center is corresponding.
5. method as claimed in claim 4, wherein, said comparison film set is carried out multi-modal cluster and is comprised:
It is that master mode, described contextual information are auxilliary mode that described visual signature is set, and the relation between definite different modalities;
Based on described auxilliary mode characteristic, type of selection center, exploratory ground;
In conjunction with described master mode and auxilliary mode characteristic, merge the high class center of similarity degree;
Based on similarity, grade subordinate relation, the symbiosis information at each type center, the residue photo is assigned in the class under each type center;
Based on the local density of master mode, adjust the class center of each type;
Said distribution residue photo can repeat with adjustment type two processes in center, till satisfying the cluster condition; And
Obtain final cluster result, and the corresponding photo in type of returning center.
6. the method for claim 1, wherein said acquisition parameters includes but not limited to camera manufacturer, camera model, exposure value, light sensitivity, aperture size, focal length, shutter speed, shooting date and time.
7. system that the guide that is used to take pictures generates automatically comprises:
The demand receiver is configured to the input based on the residing environment of user, position and/or user, confirms user's request;
The photo data storehouse; Be configured to store the database of photo data; The network image that for example obtains based on photo sharing website or online image search engine perhaps based on the collection of photographs of the shared collection of private photograph album, is used for therefrom selecting high-quality photos with the take pictures resource of guide of generation;
The guide maker of taking pictures; Be configured to based on described user's request; Extract the visual signature and the contextual information of photo; From said photo data storehouse, obtain representative photo and acquisition parameters thereof, for example camera manufacturer, camera model, exposure value, light sensitivity, aperture size, focal length, shutter speed, shooting date and time; And
The guide show stand of taking pictures is configured to show high-quality photos and each item acquisition parameters thereof to the user.
8. system as claimed in claim 7, wherein, the said guide maker of taking pictures comprises:
The information retrieval device is configured to based on described user's request, from the photo data storehouse, obtains relevant photo; And
The representative photo selector switch is configured to visual signature and contextual information according to said photo, from the said relevant picture of obtaining, selects representative photo.
9. system as claimed in claim 8, wherein, said information retrieval device comprises:
The demand analysis device is configured to confirm the kind of user's request and it is carried out corresponding pre-service;
Feature extractor, the every secondary photo that is configured in the comparison film data bank extracts text feature and visual signature;
The measuring similarity device, the degree of correlation that is configured to measure the every secondary photo in user's request and the photo data storehouse; And
Sorting unit is configured to the degree of correlation based on said every secondary photo, and all photos in the comparison film data bank sort.
10. system as claimed in claim 8, wherein, said representative photo selector switch comprises:
Context information collector is configured to collect and the related information of photo, includes but not limited to photo description, mark, spectators' comment, number of visits, comment number of times;
The clustering processing device is configured to contextual information and visual signature based on said photo, the set of comparison film carry out multi-modal cluster; And
Class center extraction apparatus is configured to based on cluster result the photo that the compute classes center is corresponding.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110003813.2A CN102591868B (en) | 2011-01-10 | 2011-01-10 | System and method for automatic generation of photograph guide |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201110003813.2A CN102591868B (en) | 2011-01-10 | 2011-01-10 | System and method for automatic generation of photograph guide |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102591868A true CN102591868A (en) | 2012-07-18 |
