CN101206749B - Merchandise recommending system and method using multi-path image retrieval module thereof - Google Patents
Merchandise recommending system and method using multi-path image retrieval module thereof Download PDFInfo
- Publication number
- CN101206749B CN101206749B CN 200710106991 CN200710106991A CN101206749B CN 101206749 B CN101206749 B CN 101206749B CN 200710106991 CN200710106991 CN 200710106991 CN 200710106991 A CN200710106991 A CN 200710106991A CN 101206749 B CN101206749 B CN 101206749B
- Authority
- CN
- China
- Prior art keywords
- image
- user
- search
- information
- commodity
- 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.)
- Expired - Fee Related
Links
Images
Landscapes
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
Abstract
The present invention relates to a merchandise recommending system, and it is an object of the present invention is to derive recommended merchandise through a multiple image search, in which image characteristic information is extracted through a text search or an image search, thereby deriving recommended merchandise. To accomplish the above object, according to one aspect of the present invention, there is provided an operator server comprising a data receiving unit for receiving a 'merchandise search request signal' containing a text search or an image search (request) from the user computer and receiving a unique identification number of each user input information and merchandise information together with a corresponding matching table from the manager computer, a matching process module unit for sequentially arranging images by performing a command processing on search keywords that are searched through the characteristic information extraction module unit, a merchandise recommendation module unit for deriving recommended merchandise using the image characteristic information according to a search result, a data transmission unit for transmitting the merchandise extracted through the merchandise recommendation module unit to the user computer, and a data storage unit stores the user input information, merchandise information, unique identification numbers, and matching table.
Description
Technical field
The present invention relates to a kind of commercial product recommending system, more particularly, relate to a kind of commercial product recommending system and method thereof that draws Recommendations by multiple image search, wherein, extract image feature information by text search or picture search, thereby draw Recommendations.
Background technology
The technology relevant with the traditional commodities commending system comprises that (use virtual portrait clothes are bought the creation commending system of information in No. the 2005-0017534th, the early stage publication application of Korea S, call " prior art " in the following text), and have multiple technologies to carry out patented claim and registration.
As shown in Figure 1, the creation commending system of prior art comprises: Internet server, and it is used for operation and the internet website that provides virtual portrait to dress up service and clothes recommendation service to the user is provided; The shopping center clothing information is collected server, and it is used for collecting the relevant toggery information of clothes of selling with the clothes shopping center, a plurality of the Internet that operates on the Internet; Virtual portrait clothing information management server, its be used for according to virtual portrait dress up service make collect the virtual portrait clothes and buy information; Be used for providing the clothes recommendation server of clothes recommendation service, it is analyzed the dress up clothes of service-user of the virtual portrait that uses virtual portrait clothes to buy information and buys preference, buys the closely-related clothes of preference and recommend to be fit to the true clothes that the user samples from toggery information search and this clothes; And database, it is used for storing virtual personage clothes and buys information, toggery information and clothes purchase preference information.
As mentioned above, if the user buys virtual portrait clothes on the Internet virtual portrait of dressing up, so that it looks like he or she oneself, prior art provides suitable user's creation recommendation service by virtual portrait clothes purchase information.But, owing to recommending corresponding clothes to the user after the virtual portrait of dressing up, therefore have a problem, namely the clothes scope that can select of user only limits to the virtual portrait clothes.
In addition, in the prior art, owing to can only confirming by virtual portrait clothes purchase information user's clothes preference, so should use the virtual portrait service of dressing up, so that the creation recommendation service to be provided, therefore have a problem, namely need extra cost and time.
Simultaneously, although in cyberspace at present virtual portrait used represent to represent user in cyberspace the user by icon (comprising jewelry, clothes etc.), rather than represent truly the user by user's the bodily form, taste and appearance.Therefore, even recommend to be fit to the clothes of user's taste to the user according to the virtual portrait clothes, can not determine whether the clothes of recommending are fit to user (buyer).
In other words, a problem is arranged, although namely the user buys clothes according to the preference of oneself, can not confirm in advance whether the clothes of buying are fit to user's the bodily form.
Summary of the invention
Therefore, proposed the present invention in order addressing the above problem, and first purpose of the present invention to be to extract image feature information by text search or picture search, thereby to draw Recommendations.
Second purpose of the present invention is that the virtual portrait by the expression user bodily form makes user (buyer) be easy to confirm the applicability of the commodity recommended.
The 3rd purpose of the present invention is the execution stack program, in order to can draw successively according to the order of priority the Recommendations that extract according to image feature information.
Description of drawings
Preferred embodiment according to hereinafter describing in detail by reference to the accompanying drawings can be well understood to above-mentioned and other objects, features and advantages of the present invention.Appended figure comprises:
Fig. 1 is the view that shows the configuration of the traditional creation commending system that uses the virtual portrait clothes to buy information;
Fig. 2 a shows to draw the block diagram of illustrative configurations of the commercial product recommending system of Recommendations according to the first embodiment of the present invention by multiple image search;
Fig. 2 b shows to draw the view of configuration of the commercial product recommending system of Recommendations according to the first embodiment of the present invention by multiple image search;
Fig. 3 is the view of the virtual portrait that shows that the essential information based on the user according to the first embodiment of the present invention creates;
What Fig. 4 was general description according to the first embodiment of the present invention draws the process flow diagram of the Method of Commodity Recommendation of Recommendations by multiple image search;
Fig. 5 is the view that shows according to the screen of uploading pattern of the first embodiment of the present invention;
Fig. 6 is that detailed description is according to the process flow diagram of the step of the reception " commercial articles searching request signal " of the first embodiment of the present invention;
Fig. 7 a is the view that shows according to the contour images registration tool of the first embodiment of the present invention;
Fig. 7 b shows the view that creates instrument according to the profile of the first embodiment of the present invention;
Fig. 8 is the view that shows according to the profile projects combo instrument of the first embodiment of the present invention;
Fig. 9 is the view that shows according to the pattern image registration tool of the first embodiment of the present invention;
Figure 10 is the view that shows according to the commodity registration tool of the first embodiment of the present invention;
Figure 11 shows to draw the view of the screen of commodity according to the first embodiment of the present invention with text search;
Figure 12 is the view that shows according to the screen that draws Recommendations of the first embodiment of the present invention;
Figure 13 is the view that shows according to the virtual portrait of wearing Recommendations of the first embodiment of the present invention;
Figure 14 shows to draw the view of the screen of commodity according to the first embodiment of the present invention with picture search;
Figure 15 is the view that the illustrative configurations of the system of preferentially recommending the commodity that profile, pattern and color be complementary according to a second embodiment of the present invention is described;
Figure 16 is the process flow diagram that the method for preferentially recommending the commodity that profile, pattern and color be complementary according to a second embodiment of the present invention is described.
Embodiment
By the preferred embodiment that describes in detail referring to accompanying drawing, the features and advantages of the invention will become apparent.Implication and concept that in whole part instructions and claims, term used and wording must be interpreted as technology purport according to the invention, and abide by the principle: the inventor can suitably define term in order to explain its invention in the mode of the best.In addition, in the following description, it should be noted that and no longer known function and structure to be elaborated, in order to avoid unnecessarily obscure purport of the present invention.
The below is with reference to the commercial product recommending system of Fig. 2 a to 3 explanation according to the first embodiment of the present invention.
Fig. 2 a shows to draw the block diagram of illustrative configurations of the commercial product recommending system of Recommendations according to the first embodiment of the present invention by multiple image search, Fig. 2 b be show according to the first embodiment of the present invention draw the view of configuration of the commercial product recommending system of Recommendations by multiple image search, and Fig. 3 is the view of the virtual portrait that shows that the essential information based on the user according to the first embodiment of the present invention creates.
As shown in Fig. 2 a, the commercial product recommending system that draws Recommendations by multiple image search according to the present invention comprises subscriber computer 100, seller's computing machine 200, administrator computer 300 and operator's server 400.The below illustrates its function separately with reference to Fig. 2 b.
At first, in the present embodiment, subscriber computer 100 comprises a plurality of users or the buyer who searches for or buy commodity by commercial product recommending system.The user can be at commodity in use commending system after membership.The membership program is well-known, will no longer describe.Simultaneously, subscriber computer 100 also comprises a plurality of sellers, and can upload seller's merchandise news.
As shown in Figure 2, subscriber computer 100 receives the user's input information that comprises personal information data, search condition data and uploading data from the user, come search commercial articles by text search or picture search, and show the virtual portrait view of wearing the Recommendations image, so that the user can confirm commodity.Subscriber computer comprises data input module unit 110, commercial articles searching modular unit 120, virtual portrait creating unit 130, display unit 140, storage unit 150 and transmission unit 160.
Specifically, data input module unit 110 receives user's input information from the user, receives " the buying the approval confirmation signal " that is used for user's input information from administrator computer 300, and receives from operator's server 400 the commodity image that institute extracts or recommends.
In addition, user's input information also comprises personal information data, search condition data and uploading data.The personal information data comprise user's personal information, such as age, height, body weight, chest measurement, waistline, hip circumference, sex, the colour of skin, the bodily form etc.The search condition data comprise application target, field of employment, use object, concept of commodity etc.Uploading data comprises color, pattern, profile, image, commodity etc.
On this point, when data input module unit 110 receives user's input information from the user, preferably receive simultaneously " commercial articles searching request signal ", because as mentioned above, the user inputs personal information data, search condition data and uploading data, with search commercial articles.
Here, because profile included in uploading data, pattern, color etc. can be proposed by the user, and also can order, therefore separately need keeper's approval so that it is uploaded.Specifically, contour images and the pattern data with expression commodity common version comes structural configuration database (DB).Contour images and pattern data are made by keeper or user, and are registered in database.During the registration profile, register contour images together with information such as sex, age, season, styles.During registered design, together with information registered design images such as material, pattern characteristics (horizontal line, decorative pattern, conventional symbols, English letter etc.), overall main colors.In addition, the essential information that the merchandising database registration is relevant with the commodity of real trade, such as photo, price, quantity, detailed description etc., selected corresponding contour images or pattern image, and additional information, such as sex, age, season, style, material, pattern characteristics (horizontal line, decorative pattern, conventional symbols, English letter etc.), overall main color etc.If have the website of the detailed classification of suitable profile, pattern, color etc., analyze in advance corresponding website, and the identical code system of the code of registering in use and outline data storehouse is registered in classification information in category database.It is such analyzing corresponding website and the method for constructs database in advance, once a day or after connecting to collect the image and data of corresponding website by the TCP/IP socket with the specific time interval, in the matching list that will illustrate below data content is stored in, and store the image in corresponding file.Simultaneously, if register the merchandise news registered in other websites etc., connect by above-mentioned TCP/IP socket and the key element mark of the commodity in use page is collected commodity data, and register according to key class.
Commercial articles searching modular unit 120 comes search commercial articles by text search or picture search.Specifically, for the execution contexts search, at first carry out key word according to user's input information and search for first, then search comprises the classification of key word.For the carries out image search, select image by key word or classification, thus the search commercial articles image.
Simultaneously, in picture search, can come searching image with the image that draws from this paper search.That is to say, can be according to drawn Search Results from image feature information search key and classification.On this point, if drawn a plurality of images, can and toply draw a plurality of optimized images with the quantity of key word.
Virtual portrait creating unit 130 creates user adaptation virtual portrait images, and it represents the user according to personal information data (comprising user's sex, age, height, body weight, chest measurement, waistline, hip circumference etc.).
In other words, as shown in Figure 3, virtual portrait creating unit 130 is by selecting or age of input user, height, body weight, chest measurement, waistline, hip circumference, sex, the bodily form etc. create user adaptation health virtual portrait, and by selecting the shape of face, hair style, the colour of skin etc. or coming the virtual portrait of completing user oneself by the facial photo of uploading user oneself.
Display unit 140 has shown the virtual portrait that creates, the commodity image that its dress is recommended by commercial articles searching, so that the user can confirm the commodity image recommended, this also can confirm by independent output unit.That is to say, represent similarly user's the shape of face and the bodily form due to virtual portrait, so the user can determine whether the commodity of recommending are suitable by the virtual portrait that the commodity image is worn in observation.
The virtual portrait that storage unit 150 storages create and the user's input information that receives from the user.
Transmission unit 160 receives the request of text search or picture search from the user, and transmission " commercial articles searching request signal " and user's input information.
That is to say, transmission unit transfers to " commercial articles searching request signal " operator's server 400 and user's input information is transferred to administrator computer 300.
Then, seller's computing machine 200 receives merchandise news from the seller, so that a plurality of sellers can vending articles, will corresponding " marketing request signal " and merchandise news be transferred to administrator computer 300, and receive " sell and ratify confirmation signal " according to the registration of merchandise news from administrator computer 300.
On this point, use independent commodity registration tool, coming registration goods by the merchandise news of selecting to define in this instrument, and merchandise news comprises sex, age, season, style, material, pattern-information and other details.
Specifically, other details comprise the information relevant with the Ms, such as jewelry, trousers, upper garment, overcoat, suitcase, cardigan, jacket, vest, jacket, knit garment, cap, shoes, shirt, skirt, socks etc., perhaps comprise the information relevant with the man, such as jewelry, trousers, overcoat, suitcase, cardigan, jacket, vest, jacket, knit garment, cap, shoes, shirt, socks etc.
Simultaneously, during the registration profile, register contour images together with information such as sex, age, season, styles.During registered design, together with information registered design images such as material, pattern characteristics (horizontal line, decorative pattern, Taiji symbol, English letter etc.), overall main colors.During registration goods, register the essential information relevant with commodity, such as photo, price, quantity, detailed description etc., and selection is fit to contour images and the pattern image of commodity, its registered in advance is in operator's server 400, and the information such as extra registration sex, age, season, style, material, pattern, color, classification.
Then, administrator computer 300 is approved user's input information and the merchandise news of subscriber computer 100 and 200 inputs of seller's computing machine, for the unique identification number of each information creating, and the structure matching list, so that user's input information and merchandise news are mated mutually.Administrator computer comprises: boundary element 310; Approval process unit 320, it is used for determining whether to ratify user's input information and merchandise news; And matching list modular unit 330.
At first, boundary element 310 receives user's input information and the merchandise news by subscriber computer 100 and 200 inputs of seller's computing machine, to reach " selling the approval confirmation signal " of being used for merchandise news for " buying the approval confirmation signal " of user's input information and be transferred to subscriber computer 100 and seller's computing machine 200, and user's input information and each unique identification number that creates of merchandise news will be transferred to operator's server 400 together with corresponding matching list.
Matching list modular unit 330 is for being contained in each item number in user's input information and merchandise news according to creating unique identification number, and the structure matching list, with search commercial articles swimmingly.The matching list modular unit comprises: table module 331, and it is used for comprising classification information, season, color, pattern, sex, purposes, material etc.; Unique identification number creation module 332, it is used to corresponding to each item number of user's input information and merchandise news according to creating recognizable unique identification number; Image processing module 333, the image information that comprises in its user's input message and merchandise news is inserted image feature information in matching list; And matching module 334, it is used for showing module 331 and mutually mates with unique identification number.
The below shows detailed description the information registered in module 331.During the registration profile, register contour images together with information such as sex, age, season, styles.During registered design, together with information registered design images such as material, pattern characteristics (horizontal line, decorative pattern, Taiji symbol, English letter etc.), overall main colors.During registration goods, register the essential information relevant with commodity, such as photo, price, quantity, detailed description etc., and selection is fit to contour images and the pattern image of commodity, its registered in advance is in operator's server 400, and the information such as extra registration sex, age, season, style, material, pattern, color, classification.Therefore, when user by selecting profile, pattern and color are come construct image, recommend to be fit to the commodity of this image.That is to say, if having the website of the detailed classification of suitable profile, pattern, color etc., analyze in advance the classification of corresponding website and it is configured to database, and the image under the classification of automatic recommendation and the handled images match of user.
Image processing module 333 inserts characteristic information by scrambler, and comes analytical characteristic information with demoder.
That is to say, image processing module comprises: the scrambler (not shown), and it is used for inserting characteristic information; And the demoder (not shown), it is used for analytical characteristic information.As a reference, technology that be used for to insert characteristic information comprises that watermark, image header information are processed, the location mode of data file and layered approach etc.
Simultaneously, operator's server 400 receives " commercial articles searching request signal, and draw the commodity image according to the image information that text search or picture search obtain from subscriber computer 100.Operator's server comprises data receiver unit 410, characteristic information extracting module unit 420, matcher modular unit 430, commercial product recommending modular unit 440, data transmission unit 450 and data storage cell 460.
Data receiver unit 410 receives from subscriber computer 100 " the commercial articles searching request signal " that comprises text search or picture search (request), and receives unique identification number and the corresponding matching list of each user's input information and merchandise news from administrator computer 300.
420 search of response " commercial articles searching request signal " execution contexts or the picture search of characteristic information extracting module unit, and extract image feature information.The characteristic information extracting module unit comprises: text search module 421, it is carried out key word according to user's input information and searches for first, and according to the text input from corresponding to search text word information the table of definition command, to extract image feature information; And image search module 422, it is in the situation that picture search is extracted image feature information from the matching list that comprises image information and unique identification number.
In addition, image search module 422 can be come searching image with the image that draws by text search, and namely image search module can be according to drawn Search Results from image feature information search key and classification.
On this point, if draw a plurality of images, can come these images of comparison by layered approach, to draw a plurality of optimized images.
Then, 430 pairs of matcher modular units are processed by the search key fill order that characteristic information extracting module unit 420 searches, to be arranged in order image.The matcher modular unit comprises matcher module 431, and it is used for searching for unique identification number according to search information; The key word of search subscriber input again in Search Results, and process order corresponding to these key words; And storehouse module 432, it uses Characteristic of Image information and layer value to begin to be arranged in order from the optimized image that searches.
As a reference, storehouse has the parallel and serial structure that comprises image feature information, key word and image layer value.That is to say, basically, the quantity of (coupling) search key that has according to each image at first is stored in the image in classification in storehouse.Purpose is come co-treatment and show in an orderly manner the image that searches by storehouse, rather than the image that random display searches for the purpose of fast search.Therefore, store by storehouse the image that searches according to these key words in classification at first.In storehouse, use each Characteristic of Image information and the layer value of storing, the optimized image that searches is stored in the front portion of the region of search in storehouse, and will have identical layer value but the image feature information less image corresponding with these key words is placed on the rear portion of stack region, thereby be arranged in order image.The storehouse of the image that is arranged in order is transferred to search system, thereby jointly send the image that arranges to user.On this point, quantity according to keywords, order top and image feature information are determined priority, but be the invention is not restricted to this.
Commercial product recommending modular unit 440 draws Recommendations according to Search Results with image feature information.
In other words, come the comparison information relevant with registration goods with detection in conjunction with selected image information.Come search commercial articles according to precision and similarity, and draw commodity by the order of degree of precision and similarity.
Data transmission unit 450 will transfer to subscriber computer 100 by the commodity that commercial product recommending modular unit 440 extracts.
Data storage cell 460 storage user's input informations, merchandise news, unique identification number and matching list.
Hereinafter with reference to a kind of Method of Commodity Recommendation of Fig. 4 to 14 explanation, the multiple image search in its system that configures as shown in Figure 2 by the first embodiment according to the invention described above draws Recommendations.
What Fig. 4 was general description according to the first embodiment of the present invention draws the process flow diagram of the Method of Commodity Recommendation of Recommendations by multiple image search; Fig. 5 is the view that shows according to the screen of uploading pattern of the first embodiment of the present invention; Fig. 6 is that detailed description is according to the process flow diagram of the step of the reception " commercial articles searching request signal " of the first embodiment of the present invention; Fig. 7 a is the view that shows according to the contour images registration tool of the first embodiment of the present invention; Fig. 7 b shows the view that creates instrument according to the profile of the first embodiment of the present invention; Fig. 8 is the view that shows according to the profile projects combo instrument of the first embodiment of the present invention; Fig. 9 is the view that shows according to the pattern image registration tool of the first embodiment of the present invention; Figure 10 is the view that shows according to the commodity registration tool of the first embodiment of the present invention; Figure 11 shows to draw the view of the screen of commodity according to the first embodiment of the present invention with text search; Figure 12 is the view that shows according to the screen that draws Recommendations of the first embodiment of the present invention; Figure 13 is the view that shows according to the virtual portrait of wearing Recommendations of the first embodiment of the present invention; Figure 14 shows to draw the view of the screen of commodity according to the first embodiment of the present invention with picture search.
At first, as shown in Figure 4, operator's server 400 receives " commercial articles searching request signal " S2 from subscriber computer 100.
Specifically, in the step S2 that receives " commercial articles searching request signal ", subscriber computer 100 receives from the user and comprises the user's input information of personal information data, search condition data and uploading data, and user's input information is transferred to administrator computer 300.
As a reference, in after the user obtains keeper's approval, pattern data being uploaded to the step of operator's server 400, as shown in Figure 5, if user's registered design, pattern is shown in the pattern list as element " user's pattern ", thereby the user can select pattern.The user also can select position and the pattern itself of pattern.
Here, because profile included in uploading data, pattern, color etc. can be proposed by the user, and also can order, therefore separately need keeper's approval so that it is uploaded.
Simultaneously, for vending articles, seller's computing machine 200 is transferred to administrator computer 300 with corresponding " marketing request signal " and merchandise news.
That is to say, as shown in Figure 6, administrator computer 300 receives user profile and merchandise news S2a.
Specifically, in the step S2e of structure matching list, at first, for structural configuration and pattern database and merchandising database, carry out the program of registration contour images registration tool, profile establishment instrument, profile projects combo instrument, pattern image registration tool and commodity registration tool.
As shown in Figure 7a, the contour images registration tool comprises contour images log-in window 202, is used for showing classification window 204 and the log-on message selection window 206 of corresponding contour images classification.
On this point, log-on message preferably includes sex, age, season, style etc.Therefore, want the user who registers contour images to use this instrument input contour images and characteristic information thereof, and the registration contour images.
Registration process is such, namely makes contour images, by contour images log-in window 202, the contour images of making is uploaded, and selects contour feature information by log-on message selection window 206, and presses registration button and complete registration.On the other hand, as shown in Figure 7b, profile creates instrument not only can create the front side 220 of profile, and can create rear side 221, left side 222 and the right side 223 of profile.
In addition, profile projects combo instrument shown in Figure 8 comprises basic clothes selection window 224, selected clothes display window 225, detailed selection window 226 and exhaustive division selection window 227.
In addition, pattern image registration tool shown in Figure 9 comprises pattern registration window 208, types of patterns and size Selection window 210 and sex and age selection window 212.Therefore, want the user of registered design image to use this instrument selection to want pattern image and the characteristic information thereof of registering, and the registered design image.
As shown in figure 10, the commodity registration tool comprises commodity image registration window 214, characteristic information selection window 216 and sex and age selection window 218.Therefore, want the user of registration goods to come registration goods and the information relevant with commodity, such as commodity image, price, quantity, profile, classification, material, pattern, color, sex, age, season, style etc. with this instrument.That is to say, obtain the commodity image by taking commodity, come the photo of registration goods by commodity image registration window 214 with the image registration instrument, select the characteristic information of commodity by characteristic information selection window 216 and sex and age selection window 218, and the input details, to complete registration.
That is to say, as mentioned above, the essential information that the merchandising database registration is relevant with the commodity of real trade, such as photo, price, quantity, detailed description etc., selected corresponding contour images and pattern image, and additional information, such as sex, age, season, style, material, pattern characteristics (horizontal line, decorative pattern, conventional symbols, English letter etc.), overall main color etc.
If confirmed approval, will " buy the approval confirmation signal " from the administrator computer 300 of subscriber computer reception user's input information and transfer to subscriber computer 100, to " sell the approval confirmation signal " and transfer to seller's computing machine 200, and unique identification number and the matching list that creates transferred to operator's server 400.
Then, operator's server 400 determines that received " commercial articles searching request signal " is text search signal or picture search signal S4.
If receive the text search signal in step S4, as shown in figure 11, operator's server 400 is by text search detected image characteristic information S6.
Specifically, in the step S6 of execution contexts search, operator's server 400 is carried out key word according to user's input information and is searched for first S6a, and search comprises the classification S6b of key word, and draws image feature information S6c from the image that comprises key word or the image in these classifications.
That is to say, if the user selects detail image from the image that searches, the characteristic information according to detail image comes search key and classification by picture search, to detect one group of image, and draw image feature information from the image that detects, thereby come Recommendations with matching list.
For example, come searching image if " climb the mountain " by input characters, result shows below.
Climb the mountain-climbing gear-(upper strata climbing gear, climbing trousers, cap, socks, shoes etc.)
-camp-(camp lamp, alpin-stock, tent, equipment etc.)
-knapsack-(carry bag, doggie bag etc.)
As mentioned above, search for be included in " climbing the mountain " in all images corresponding to classification.
That is to say, come searching image if use key word " to climb the mountain ", search for representational image " climbing gear ", " camping " and " knapsack ", then the search presentation graphics corresponding with image under each classification " climbing gear ", " camping " and " knapsack ".
Next, operator's server 400 uses the image feature information that has extracted to be arranged in order the image S8 that all draw.
Specifically, operator's server 400 is searched for unique identification number to draw image S8a according to user's input information, the key word S8b of search subscriber input again in Search Results, process the order S8c corresponding with these key words, and use Characteristic of Image information and layer value to arrange successively S8d from the optimized image that searches.
In other words, detect the characteristic information of each image sets from each presentation graphics that searches.The condition that use detects, and key word and classification search condition take the key word of user input as the basis are arranged characteristic information in image sets and key word and the classification of similar image by storehouse.
As shown in figure 12, operator's server 400 uses image feature information to draw Recommendations S10 according to Search Results.
In step S10, according to priority with the Characteristic of Image communication arranged in step S8 to matching list.In matching list, according to the characteristic information of the image sets of transmitting carry out with operator's server 400 in the commodity of registration or the coupling of the commodity group that register in electronic market or shopping center.
Next, operator's server 400 transfers to subscriber computer 100 with drawn similar clause, so that the user can confirm commodity S12.
Specifically, in step S12, subscriber computer 100 creates user adaptation virtual portrait images, and it uses personal information data (comprising user's sex, age, height, body weight, chest measurement, waistline, hip circumference etc.) to represent user S12a.Subscriber computer receives drawn commodity from operator's server 40, commodity is through with it virtual portrait, and shows as shown in figure 13 virtual portrait S12b.
On this point, use the characteristic information of selected detail image that the detail items such as pattern, color and virtual portrait are combined, to show the collocation form (namely dressing up) that user's virtual portrait is carried out, and the characteristic information of transmitting image accordingly, to recommend similar commodity by matching list.
On the other hand, if receive the picture search signal in step S4, as shown in figure 14, operator's server 400 comes searching image characteristic information S14 by picture search.
For example, search for if selection belongs to " climbing gear " of the presentation graphics of " climbing the mountain ", result shows below:
Climb the mountain-climbing gear-the upper strata climbing gear
-climbing trousers
-cap
-socks
-shoes
That is to say, the image that comprises in search " climbing gear " table and the upper strata classification of " climbing gear ", the classification of namely " climbing the mountain ", and can search for similar image sets in this classification according to the characteristic information of " climbing gear " profile by the keyword search in this classification.
On this point, in picture search, can come searching image with search for the image that draws by this paper.That is to say, can be according to drawn Search Results from image feature information search key and classification.On this point, if drawn a plurality of images, can and toply draw a plurality of optimized images with the quantity of key word.
Hereinafter with reference to Figure 15, the second embodiment of the present invention is described.The second embodiment adopts and the similar form of the system and method for the first embodiment basically.Therefore, with the difference that stresses with the first embodiment.Figure 15 is the view that the illustrative configurations of the system of preferentially recommending the commodity that profile, pattern and color be complementary according to a second embodiment of the present invention is described.
At first, as shown in figure 15, the subscriber computer 100 of the second embodiment further comprises codi image creation modular unit 170, it is by selecting profile, pattern, color etc. to create the codi image that is fit to user's taste, purpose and the bodily form, and similarity searching modular unit 180, it is used for relatively and analyzes drawn image feature information and registration goods information, and presses the commodity sequential delivery commodity of optimum matching.
Here, preferably, use codi as the abbreviation of coordination (collocation).
Codi image creation modular unit 170 comprises selects module 171, and it is used for selecting registration profile, pattern and color; Conversion and correction module 172, it is used for proofreading and correct profile, pattern and color by selecting module 171 to select; And codi image creation module 173, it creates with profile, pattern or the color after proofreading and correct the codi image that is fit to user's taste, purpose and the bodily form.
Specifically, conversion and correction module 172 use 2.9D image transitions engines by the pattern, color, size, gauffer, rotation etc. of converted image in many ways such as pattern Mapping, color correction, size conversion, rotation conversion, pleating (form of expression of rivel), and described 2.9D image transitions engine adopts 3 dimensionizations 2 to tie up (2D) images (being called 2.9D in the present invention).On this point, the actuating speed of image is identical with the situation of 2D with capacity.If use existing 2D image or photograph image, can not carry out inherently pattern Mapping, color correction, size conversion etc., thereby a plurality of images in the usage data storehouse are changed pattern, color etc.
As a reference, 2.9D image transitions engine is the 2.9D image with the 2D image improvement, and it has virtual z value, to implement the function of 3D model.The 2.9D image that passes through to create, 2.9D image transitions engine can be used pattern Mapping, color correction, size conversion, rotation conversion etc.With regard to existing 3D model, in order to carry out pattern Mapping, color correction, size conversion etc., need the sources such as Planar Mapping model, 3D model, coordinate data of 3D form.Therefore, carry out mapping very consuming time, and file is too large, so 2.9D image transitions engine is not suitable for being used on Web.But improvement type mapping algorithm of the present invention can only use the 2.9D view data to carry out mapping, so its advantage is, can carry out quickly pattern Mapping, color correction, size conversion etc. than 3D model.And because improved mapping algorithm has the low capacity identical with the 2D image and high-speed, so it is easy to be applied on Web.And, therefore because the 2.9D image itself has virtual z value, the rotation conversion can be carried out, and pattern Mapping, color correction, size conversion, pleating etc. can be carried out, carry out simultaneously the rotation conversion.The mapping of the 3D model of carrying out when simultaneously, those operations are than rotation is faster.
That is to say, conversion and correction module 172 are to be different from the mode of existing 3D codi solution, and the 2.9D image of use itself (not being 3D rendering) is carried out pattern transfer, color conversion, size conversion, rotation conversion and pleating function.
On the other hand, similarity searching modular unit 180 is recommended the commodity that profile, pattern and color all be complementary with image feature information, the commodity of recommending profile and pattern match in next position in first position, and recommends the commodity of outline in next position again.
The concrete operations of similarity searching modular unit 180 are: the similarity searching modular unit is selected the representative contour images corresponding with each brand name in contour images.In the situation that carry out codi, the similarity searching modular unit is placed on limit priority on the brand name or electronic market trade name of keyword message when extracting the characteristic information of contour images, draw included similar clause in brand article group or electronic market commodity group in first position, and show similar commodity by virtual portrait to the user.
For example, in the situation that user-selected codi image means the contour images of jeans, if key word Levi ' s detected when extracting the characteristic information of contour images, in the brand article group of first location finding Levi ' s, and draw the recommended brands commodity corresponding with Levi ' s.On this point, the characteristic information of contour images has the keyword message of Levi ' s.
On this point, the commodity group of electronic market commodity is arranged, can pass through text search or picture search to user's recommendation name brands or electronic market commodity wherein, therefore draw brand article according to user-selected codi project.
Then, user-selected contour images is placed on user's virtual portrait.When virtual portrait is worn contour images, can freely change pattern or color according to user's needs, and contour images can be adjusted to the size of virtual portrait.And, can rotate the view of the virtual portrait of wearing contour images, to show left and right, the front and rear side view of virtual portrait.In addition, when carrying out pattern Mapping, also can change gauffer (form of expression of rivel), make the shape of its suitable contour images, and the view of wearing the virtual portrait of switched contour images can be shown on subscriber computer.
The below will by the above-mentioned system of structure as shown in figure 15 according to a second embodiment of the present invention, illustrate a kind of commodity of preferentially recommending profile, pattern and color to mate with reference to Figure 16.Here, Figure 16 is the process flow diagram that the method for preferentially recommending the commodity that profile, pattern and color be complementary according to a second embodiment of the present invention is described.
At first, as shown in figure 16, codi image creation modular unit 170 has created the codi image of suitable user's taste, purpose and the bodily form, to recommend S22 together with commodity.
Namely specifically, in the step S20 that creates the codi image, select module 171 user's input messages to select profile, pattern and color S22a.
Conversion and correction module 172 are by carrying out pattern Mapping, color correction, size conversion, rotation and change and pleating (form of expression of rivel) and the pattern of converted image, color, size, gauffer and rotation S22b in many ways to selected profile, pattern and color.
Then, codi image creation module 173 has created the codi image S22c of suitable user's taste, purpose and the bodily form.
Codi image creation modular unit 170 shows the codi image that has created by user's virtual portrait, so that the user can confirm codi image S24.
Here, when the contour images of user's virtual portrait dress combination, codi image creation modular unit 170 can be according to user's freely conversion of needs and correcting pattern or color, and contour images can be adjusted to the size of virtual portrait.Can rotate the view of the virtual portrait of wearing contour images, to show left and right, the front and rear side view of virtual portrait.
Similarity searching modular unit 180 use be stored in merchandising database or category database in the relevant essential information of the commodity of real trade, and corresponding to contour images, pattern image and the profile information of essential information, search for and the same or similar commodity image of the codi image S26 that creates.
Here, specifically, in the step S26 of search commercial articles image, similarity searching modular unit 180 draws characteristic information S26a from contour images, with the characteristic information that the draws S26b that compares with commodity Characteristic of Image information, and according to the sequential search same or similar commodity S26c of comparative result by similarity.
That is to say, the contour images that the user is made up is transferred to operator's server 400, and operator's server 400 above-mentioned matching list of use are extracted Characteristic of Image information and recommended similar commodity.
Therefore, the commodity that profile, pattern and color all be complementary with image feature information, the commodity of recommending profile and pattern match in next position are recommended and shown to similarity searching modular unit 180 in first position, and recommend the commodity S28 of outline in next position again.
In addition, the similarity searching modular unit is selected the representative contour images S28a corresponding to each brand name in contour images.In the situation that carry out codi, the similarity searching modular unit is placed on limit priority S28b on the brand name of keyword message or electronic market trade name when extracting the characteristic information of contour images, draw included similar clause S28c in brand article group or electronic market commodity group in first position, and show similar commodity S28d by virtual portrait to the user.
Although the present invention has illustrated a kind of commercial product recommending system and method thereof that stresses virtual portrait and toggery with reference to specific description embodiment, it is not subjected to the restriction of these embodiment, and only is subjected to the restriction of appended claims.It is to be appreciated that those skilled in the art that and to change these embodiments or to revise on the basis that does not exceed scope of the present invention and purport.
According to the invention described above, the user can buy commodity after extracting image feature information by text search or picture search and confirming to have the Recommendations of virtual portrait dress of user's bodily form, therefore it is very effective, because the user can buy the commodity that are fit to user's bodily form or the commodity that the user wants.
That is to say, can use the commodity of recommendation that the virtual portrait that is similar to user's bodily form is arranged in pairs or groups due to the user, and the commodity of range estimation confirmation recommendation, so it is very effective, because the user can buy the commodity that be similar to required commodity by text search or picture search.
Claims (19)
1. commercial product recommending system, it comprises:
Subscriber computer (100), it is used for receiving the user's input information that comprises personal information data, search condition data and uploading data, the user comes search commercial articles by text search or picture search, and shows the virtual portrait view of wearing the Recommendations image, so that can confirm commodity;
Seller's computing machine (200), it is used for receiving merchandise news from the seller who needs vending articles, transmission corresponding " marketing request signal " and merchandise news, and receive " selling the approval confirmation signal " according to the registration of merchandise news;
Administrator computer (300), it is used for approving user's input information and the merchandise news of subscriber computer and seller's computer input, create unique identification number for each user's input information and merchandise news, and the structure matching list, so that user's input information and merchandise news are mated mutually;
Operator's server (400), it is used for receiving " commercial articles searching request signal " from subscriber computer, and draws the commodity image according to the image information that obtains from text search or picture search;
Wherein, operator's server (400) comprising:
Data receiver unit (410), it is used for receiving from subscriber computer " the commercial articles searching request signal " that comprises text search or image search requests, and receives unique identification number and the corresponding matching list of each user's input information and merchandise news from administrator computer;
Characteristic information extracting module unit (420), it extracts image feature information by response " commercial articles searching request signal " execution contexts search or picture search; Characteristic information extracting module unit (420) comprises: text search module (421), it is used for carrying out key word according to user's input information searches for first, and inputs from searching for text word information to extract image feature information corresponding to the table of definition command according to text; And comprising image search module (422), it is used in the situation that searching image extracts image feature information from the matching list that comprises image information and unique identification number;
Matcher modular unit (430), it is processed the search key fill order that arrives by the characteristic information extracting module unit searches, to be arranged in order image; Matcher modular unit (430) comprises: matcher module (431), and it is used for searching for unique identification number according to search information, the key word of search subscriber input again in Search Results, and process the order corresponding with these key words; And comprising storehouse module (432), it begins to be arranged in order from the optimized image that searches with Characteristic of Image information and layer value;
Commercial product recommending modular unit (440), it is used for drawing Recommendations according to Search Results with image feature information;
Data transmission unit (450), it is used for and will transfers to subscriber computer by the commodity that the commercial product recommending modular unit extracts; And
Data storage cell (460), it is used for storage user's input information, merchandise news, unique identification number and matching list.
2. the system as claimed in claim 1, wherein, this subscriber computer (100) comprising:
Data input module unit (110), it is used for receiving user's input information from the user, receive " the buying the approval confirmation signal " that is used for user's input information from administrator computer (300), and receive from operator's server (400) the commodity image that institute extracts or recommends;
Commercial articles searching modular unit (120), it is used for searching for the classification that comprises the key word of having searched for after searching for first according to user's input information execution key word and searches for execution contexts, and select image and the search commercial articles image is searched for carries out image by key word or classification;
Virtual portrait creating unit (130), it creates the user adaptation virtual portrait image that represents the user with the personal information data;
Display unit (140), it be used for to show the virtual portrait of the commodity image that the dress that created is recommended by commercial articles searching, so that the user can confirm the commodity image recommended;
Storage unit (150), it is used for storing the virtual portrait that creates and the user's input information that receives from the user;
Transmission unit (160), it is used for receiving text search or image search requests from the user, and transmission " commercial articles searching request signal " and user's input information.
3. the system as claimed in claim 1, wherein, described user's input information comprises: the personal information data, it comprises user's age, height, body weight, chest measurement, waistline, hip circumference, sex, the colour of skin, the bodily form; The search condition data, it comprises application target, field of employment, use object or conceptual data; And uploading data, it comprises color, pattern, profile, image or commodity data.
4. system as claimed in claim 3, wherein, with regard to commodity data, registering in the situation that the merchandise news of other websites registrations, classification or key element mark by the connection of TCP/IP socket, the commodity in use page are collected commodity data, and come registration goods information according to classification or key class.
5. the system as claimed in claim 1, wherein, this administrator computer (300) comprising:
Boundary element (310), it is used for receiving user's input information and merchandise news by subscriber computer and seller's computer input, to reach " selling the approval confirmation signal " of being used for merchandise news for " buying the approval confirmation signal " of user's input information and be transferred to subscriber computer and seller's computing machine, and user's input information and each unique identification number that creates of merchandise news will be transferred to operator's server together with corresponding matching list;
Approval process unit (320), it is used for determining whether to ratify user's input information and merchandise news; And
Matching list modular unit (330), it is used for for each item number that is contained in user's input information and merchandise news according to creating unique identification number, and the structure matching list, to carry out swimmingly commercial articles searching;
This matching list modular unit (330) comprises: table module (331), and it is used for comprising classification information, season, color, pattern, sex, purposes, material; Unique identification number creation module (332), it is used to corresponding to each item number of user's input information and merchandise news according to creating recognizable unique identification number; The image information that image processing module (333), its use are contained in user's input information and merchandise news is inserted image feature information in matching list; And matching module (334), it is used for showing module (331) and unique identification number mates mutually.
6. system as claimed in claim 5, wherein, this table module (331) is constructed matching list by the process that registration contour images registration tool, profile create instrument, profile projects combo instrument, pattern image registration tool and commodity registration tool, with structural configuration and pattern database and merchandising database.
7. commercial product recommending system, it comprises:
Subscriber computer (100), it is used for receiving the user's input information that comprises personal information data, search condition data and uploading data, the user comes search commercial articles by text search or picture search, and shows the virtual portrait view of wearing the Recommendations image, so that can confirm commodity;
Seller's computing machine (200), it is used for receiving merchandise news from the seller who needs vending articles, transmission corresponding " marketing request signal " and merchandise news, and receive " selling the approval confirmation signal " according to the registration of merchandise news;
Administrator computer (300), it is used for approving user's input information and the merchandise news of subscriber computer and seller's computer input, create unique identification number for each user's input information and merchandise news, and the structure matching list, so that user's input information and merchandise news are mated mutually;
Operator's server (400), it is used for receiving " commercial articles searching request signal " from subscriber computer, and draws the commodity image according to the image information that obtains from text search or picture search;
Wherein, subscriber computer (100) comprising:
Codi image creation modular unit (170), it comprises:
Select module (171), it is used for selecting registered profile, pattern and color;
Conversion and correction module (172), it is used for changing selected profile, pattern and color;
Codi image creation module (173), it creates the codi image that is fit to user's taste, purpose and the bodily form with switched profile, pattern or color; And
Similarity searching modular unit (180), the image feature information that it is used for relatively and analysis has drawn and the merchandise news of registration, and transmit commodity according to the order of optimum matching commodity;
This conversion and correction module (172) are carried out pattern transfer, color conversion, size conversion, rotation conversion and pleating function to the 2.9D image;
This similarity searching modular unit (180) recommend in first position profile, pattern and color all with the commodity of image feature information coupling, recommend the commodity of profile and pattern match in next position, and recommend the commodity of outline in next position again.
8. a right to use requires 1 described commercial product recommending system to come the method for Recommendations, and the method comprises the following steps:
(1-a) receive " commercial articles searching request signal " by operator's server (400) from subscriber computer (100);
" the commercial articles searching request signal " that (1-b) receives by operator's server response and execution contexts search;
(1-c) by operator's server, to carry out key word according to user's input information and search for first, search comprises the classification of the key word of having searched for, and extracts image feature information from image or the image in these classifications that comprises these key words;
(1-d) by operator's server, use the image feature information that extracts by text search to draw and the Pareto diagram picture; And
(1-e) by operator's server, the Characteristic of Image communication that will arrange according to priority is to matching list, uses matching list that the characteristic information of the image sets transmitted is mated with registered commodity group, and recommends similar commodity.
9. method as claimed in claim 8, it is further comprising the steps before in step (1-a):
(1-a-1) receive user's input information and merchandise news by administrator computer (300);
(1-a-2) determine whether to ratify user's input information and merchandise news by administrator computer;
(1-a-3) create by administrator computer the table that comprises classification information, season, color, pattern, sex, purposes and material;
Be (1-a-4) according to creating recognizable unique identification number corresponding to each item number of user's input information and merchandise news by administrator computer; And
(1-a-5) construct matching list by administrator computer, should show and unique identification number coupling.
10. method as claimed in claim 8, wherein, (1-b) is further comprising the steps for step:
Search for by " commercial articles searching request signal " carries out image that the response of operator's server receives.
11. method as described in claim 8 or 10, wherein, (1-b) is further comprising the steps for step:
Use the image that draws from text search to come the carries out image search by operator's server; And
Search for described key word and classification according to the Search Results that has drawn from image feature information by operator's server.
12. method as claimed in claim 8, wherein, (1-d) is further comprising the steps for step:
(1-d-1) search for unique identification number by operator's server according to user's input information;
(1-d-2) again search in Search Results by the key word of operator's server user input;
(1-d-3) by the order of operator's server process corresponding to key word; And
(1-d-4) use Characteristic of Image information and layer value to begin to be arranged in order from the optimized image that searches by operator's server.
13. method as claimed in claim 8, it is further comprising the steps afterwards in step (1-e):
(1-e-1) receive the commodity of recommending from operator's server by subscriber computer;
(1-e-2) create expression user's user adaptation virtual portrait image by subscriber computer; And
(1-e-3) put on the commodity of recommendation and show virtual portrait to virtual portrait.
14. method as described in claim 8 or 13, it is further comprising the steps afterwards in step (1-e):
If drawn a plurality of images, can and toply draw a plurality of optimized images with the quantity of key word.
15. a right to use requires 7 described commercial product recommending systems to come the method for Recommendations, the method comprises the following steps:
(2-a) by the user adaptation virtual portrait creation of image codi image of codi image creation modular unit (170) according to suitable user's bodily form, this virtual portrait image user input message represents the user;
The user (2-b) shows the codi image that has created with user's virtual portrait by codi image creation modular unit, so that can confirm the codi image;
(2-c) by similarity searching modular unit (180), use contour images, pattern image and profile information corresponding to merchandise news to search for and the same or similar commodity image of the codi image that has created; And
(2-d) by the similarity searching modular unit, recommend and show the commodity that profile, pattern and color all mate, the commodity of recommending profile and pattern match in next position in first position, and recommend the commodity of outline in next position again.
16. method as claimed in claim 15, wherein, step (2-a) comprises the following steps:
(2-a-1) create the codi image by codi image creation modular unit;
(2-a-2) select profile, pattern and color by codi image creation modular unit;
(2-a-3) by codi image creation modular unit, adopt selected profile, pattern and the color of pattern, color, size, gauffer and rotation conversion conversion in many ways and correcting image; And
(2-a-4) create the codi image of the taste, purpose and the bodily form that are fit to the user by codi image creation modular unit.
17. method as claimed in claim 15, wherein, (2-b) is further comprising the steps for step:
By codi image creation modular unit, when the contour images of user's virtual portrait dress combination, according to user's needs conversion and correcting pattern or color; And
Wear the view of the virtual portrait of contour images by the rotation of codi image creation modular unit,, front and rear side view left and right to show.
18. method as claimed in claim 15, wherein, step (2-c) comprises the following steps:
Draw characteristic information by the similarity searching modular unit from contour images;
The characteristic information and the commodity Characteristic of Image information that draw are compared; And
According to comparative result, according to the same or similar commodity of the sequential search of similarity.
19. method as claimed in claim 15, wherein, step (2-d) comprises the following steps:
(2-d-1) select the representative contour images corresponding with brand name by the similarity searching modular unit in contour images;
(2-d-2) by the similarity searching modular unit, when extracting the characteristic information of profile, according to keyword message, priority is placed on brand name or electronic market trade name;
(2-d-3) draw included similar clause in brand article group or electronic market commodity group by the similarity searching modular unit in first position; And
(2-d-4) by the similarity searching modular unit, after virtual portrait is put on the commodity that draw, show similar commodity to the user.
Applications Claiming Priority (3)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| KR10-2006-0129744 | 2006-12-19 | ||
| KR1020060129744 | 2006-12-19 | ||
| KR1020060129744A KR100687906B1 (en) | 2005-12-31 | 2006-12-19 | System for recommendation the goods and method therefor |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101206749A CN101206749A (en) | 2008-06-25 |
| CN101206749B true CN101206749B (en) | 2013-06-05 |
Family
ID=39566926
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 200710106991 Expired - Fee Related CN101206749B (en) | 2006-12-19 | 2007-05-16 | Merchandise recommending system and method using multi-path image retrieval module thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101206749B (en) |
Families Citing this family (32)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP5811849B2 (en) * | 2009-12-24 | 2015-11-11 | 株式会社ニコン | Search support system, search support method, and search support program |
| US9792638B2 (en) | 2010-03-29 | 2017-10-17 | Ebay Inc. | Using silhouette images to reduce product selection error in an e-commerce environment |
| US8861844B2 (en) | 2010-03-29 | 2014-10-14 | Ebay Inc. | Pre-computing digests for image similarity searching of image-based listings in a network-based publication system |
| US8412594B2 (en) * | 2010-08-28 | 2013-04-02 | Ebay Inc. | Multilevel silhouettes in an online shopping environment |
| KR101440254B1 (en) * | 2011-02-04 | 2014-09-16 | 라쿠텐 인코포레이티드 | Information supply device |
| JP5203490B2 (en) * | 2011-07-20 | 2013-06-05 | 楽天株式会社 | SEARCH DEVICE, SEARCH PROGRAM, COMPUTER-READABLE RECORDING MEDIUM CONTAINING SEARCH PROGRAM, AND SEARCH METHOD |
| JP5044035B1 (en) * | 2011-07-29 | 2012-10-10 | 楽天株式会社 | Information providing apparatus, information providing method, information providing program, and computer-readable recording medium storing the program |
| WO2013023358A1 (en) * | 2011-08-16 | 2013-02-21 | Google Inc. | Using image match technology to improve image advertisement quality |
| CN102567543B (en) * | 2012-01-12 | 2015-02-18 | 北京搜狗信息服务有限公司 | Clothing picture search method and clothing picture search device |
| CN103325047B (en) * | 2012-03-22 | 2018-01-16 | 北京京东尚科信息技术有限公司 | Net purchase guide device and method |
| JP5256362B1 (en) * | 2012-03-30 | 2013-08-07 | 楽天株式会社 | Information processing apparatus, information processing method, information processing program, and recording medium |
| US20150058170A1 (en) * | 2012-03-30 | 2015-02-26 | Nec Corporation | Digital signage system, digital signage, article information presenting method and program |
| US20140075393A1 (en) * | 2012-09-11 | 2014-03-13 | Microsoft Corporation | Gesture-Based Search Queries |
| CN103106240A (en) * | 2012-12-12 | 2013-05-15 | 江苏乐买到网络科技有限公司 | Method of searching products in online shopping |
| JP6129600B2 (en) * | 2013-03-18 | 2017-05-17 | 株式会社東芝 | Information processing method and information processing system |
| CN103412937B (en) * | 2013-08-22 | 2016-12-28 | 成都数之联科技有限公司 | A kind of search purchase method based on handheld terminal |
| CN103646120A (en) * | 2013-12-27 | 2014-03-19 | 联想(北京)有限公司 | Information processing method and electronic device |
| CN103927348B (en) * | 2014-04-02 | 2018-09-07 | 华为技术有限公司 | Image processing method, information acquisition method and device |
| CN105157742B (en) * | 2014-04-29 | 2021-03-23 | 杭州美盛红外光电技术有限公司 | Identification device and identification method |
| CN104599153B (en) * | 2014-08-29 | 2017-02-22 | 腾讯科技（深圳）有限公司 | Commodity recommendation method, commodity recommendation server and commodity recommendation terminal |
| KR20160043588A (en) * | 2014-10-13 | 2016-04-22 | 삼성전자주식회사 | Method and apparatus for providing a content sevice |
| WO2016088920A1 (en) * | 2014-12-05 | 2016-06-09 | (주)위셔리 | System and method for recommending social commerce-based product |
| CN106056120B (en) * | 2016-06-14 | 2020-07-24 | 浙江理工大学 | Method for realizing clothes style identification and recommendation based on contour shape tracking of clothes |
| CN106874330A (en) * | 2016-07-27 | 2017-06-20 | 阿里巴巴集团控股有限公司 | A kind of resource supplying method and apparatus |
| WO2018023449A1 (en) * | 2016-08-02 | 2018-02-08 | 步晓芳 | Method for acquiring usage condition of clothing recommendation technique, and push system |
| WO2018023450A1 (en) * | 2016-08-02 | 2018-02-08 | 步晓芳 | Method for pushing ownership information when garment information is pushed, and pushing system |
| WO2018023451A1 (en) * | 2016-08-02 | 2018-02-08 | 步晓芳 | Method for pushing garment information by means of clothing store, and pushing system |
| CN112700296B (en) * | 2019-10-23 | 2022-05-27 | 阿里巴巴集团控股有限公司 | Method, device, system and equipment for searching/determining business object |
| CN113821662A (en) * | 2020-06-18 | 2021-12-21 | 北京小米移动软件有限公司 | Picture searching method and device and storage medium |
| CN112767063A (en) * | 2020-09-07 | 2021-05-07 | 肆嘉(上海)商务咨询有限公司 | Method and system for realizing virtuous circle network platform by user and seller |
| CN111951080A (en) * | 2020-09-22 | 2020-11-17 | 肆嘉(上海)商务咨询有限公司 | Method and system for integrating artificial intelligence into platform |
| CN113065922B (en) * | 2021-04-13 | 2021-12-07 | 南京莫愁智慧信息科技有限公司 | Electronic commerce shopping platform commodity intelligent matching recommendation method based on mobile internet and big data analysis |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1388461A (en) * | 2001-05-29 | 2003-01-01 | 黄俊诚 | Commodity recommending device and electronic transaction method |
| CN1588400A (en) * | 2004-10-21 | 2005-03-02 | 上海交通大学 | Method for recommending goods in electronic business |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US7904348B2 (en) * | 2004-05-05 | 2011-03-08 | Eplus Systems, Inc. | System and method for eCatalog supplier portal |
| US20050253840A1 (en) * | 2004-05-11 | 2005-11-17 | Kwon Ryan Y W | Method and system for interactive three-dimensional item display |
| US20060031128A1 (en) * | 2004-08-09 | 2006-02-09 | Lamitie Rickey K | System and associated method of marketing customized articles of clothing |
-
2007
- 2007-05-16 CN CN 200710106991 patent/CN101206749B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1388461A (en) * | 2001-05-29 | 2003-01-01 | 黄俊诚 | Commodity recommending device and electronic transaction method |
| CN1588400A (en) * | 2004-10-21 | 2005-03-02 | 上海交通大学 | Method for recommending goods in electronic business |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101206749A (en) | 2008-06-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101206749B (en) | Merchandise recommending system and method using multi-path image retrieval module thereof | |
| KR100687906B1 (en) | System for recommendation the goods and method therefor | |
| JP7196885B2 (en) | Search system, search method, and program | |
| JP6390693B2 (en) | Search support system, search support method, and search support program | |
| US7617016B2 (en) | Computer system for rule-based clothing matching and filtering considering fit rules and fashion rules | |
| US8370360B2 (en) | Merchandise recommending system and method thereof | |
| CN106649383A (en) | Clothes management method and system | |
| KR20200046924A (en) | The Automatic Recommendation System and Method of the Fashion Coordination | |
| US20150186965A1 (en) | Computerized device, system, and method for coordinating an outfit | |
| US20110184832A1 (en) | System and Method for Networking Shops Online and Offline | |
| US20090043665A1 (en) | Method and system of providing recommendations during online shopping | |
| US20090210320A1 (en) | System and method for comparative sizing between a well-fitting source item and a target item | |
| US20230140225A1 (en) | System for constructing virtual closet and creating coordinated combination, and method therefor | |
| CN106294419B (en) | Method and device for collecting service object information | |
| KR102307095B1 (en) | The Automatic Recommendation System and Method of the Fashion Coordination |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20130605 Termination date: 20170516 |