| CN102591868B CN102591868B (en) | 2015-03-18 |
Family
ID=46480537
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201110003813.2A Active CN102591868B (en) | 2011-01-10 | 2011-01-10 | System and method for automatic generation of photograph guide |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102591868B (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104038696A (en) * | 2014-06-10 | 2014-09-10 | 小米科技有限责任公司 | Shooting apparatus recommendation method and device |
| CN104735259A (en) * | 2015-03-31 | 2015-06-24 | 努比亚技术有限公司 | Mobile terminal shooting parameter setting method and device and mobile terminal |
| CN104767941A (en) * | 2015-04-15 | 2015-07-08 | 努比亚技术有限公司 | Photography method and device |
| WO2015180431A1 (en) * | 2014-05-30 | 2015-12-03 | 西安中兴新软件有限责任公司 | Shooting parameter setting method and device, terminal and storage medium |
| CN105488526A (en) * | 2015-11-26 | 2016-04-13 | 熊桂荣 | Automatic screening method of group sharing photos |
| CN105791700A (en) * | 2016-04-26 | 2016-07-20 | 努比亚技术有限公司 | Method and terminal of realizing image shooting |
| CN105915816A (en) * | 2016-07-06 | 2016-08-31 | 穆德远 | Method and equipment for determining brightness of given scene |
| CN108471499A (en) * | 2018-05-06 | 2018-08-31 | 深圳市保千里电子有限公司 | A method of taking pictures and select photo automatically based on artificial intelligence |
| WO2018170727A1 (en) * | 2017-03-21 | 2018-09-27 | Sony Mobile Communications Inc. | Method and apparatus for determining pickup parameters and electronic equipment |
| CN108848300A (en) * | 2018-05-08 | 2018-11-20 | 百度在线网络技术（北京）有限公司 | Method and apparatus for output information |
| CN108846375A (en) * | 2018-06-29 | 2018-11-20 | 山东大学 | A kind of multi-modal Cooperative Study method and device neural network based |
| CN108989683A (en) * | 2018-08-20 | 2018-12-11 | 崔跃 | A kind of automatic shooting system for children |
| CN110413806A (en) * | 2018-04-27 | 2019-11-05 | 发那科株式会社 | Image management apparatus |
| CN110832838A (en) * | 2018-09-28 | 2020-02-21 | 深圳市大疆创新科技有限公司 | Shooting method, shooting system and terminal |
| CN110868543A (en) * | 2019-11-25 | 2020-03-06 | 三星电子（中国）研发中心 | Intelligent photographing method and device and computer readable storage medium |
| US11431895B2 (en) | 2020-06-24 | 2022-08-30 | International Business Machines Corporation | Photography guidance based on crowdsourced photographs |
| CN116467607A (en) * | 2023-03-28 | 2023-07-21 | 阿里巴巴（中国）有限公司 | Information matching method and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060164536A1 (en) * | 2005-01-24 | 2006-07-27 | Casio Computer Co., Ltd. | Imaging device, imaging method and program thereof |
| CN101141569A (en) * | 2006-09-04 | 2008-03-12 | 佛山普立华科技有限公司 | Method of performing pick-up image through fast switching combination of different pick-up parameters |
| CN101193206A (en) * | 2006-11-30 | 2008-06-04 | 华晶科技股份有限公司 | Method for automatically adjusting shooting parameters according to shooting operation habit of the user |
| CN101398832A (en) * | 2007-09-30 | 2009-04-01 | 国际商业机器公司 | Image searching method and system by utilizing human face detection |
-
2011
- 2011-01-10 CN CN201110003813.2A patent/CN102591868B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20060164536A1 (en) * | 2005-01-24 | 2006-07-27 | Casio Computer Co., Ltd. | Imaging device, imaging method and program thereof |
| CN101141569A (en) * | 2006-09-04 | 2008-03-12 | 佛山普立华科技有限公司 | Method of performing pick-up image through fast switching combination of different pick-up parameters |
| CN101193206A (en) * | 2006-11-30 | 2008-06-04 | 华晶科技股份有限公司 | Method for automatically adjusting shooting parameters according to shooting operation habit of the user |
| CN101398832A (en) * | 2007-09-30 | 2009-04-01 | 国际商业机器公司 | Image searching method and system by utilizing human face detection |
Cited By (23)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2015180431A1 (en) * | 2014-05-30 | 2015-12-03 | 西安中兴新软件有限责任公司 | Shooting parameter setting method and device, terminal and storage medium |
| CN104038696A (en) * | 2014-06-10 | 2014-09-10 | 小米科技有限责任公司 | Shooting apparatus recommendation method and device |
| CN104735259B (en) * | 2015-03-31 | 2016-09-28 | 努比亚技术有限公司 | Mobile terminal acquisition parameters method to set up, device and mobile terminal |
| CN104735259A (en) * | 2015-03-31 | 2015-06-24 | 努比亚技术有限公司 | Mobile terminal shooting parameter setting method and device and mobile terminal |
| CN104767941A (en) * | 2015-04-15 | 2015-07-08 | 努比亚技术有限公司 | Photography method and device |
| CN105488526A (en) * | 2015-11-26 | 2016-04-13 | 熊桂荣 | Automatic screening method of group sharing photos |
| CN105488526B (en) * | 2015-11-26 | 2019-07-09 | 嵊州明智科技服务有限公司 | The auto-screening method of the shared photo of group |
| CN105791700A (en) * | 2016-04-26 | 2016-07-20 | 努比亚技术有限公司 | Method and terminal of realizing image shooting |
| CN105791700B (en) * | 2016-04-26 | 2019-05-21 | 努比亚技术有限公司 | A kind of method and terminal for realizing image taking |
| CN105915816A (en) * | 2016-07-06 | 2016-08-31 | 穆德远 | Method and equipment for determining brightness of given scene |
| CN105915816B (en) * | 2016-07-06 | 2018-12-07 | 穆德远 | Method and apparatus for determining the brightness of given scenario |
| WO2018170727A1 (en) * | 2017-03-21 | 2018-09-27 | Sony Mobile Communications Inc. | Method and apparatus for determining pickup parameters and electronic equipment |
| CN110413806A (en) * | 2018-04-27 | 2019-11-05 | 发那科株式会社 | Image management apparatus |
| CN108471499A (en) * | 2018-05-06 | 2018-08-31 | 深圳市保千里电子有限公司 | A method of taking pictures and select photo automatically based on artificial intelligence |
| CN108848300A (en) * | 2018-05-08 | 2018-11-20 | 百度在线网络技术（北京）有限公司 | Method and apparatus for output information |
| CN108846375A (en) * | 2018-06-29 | 2018-11-20 | 山东大学 | A kind of multi-modal Cooperative Study method and device neural network based |
| CN108989683A (en) * | 2018-08-20 | 2018-12-11 | 崔跃 | A kind of automatic shooting system for children |
| CN110832838A (en) * | 2018-09-28 | 2020-02-21 | 深圳市大疆创新科技有限公司 | Shooting method, shooting system and terminal |
| WO2020061993A1 (en) * | 2018-09-28 | 2020-04-02 | 深圳市大疆创新科技有限公司 | Photographing method, photographing system, and terminal |
| CN110868543A (en) * | 2019-11-25 | 2020-03-06 | 三星电子（中国）研发中心 | Intelligent photographing method and device and computer readable storage medium |
| US11431895B2 (en) | 2020-06-24 | 2022-08-30 | International Business Machines Corporation | Photography guidance based on crowdsourced photographs |
| CN116467607A (en) * | 2023-03-28 | 2023-07-21 | 阿里巴巴（中国）有限公司 | Information matching method and storage medium |
| CN116467607B (en) * | 2023-03-28 | 2024-03-01 | 阿里巴巴（中国）有限公司 | Information matching method and storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102591868B (en) | 2015-03-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102591868B (en) | System and method for automatic generation of photograph guide | |
| Li et al. | GPS estimation for places of interest from social users' uploaded photos | |
| CN103069415B (en) | Computer-implemented method, computer program and computer system for image procossing | |
| US8731308B2 (en) | Interactive image selection method | |
| US20140355907A1 (en) | Photo and video search | |
| US20110184953A1 (en) | On-location recommendation for photo composition | |
| Luo et al. | Event recognition: viewing the world with a third eye | |
| EP3164811B1 (en) | Method for adding images for navigating through a set of images | |
| EP4057163B1 (en) | Facilitating use of images as search queries | |
| US20120027311A1 (en) | Automated image-selection method | |
| US20120030575A1 (en) | Automated image-selection system | |
| JP2011511365A (en) | Recognition of image environment by image and position | |
| US20150161169A1 (en) | Interactive image selection method | |
| Zhuang et al. | Anaba: An obscure sightseeing spots discovering system | |
| Cao et al. | Learning human photo shooting patterns from large-scale community photo collections | |
| Ivanov et al. | Geotag propagation in social networks based on user trust model | |
| Qian et al. | On combining social media and spatial technology for POI cognition and image localization | |
| US20120027293A1 (en) | Automated multiple image product method | |
| US20120027303A1 (en) | Automated multiple image product system | |
| Li et al. | Geo-visual ranking for location prediction of social images | |
| Kim et al. | Classification and indexing scheme of large-scale image repository for spatio-temporal landmark recognition | |
| Li et al. | Global-scale location prediction for social images using geo-visual ranking | |
| Phan et al. | Collaborative recommendation of photo-taking geolocations | |
| CN105095215B (en) | Information acquisition device, method and server | |
| Jones et al. | Automated annotation of landmark images using community contributed datasets and web resources |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |