WO2019120031A1 - 服饰搭配推荐方法、装置、存储介质及移动终端 - Google Patents

服饰搭配推荐方法、装置、存储介质及移动终端 Download PDF

Info

Publication number
WO2019120031A1
WO2019120031A1 PCT/CN2018/116799 CN2018116799W WO2019120031A1 WO 2019120031 A1 WO2019120031 A1 WO 2019120031A1 CN 2018116799 W CN2018116799 W CN 2018116799W WO 2019120031 A1 WO2019120031 A1 WO 2019120031A1
Authority
WO
WIPO (PCT)
Prior art keywords
matching
clothing
suggestion
picture
information
Prior art date
Application number
PCT/CN2018/116799
Other languages
English (en)
French (fr)
Inventor
刘耀勇
陈岩
Original Assignee
Oppo广东移动通信有限公司
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Oppo广东移动通信有限公司 filed Critical Oppo广东移动通信有限公司
Publication of WO2019120031A1 publication Critical patent/WO2019120031A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
    • G06Q30/00—Commerce
    • G06Q30/06—Buying, selling or leasing transactions

Definitions

  • the embodiments of the present application relate to a mobile terminal technology, for example, to a clothing matching recommendation method, device, storage medium, and mobile terminal.
  • the user's clothing matching knowledge is generally obtained by reading a fashion magazine; some apparel providers try to display the clothing matching scheme provided by the designer of the clothing brand through a display screen set in the shopping mall, thereby providing the user with the clothing matching solution.
  • the brand's clothing recommendations, but this collocation proposal has limitations, and the real-time performance is not good.
  • the embodiment of the present invention provides a clothing matching recommendation method, device, storage medium and mobile terminal, which can effectively optimize the clothing matching recommendation scheme.
  • the embodiment of the present application provides a clothing matching recommendation method, including:
  • the classification model is a deep learning model trained according to different categories of apparel picture samples
  • the clothing matching suggestion is generated according to the category information, and the clothing matching suggestion is displayed.
  • the embodiment of the present application further provides a clothing matching recommendation device, and the device includes:
  • a picture capturing module configured to obtain a picture taken with a clothing item as a subject
  • a clothing identification module configured to input the captured picture into a pre-configured classification model to identify category information of the clothing item in the captured picture, wherein the classification model is trained according to different categories of clothing picture samples Deep learning model
  • the suggestion generation module is configured to generate a clothing matching suggestion according to the category information, and display the clothing matching suggestion.
  • the embodiment of the present application further provides a computer readable storage medium, where the computer program is stored, and the program is implemented by the processor to implement the clothing matching recommendation method.
  • the embodiment of the present application further provides a mobile terminal, including a memory, a processor, and a computer program stored on the memory and operable by the processor, where the processor implements the computer program Recommended clothing with matching methods.
  • FIG. 1 is a flowchart of a method for recommending clothing matching according to an embodiment of the present application
  • FIG. 2 is a schematic diagram of a clothing matching recommendation interface provided by an embodiment of the present application.
  • FIG. 3 is a flowchart of another method for recommending clothing matching according to an embodiment of the present application.
  • FIG. 4 is a flowchart of still another method for recommending clothing matching according to an embodiment of the present application.
  • FIG. 5 is a clothing matching recommendation device according to an embodiment of the present application.
  • FIG. 6 is a structural block diagram of a mobile terminal according to an embodiment of the present application.
  • FIG. 7 is a structural block diagram of a smart phone according to an embodiment of the present application.
  • the traditional classification algorithm generally adopts a two-step method.
  • the first step is to calculate the artificially set features from the input image.
  • the second step trains a classifier based on the calculated features, and the classifier is set to classify the test data. Due to the limitations of the features of artificial design, the effect of traditional methods depends to a large extent on whether the characteristics of artificial selection are reasonable, has great blindness, and generally has the problem of low classification accuracy.
  • the present application proposes a clothing matching recommendation scheme, which can quickly and accurately identify the type of clothing, and provide clothing matching suggestions in real time.
  • FIG. 1 is a flowchart of a clothing matching recommendation method according to an embodiment of the present application.
  • the method may be implemented by a clothing matching recommendation device, where the device may be implemented by software and/or hardware, and generally integrated in a mobile terminal.
  • a mobile terminal such as a mobile terminal with a camera.
  • the method includes:
  • Step 110 Acquire a photograph taken with the costume item as a subject.
  • apparel items can be a general term for items that decorate the human body, including but not limited to tops, pants, skirts, shoes, hats, scarves, ties, bags, belts, jewelry, and the like.
  • the captured picture may include a picture that the user has selected the clothing item as the subject, and presses the camera button to capture through the camera.
  • the photograph taken may be a photograph obtained by photographing a user's existing clothing item, or may be a picture obtained by photographing a clothing item that the user is trying on, or a clothing list hanging or tiling in the specialty store.
  • the captured picture may also be a historical captured picture obtained by the user clicking on the picture library (or album) under the shooting interface. For example, in the shooting interface, the user clicks on the picture located near the photo button to enter the photo library or album, and then selects the historical shooting picture including the clothing item (ie, the historical shooting picture with the clothing item as the shooting object).
  • the captured picture includes a picture currently taken by the user or a historical captured picture acquired from the picture library.
  • Step 120 Input the captured picture into a pre-configured classification model to identify category information of the apparel item in the captured picture.
  • the classification model is a deep learning model trained according to different categories of clothing picture samples.
  • the deep learning model may select a convolutional neural network model.
  • the above classification model is built, trained, and optimized in the server.
  • network parameters such as the number of layers of the convolutional neural network model, the number of neurons, the convolution kernel, and/or the weight are not limited.
  • the convolutional neural network model is trained by using different types of clothing picture samples to obtain a classification model.
  • the clothing picture samples used to train the convolutional neural network model include clothing items and jewelry single product pictures, including but not limited to shirt pictures, pants pictures, skirt pictures, shoe pictures, hat pictures, scarf pictures, tie pictures, Bag pictures, belt pictures, jewelry pictures.
  • the clothing item picture may be downloaded from the network platform through a web crawler, wherein the source of the clothing item picture in the network platform may be a shopping website, an electronic version of a fashion magazine, or a user-applied clothing item picture, etc. And sort the pictures of the downloaded clothing items.
  • the classification can be based on the gender of the clothing (ie, the audience is men or women), clothing style (that is, campus, retro, office or leisure), clothing styles (including long sleeves, short sleeves, trousers, shorts, etc., popular elements And clothing texture or material, etc.) and functional information (including regular clothing or functional clothing).
  • the women's T-shirt men's T-shirt, women's shirt, men's shirt, women's suit jacket, men's suit jacket, women's windbreaker, men's windbreaker, ladies coat, men's coat, leopard print, floral, campus and office, etc. sort.
  • the pants images are classified according to women's trousers, men's trousers, ladies shorts, men's shorts, campus, vintage and texture.
  • the dress pictures, shoe pictures, jewelry pictures and other clothing pictures can be classified according to actual needs.
  • the classification method is similar to the above-mentioned top picture and pants picture classification, and will not be enumerated here. Then, the classified clothing picture and its classification information are stored as a clothing picture sample.
  • model construction is performed, that is, the number of hidden layers and the number of nodes in each of the input layer, the hidden layer, and the output layer are preset, and the first parameter of the convolutional neural network is initialized, wherein the first parameter includes the bias of each layer.
  • the value of each edge and the weight of each edge are initially obtained as a framework for the convolutional neural network model.
  • the pre-set convolutional neural network model is trained in two stages of forward propagation and backward propagation.
  • the convolutional neural network model may be trained by using a forward propagation algorithm and a backward propagation algorithm according to the clothing picture sample, and the second parameter of the neural network model is learned, wherein the second parameter is the calculation of the clothing picture.
  • the first parameter is updated with the second parameter, and the above training process is continued.
  • the model error is calculated, wherein the model error can be determined according to the deviation between the actual output of the clothing picture sample and the expected output, and when the model error reaches the expected error value, the training ends, and the classification model is obtained.
  • the classification model is transplanted into the mobile terminal in response to the download request of the mobile terminal.
  • the classification model is optimized before the classification model is transplanted to the mobile terminal.
  • the convolutional neural network model is optimized by using a preset optimization strategy, wherein optimization of the convolutional neural network model includes optimization of internal network structure, optimization of implementation of convolutional layer, and pooling layer At least one of the implementation optimizations. For example, adding a residual block to construct a residual neural network model, or adjusting the structure of the residual block.
  • the optimization of the implementation of the convolution layer may be to reduce the number of connections of the output channel and the input channel, that is, the output channel is no longer related to all input channels, but only to adjacent input channels.
  • the base layer is added to the implementation of the convolution layer, and the volume is integrated into two steps: first, each channel of the input is operated separately, and under the action of the same size convolution kernel, each channel obtains an intermediate calculation result, which will Each channel of the intermediate calculation result is called a base layer; then, multiple channels are combined to obtain the output result of the convolution layer.
  • a matrix for image compression in the pooling layer is designed by image compression coefficients.
  • the clothing matching function is enabled, so as to provide a matching suggestion with the clothing item in the captured picture according to the specified picture taken by the user.
  • the acquired shooting picture is input into the classification model, and the category of the clothing item in the captured picture can be quickly and accurately identified, and the category information is obtained.
  • the category information includes the gender, style, style and function information corresponding to the clothing item. Since the features of the image are implicitly extracted by the classification model, the feature extraction by the artificial design in the related art is faster and more accurate.
  • Step 130 Generate a clothing matching suggestion according to the category information, and display the clothing matching suggestion.
  • the clothing matching suggestions include suggestion information such as a suit or a jewelry that matches the clothing items in the photograph, including but not limited to the text description of the matching clothing matched with the clothing item in the photographed picture, and the photographed picture A matching picture of the matching clothing items.
  • the collocation database pre-configured in the mobile terminal is queried according to the category information of the apparel item in the captured picture, and the collocation suggestion corresponding to the category information is determined.
  • the matching proposal includes a matching record of one bar.
  • a matching record for a sports ladies' tops includes a cotton sports skirt; and a matching pair of sports ladies' tops can be paired with a cotton harem pants; or, for a sporty women's top with a pair Functional quick-drying pants, etc.
  • the matching record for the ladies' suit includes a pair of tooling trousers; and the matching record for the ladies' suit can also include a skirt with a suit fabric; or the matching record for the ladies suit can also include a suit shorts.
  • the collocation record is an abstract collocation strategy by performing statistics, analysis, and processing on the clothing collocation picture in the network platform.
  • the collocation database is pre-built by the server, and the configured collocation database is delivered to the mobile terminal.
  • the server downloads clothing matching pictures from a web site such as a shopping website, a fashion magazine, or a clothing launch through a web crawler.
  • the method of big data analysis is used to analyze the downloaded clothing matching pictures, abstract the matching strategy of clothing and accessories, and store the category information of the clothing items and the characteristic information of the matching clothing in the form of matching records, and store the matching records in the matching.
  • the method of big data analysis is used to analyze the downloaded clothing matching pictures, abstract the matching strategy of clothing and accessories, and store the category information of the clothing items and the characteristic information of the matching clothing in the form of matching records, and store the matching records in the matching.
  • the database In the database.
  • FIG. 2 is a schematic diagram of a clothing matching recommendation interface provided by the present application. As shown in FIG. 2, after the user enables the costume matching function, the camera button 220 is pressed to take a picture of the hanging ladies round-neck long-sleeved solid color cotton shirt to obtain a captured picture 210.
  • the mobile terminal After acquiring the captured picture 210, the mobile terminal automatically inputs the captured picture 210 into a preset classification model, and identifies the category information of the top by the classification model, including: ladies, a round neck, a long-sleeved shirt, cotton, and a solid color. .
  • the preset matching database After obtaining the category information of the jacket, the preset matching database is called, and the corresponding matching suggestions according to the category information are included, but not limited to a, with a solid color V-neck sweater; b, with a pair of pants A; c, with A skirt B; d, with a hat C; e, with a shoulder bag D; f, with a watch E.
  • the above matching suggestion is presented in the form of a pop-up dialog 230.
  • the mobile terminal sends the collocation proposal to the cloud server, where the collocation suggestion is used to instruct the cloud server to query location information of the product that matches the collocation proposal; and obtain the return of the cloud server.
  • the matching is suggested to match the location information of the product, and the location information is displayed on the electronic map.
  • the cloud server eg, a mall server or a third party server
  • the mobile terminal sends the feature information of the matching costume in the pop-up dialog box to the cloud server to query whether the commodity information matching the feature information is provided in the mall or in another mall adjacent to the mall through the cloud server.
  • the merchant location and the store name are pushed to the mobile terminal.
  • the mobile terminal invokes the map application to mark the location information of the product and the store name in the map.
  • the captured image of the obtained clothing item is input into a pre-configured classification model by using the clothing matching function in the shooting interface to identify the category information of the clothing item in the captured picture.
  • the preset matching database is called, the clothing matching suggestion is generated, and the clothing matching suggestion is displayed.
  • the captured image is quickly and accurately identified according to the classification model, the category of the clothing item in the captured picture is determined, and the matching suggestion matching the same is determined according to the category, and the clothing item based on the user input is realized in time. Provide professional matching suggestions to shorten the time users spend on shopping.
  • FIG. 3 is a flowchart of another method for recommending clothing matching according to an embodiment of the present application. As shown in FIG. 3, the method includes:
  • Step 310 Acquire a photograph taken with the costume item as a subject.
  • Step 320 Obtain status information of the clothing matching function.
  • the clothing matching function is used to provide a clothing matching suggestion, and the function can be pre-configured in the mobile terminal and enabled or disabled according to an operation instruction input by the user.
  • the clothing matching function is integrated in the camera application of the mobile terminal, and if the user opens the camera application, the clothing matching function can be enabled or disabled by clicking the setting interface switch.
  • the status information includes an enabled state and a disabled state, and when the setting interface switch is in an open state, determining that the state information of the clothing matching function is an enabled state, and in a case where the setting interface switch is in a closed state, determining the The status information of the clothing matching function is disabled.
  • step 330 it is determined whether the clothing matching function is enabled according to the status information. If the clothing matching function is not enabled, step 340 is performed. If the clothing matching function is enabled, step 350 is performed.
  • step 350 is performed.
  • step 340 is performed.
  • Step 340 Store the captured picture in the picture library.
  • Step 350 Input the captured picture into a pre-configured classification model to identify category information of the apparel item in the captured picture.
  • Step 360 Query a preset matching database according to the category information, and determine a matching suggestion that matches the category information.
  • the category information includes gender, style, style and function information.
  • the preset matching database is queried according to the gender, style, style and function information of the clothing item, and the matching suggestion corresponding to the clothing item is determined. Since the association record in the preset matching database has the matching record corresponding to the category information, querying the preset matching database according to the category information of the clothing item in the captured picture can obtain the characteristics of the matching clothing matching the clothing item.
  • the feature information constitutes a matching suggestion.
  • the classification model determines the category information of the clothing items in the captured picture, including ladies, chiffon dresses, round necks, florals, and the like.
  • matching matching suggestions include but not limited to: a, with a solid color long sweater S; b, with a black leather coat; c, with a black boots; d With a solid color jacket R; e, with a plaid scarf; f, with a red handbag.
  • the above matching suggestions are stored in the preset matching database in text mode.
  • the text recording method can reduce the data amount of the preset matching database, and avoid excessively occupying the storage space of the mobile terminal.
  • the preset collocation database when the mobile terminal queries the preset collocation database according to the category information, the preset collocation database is invoked to determine a matching suggestion that matches the category information, and the default sorting is used in the form of text.
  • the suggestion is displayed in the display window.
  • the product image that meets the matching suggestion in the network platform is obtained, and the display is displayed in the form of a display window.
  • product picture For example, taking the category information of the clothing items in the picture, including the ladies, the chiffon dress, the round neck, and the floral, as an example, the network platform can be queried based on the above-mentioned matching suggestions, and the product picture of the matching clothing that meets the above matching suggestions is determined, and the display is adopted.
  • the form of the window scrolls to display the product image.
  • the user's click operation may also be detected, and the detailed information of the selected product image is provided in response to the click operation.
  • Step 370 Send the collocation recommendation to the cloud server.
  • the mobile terminal establishes a communication connection with the cloud server and sends the matching proposal to the cloud server.
  • the cloud server is configured to determine location information of the product matching the collocation suggestion according to the collocation suggestion, and mark the location information on the electronic map, and design a thumbnail of the product to be displayed on each product location.
  • Step 380 Acquire location information of the commodity returned by the cloud server that matches the collocation recommendation.
  • the location information of the product matching the matching suggestion is obtained by the cloud server, and the electronic map application is called to display the location information, and the thumbnail of the product is displayed at the location corresponding to the location information.
  • Step 390 Acquire feedback information of the user for the matching suggestion.
  • the feedback information includes the user's adoption of the matching suggestion, including but not limited to whether the user input adopts the adoption information of the matching suggestion. For example, if the user selects a product image to continue displaying the detailed information of the product image, it is considered The user has adopted the collocation proposal. For example, if the location information of the product matching the matching suggestion is displayed on the electronic map, if the user inputs the path planning instruction, and the destination of the path planning instruction is the location corresponding to the product, the user is deemed to have adopted the location information. Strip with suggestions.
  • obtaining the feedback information of the user for the collocation suggestion includes: obtaining the adoption information input by the user for the collocation recommendation as the feedback information.
  • the adoption information includes any of the product images selected by the user in the form of a display window, and the user inputs a path plan indicating the location information of the product.
  • Step 3100 Determine a user's matching preference according to the feedback information, and update a recommendation priority of the clothing matching suggestion in the matching database according to the matching preference.
  • the mobile terminal analyzes and summarizes the feedback information of the user in a certain period of time, determines the user's matching preference, and adjusts the recommendation priority of the matching recommendation in the matching database according to the matching. For example, based on the analysis and induction of the user feedback information within one month, and determining that the user likes casual clothing, a higher recommendation priority is set according to the preference for matching matching in the matching database. It can also be that, based on the analysis and induction of the user feedback information within one month, it is determined that the user likes the top with the skirt, and the higher recommended priority can be configured for the skirt in the matching database.
  • the history can be determined according to the history of the matching suggestions, and the user preferences are determined, and the matching suggestions are further sorted according to the user preference, so that different priorities are assigned to the matching suggestions according to the sorting result.
  • the technical solution of the embodiment obtains a photograph taken by the clothing item, and inputs the photographed image into a pre-configured classification model to identify category information of the clothing item in the photographed picture;
  • the category information is used to query a preset matching database, and the matching suggestion matching the category information is determined;
  • the product image in the network platform that meets the matching suggestion is obtained, and the product image is displayed in the form of a display window according to the user's matching preference. That is, the closer the product is to the user, the more advanced the ranking is, and the intuitive and visual display and matching suggestions are made, so that the user can purchase based on the matching suggestion, saving the time for the user to match the costume.
  • the collocation recommendation is sent to the cloud server, so that the navigation path to the location of the commodity can be planned through the electronic map, thereby saving the shopping time of the user.
  • FIG. 4 is a flowchart of still another method for recommending clothing matching according to an embodiment of the present application, where the method includes:
  • Step 410 Acquire a photograph taken with the costume item as a subject.
  • Step 420 Obtain status information of the clothing matching function.
  • Step 430 Determine whether the clothing matching function is enabled according to the status information. If the clothing matching function is not enabled, perform step 440. If the clothing matching function has been enabled, perform step 450.
  • Step 440 Store the captured picture in the picture library.
  • Step 450 Input the captured picture into a pre-configured classification model to identify category information of the apparel item in the captured picture.
  • Step 460 Query a preset collocation database according to the category information, and determine a clothing collocation picture that matches the category information.
  • the preset matching database stores the clothing matching pictures.
  • the collocation database is pre-built by the server, and the constructed collocation database is sent to the mobile terminal.
  • the server downloads clothing matching pictures from a web site such as a shopping website, a fashion magazine, or a clothing launch through a web crawler.
  • a series of image processing is performed on the clothing matching picture and stored in the preset matching database, and the category information of the clothing items included in each clothing matching picture is correspondingly marked.
  • the image processing includes, but is not limited to, performing deduplication processing on the clothing matching pictures with only different colors, and processing the clothing matching pictures with high definition and the like for the pictures with low discrimination.
  • the matching database Since the category information in the matching database is stored in association with the clothing matching picture, after identifying the category information of the clothing item in the captured picture, the matching database may be queried according to the category information, and the clothing matching picture matching the category information is determined.
  • Step 470 Generate a clothing matching suggestion according to the clothing and/or the accessory matched with the clothing item in the clothing matching picture, and display the clothing matching suggestion in the form of a display window.
  • the costume item in the picture is a women's chiffon floral dress
  • the clothing and accessories are generated, and the display window is displayed.
  • the form shows the clothing matching suggestions.
  • detailed information of one or more apparel items may also be displayed according to a user click operation.
  • Step 480 Send the collocation recommendation to the cloud server.
  • the matching suggestion corresponding to the click operation is sent to the cloud server.
  • the picture of the black leather is sent to the cloud server, so that the cloud server searches for the location of the product or the like according to the picture.
  • Step 490 Obtain location information of the commodity returned by the cloud server that matches the collocation recommendation.
  • the technical solution of the embodiment obtains a photograph taken by the clothing item, and inputs the photographed image into a pre-configured classification model to identify category information of the clothing item in the photographed picture;
  • the category information is used to query a preset matching database, and the clothing matching picture matching the category information is determined; according to the clothing and/or accessories matched with the clothing item in the clothing matching picture, a clothing matching suggestion is generated, and the display is performed
  • the form of the window displays the clothing matching suggestion, and the picture of the matching clothing matching the clothing item is searched in the matching database of the mobile terminal, and the clothing matching picture can be quickly displayed without relying on the network.
  • the collocation recommendation is sent to the cloud server, so that the navigation path to the location of the commodity can be planned through the electronic map, thereby saving the shopping time of the user.
  • FIG. 5 is a clothing matching recommendation device provided by an embodiment of the present application.
  • the device may be implemented by software and/or hardware, and may be integrated into a mobile terminal for performing the clothing matching recommendation method provided by the embodiment of the present application. As shown in Figure 5, the device includes:
  • the picture capturing module 510 is configured to acquire a captured picture that is a subject of the clothing item
  • the clothing identification module 520 is configured to input the captured picture into a pre-configured classification model to identify category information of the clothing item in the captured picture, wherein the classification model is training according to different categories of clothing picture samples. Deep learning model;
  • the suggestion generating module 530 is configured to generate a clothing matching suggestion according to the category information, and display the clothing matching suggestion.
  • the technical solution of the embodiment provides a clothing matching recommendation device, which quickly and accurately identifies the captured image according to the classification model, determines the category of the clothing item in the captured picture, and determines the matching suggestion according to the category, thereby realizing Based on the user-entered clothing items, timely professional advice is provided to shorten the time spent on shopping.
  • the apparatus further comprises:
  • a status obtaining module configured to obtain status information of the clothing matching function before inputting the captured picture into the pre-configured classification model, wherein the clothing matching function is used to provide a clothing matching suggestion;
  • the apparel identification module 520 is configured to:
  • the apparatus may include a function determining module configured to determine, according to the status information, whether the clothing matching function is enabled;
  • the apparel identification module 520 is configured to perform an operation of inputting the captured image into a pre-configured classification model.
  • the suggestion generation module 530 is configured to:
  • the suggestion generation module 530 is configured to:
  • the preset matching database is queried according to the gender, style, style and function information of the clothing item, and the matching suggestion corresponding to the clothing item is determined.
  • the apparatus further comprises:
  • a feedback information obtaining module configured to: after displaying the product image in the form of a display window, obtain feedback information of the user for the matching suggestion;
  • the priority adjustment module is configured to determine the user's matching preference according to the feedback information, and update the recommendation priority of the clothing matching suggestion in the matching database according to the matching preference.
  • the suggestion generation module 530 is configured to:
  • the clothing matching suggestion is generated according to the clothing and/or the accessory matched with the clothing item in the clothing matching picture, and the clothing matching suggestion is displayed in the form of a display window.
  • the apparatus further comprises:
  • a sending module is configured to send the collocation suggestion to the cloud server after the collocation suggestion is displayed, where the collocation suggestion is used to instruct the cloud server to query location information of the product matched with the collocation suggestion;
  • the location obtaining module is configured to obtain location information of the commodity returned by the cloud server that matches the matching suggestion, and display the location information.
  • the feedback information acquisition module is configured to:
  • the adoption information includes any of the product images selected by the user in the form of a display window, and the user inputs a path plan indicating the location information of the product.
  • the captured picture includes a picture currently taken by the user or a historically taken picture acquired from the picture library.
  • the embodiment of the present application further provides a storage medium including computer executable instructions for executing a clothing matching recommendation method when executed by a computer processor, the method comprising:
  • the classification model is a deep learning model trained according to different categories of apparel picture samples
  • the clothing matching suggestion is generated according to the category information, and the clothing matching suggestion is displayed.
  • Storage media any type of storage device or storage device.
  • the term "storage medium” is intended to include: a mounting medium such as a Compact Disc Read-Only Memory (CD-ROM), a floppy disk or a tape device; a computer system memory or a random access memory such as a dynamic random access memory; (Dynamic Random Access Memory, DRAM), (Double Data Rate Random Access Memory, DDR RAM), Static Random Access Memory (SRAM), Extended Data Output Random Access Memory (Extended Data Output Random Access Memory) , EDO RAM), Rambus Random Access Memory (Rambus RAM), etc.; non-volatile memory such as flash memory, magnetic media (such as hard disk or optical storage); registers or other similar types of memory components Wait.
  • a mounting medium such as a Compact Disc Read-Only Memory (CD-ROM), a floppy disk or a tape device
  • a computer system memory or a random access memory such as a dynamic random access memory
  • DRAM Dynamic Random Access Memory
  • DDR RAM Dyna
  • the storage medium may also include other types of memory or a combination thereof. Additionally, the storage medium may be located in a first computer system in which the program is executed, or may be located in a different second computer system, the second computer system being coupled to the first computer system via a network, such as the Internet. The second computer system can provide program instructions to the first computer for execution.
  • the term "storage medium" can include two or more storage media that can reside in different locations (eg, in different computer systems connected through a network).
  • a storage medium may store program instructions (eg, a computer program) executable by one or more processors.
  • the storage medium containing the computer executable instructions provided by the embodiment of the present application is not limited to the clothing matching recommendation operation as described above, and may also perform the clothing matching provided by any embodiment of the present application. Related operations in the recommended method.
  • the embodiment of the present application provides a mobile terminal, where the mobile terminal has an operating system, and the clothing matching recommendation device provided by the embodiment of the present application can be integrated into the mobile terminal.
  • the mobile terminal can be a smart phone, a portable device (PAD), a smart wearable device, or the like.
  • FIG. 6 is a structural block diagram of a mobile terminal according to an embodiment of the present application. As shown in FIG. 6, the mobile terminal can include a memory 610 and a processor 620.
  • the memory 610 is configured to store a computer program, a collocation database, a classification model, and the like; the processor 620 reads and executes a computer program stored in the memory 610.
  • the processor 620 implements the following steps: acquiring a captured picture with the apparel item as a photographic subject; inputting the captured picture into a pre-configured classification model to identify the clothing in the captured picture
  • the category information of the item wherein the classification model is a deep learning model trained according to different types of clothing picture samples; generating a clothing matching suggestion according to the category information, and displaying the clothing matching suggestion.
  • FIG. 7 is a structural block diagram of a smart phone according to an embodiment of the present application.
  • the smart phone may include: a memory 701, a central processing unit (CPU) 702 (also referred to as a processor), a peripheral interface 703, a radio frequency (RF) circuit 705, and an audio circuit.
  • CPU central processing unit
  • RF radio frequency
  • CPU central processing unit
  • RF radio frequency
  • the illustrated smartphone 700 is merely one example of a mobile terminal, and the smartphone 700 may have more or fewer components than those shown in the figures, two or more components may be combined, or may have different Component configuration.
  • Each of the components shown in the figures can be implemented in hardware, software, or a combination of hardware and software, including one or more signal processing and/or application specific integrated circuits.
  • the smart phone integrated with the clothing matching recommendation device provided in this embodiment is described in detail below.
  • the memory 701 can be accessed by the CPU 702, the peripheral interface 703, etc., and the memory 701 can include a high speed random access memory, and can also include a non-volatile memory, such as one or more magnetic disk storage devices, flash memory devices. Or other volatile solid-state storage devices.
  • the computer program is stored in the memory 701, and can also be combined with a database, a classification model, and the like.
  • Peripheral interface 703, which can connect the input and output peripherals of the device to CPU 702 and memory 701.
  • the I/O subsystem 709 can connect input and output peripherals on the device, such as touch screen 712 and other input/control devices 710, to peripheral interface 703.
  • the I/O subsystem 709 can include a display controller 7071 and one or more input controllers 7092 for controlling other input/control devices 710.
  • one or more input controllers 7092 receive electrical signals from other input/control devices 710 or transmit electrical signals to other input/control devices 710, and other input/control devices 710 may include physical buttons (press buttons, rocker buttons, etc.) ), dial, slide switch, joystick, click wheel.
  • the input controller 7092 can be connected to any of the following: a keyboard, an infrared port, a Universal Serial Bus (USB) interface, and a pointing device such as a mouse.
  • USB Universal Serial Bus
  • the touch screen 712 is an input interface and an output interface between the user terminal and the user, and displays the visual output to the user.
  • the visual output may include graphics, text, icons, videos, and the like.
  • Display controller 7071 in I/O subsystem 709 receives an electrical signal from touch screen 712 or an electrical signal to touch screen 712.
  • the touch screen 712 detects the contact on the touch screen, and the display controller 7071 converts the detected contact into an interaction with the user interface object displayed on the touch screen 712, that is, realizes human-computer interaction, and the user interface object displayed on the touch screen 712 can be operated.
  • the icon of the game, the icon of the network to the corresponding network, and the like.
  • the device may also include a light mouse, which is a touch sensitive surface that does not display a visual output, or an extension of a touch sensitive surface formed by the touch screen.
  • the RF circuit 705 is configured to establish communication between the mobile phone and the wireless network (ie, the network side) to implement data reception and transmission between the mobile phone and the wireless network. For example, sending and receiving short messages, emails, and the like.
  • RF circuit 705 receives and transmits an RF signal, also referred to as an electromagnetic signal, and RF circuit 705 converts the electrical signal into an electromagnetic signal or converts the electromagnetic signal into an electrical signal, and through the electromagnetic signal and communication network And other devices to communicate.
  • RF circuitry 705 may include known circuitry for performing these functions including, but not limited to, an antenna system, an RF transceiver, one or more amplifiers, a tuner, one or more oscillators, a digital signal processor, a codec CODER-DECoder (CODEC) chipset, Subscriber Identity Module (SIM), etc.
  • CDDEC codec CODER-DECoder
  • the audio circuit 706 is arranged to receive audio data from the peripheral interface 703, convert the audio data into an electrical signal, and transmit the electrical signal to the speaker 711.
  • the speaker 711 is arranged to restore the voice signal received by the mobile phone from the wireless network through the RF circuit 705 to sound and play the sound to the user.
  • the power management chip 708 is configured to provide power and power management for the hardware connected to the CPU 702, the I/O subsystem, and the peripheral interface.
  • the mobile terminal provided by the embodiment of the present application can provide professional matching suggestions in time based on the clothing items input by the user, and shorten the time spent by the user on shopping matching.
  • the clothing matching recommendation device, the storage medium, and the mobile terminal provided in the foregoing embodiments may perform the clothing matching recommendation method provided by any embodiment of the present application, and have the corresponding function modules and effects of executing the party.
  • the clothing matching recommendation method provided by any embodiment of the present application.

Landscapes

  • Business, Economics & Management (AREA)
  • Accounting & Taxation (AREA)
  • Finance (AREA)
  • Development Economics (AREA)
  • Economics (AREA)
  • Marketing (AREA)
  • Strategic Management (AREA)
  • Physics & Mathematics (AREA)
  • General Business, Economics & Management (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Management, Administration, Business Operations System, And Electronic Commerce (AREA)

Abstract

一种服饰搭配推荐方法、装置、存储介质以及移动终端，所述方法包括：获取以服饰单品为拍摄对象的拍摄图片（110）；将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息（120），其中所述分类模型为根据不同类别的服饰图片样本训练的深度学习模型；根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议（130）。

Description

服饰搭配推荐方法、装置、存储介质及移动终端
本申请要求在2017年12月20日提交中国专利局、申请号为201711382746.3的中国专利申请的优先权，该申请的全部内容通过引用结合在本申请中。
技术领域
本申请实施例涉及移动终端技术，例如涉及一种服饰搭配推荐方法、装置、存储介质及移动终端。
背景技术
随着经济的飞速发展，服饰产品的品种、样式变得丰富而复杂，同时，人们对服装及配饰的要求也日益提高。面对品种、样式繁杂的服饰，人们通常希望得到专业而科学的搭配建议。
但是，用户在逛街时很难及时的得到专家有针对性地帮助。相关技术中，用户的穿衣搭配知识一般通过阅读时尚杂志获取；也有一些服饰提供商尝试通过设置于商场内的显示屏展示该服饰品牌的设计师提供的穿衣搭配方案，从而，为用户提供该品牌的服饰的搭配建议，但是这种搭配建议存在局限性，且实时性不佳。
发明内容
本申请实施例提供一种服饰搭配推荐方法、装置、存储介质及移动终端，可以有效地优化服饰搭配推荐方案。
在一实施例中，本申请实施例提供了一种服饰搭配推荐方法，包括：
获取以服饰单品为拍摄对象的拍摄图片；
将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息，其中，所述分类模型为根据不同类别的服饰图片样本训练的深度学习模型；
根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议。
在一实施例中，本申请实施例还提供了一种服饰搭配推荐装置，该装置包括：
图片拍摄模块，设置为获取以服饰单品为拍摄对象的拍摄图片；
服饰识别模块，设置为将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息，其中，所述分类模型为根据不同类别的服饰图片样本训练的深度学习模型；
建议生成模块，设置为根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议。
在一实施例中，本申请实施例还提供了一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现上述的服饰搭配推荐方法。
在一实施例中，本申请实施例还提供了一种移动终端，包括存储器，处理器及存储在存储器上并可在处理器运行的计算机程序，所述处理器执行所述计算机程序时实现上述的服饰搭配推荐方法。
附图说明
图1是本申请实施例提供的一种服饰搭配推荐方法的流程图；
图2是本申请实施例提供的一种服饰搭配推荐界面示意图；
图3是本申请实施例提供的另一种服饰搭配推荐方法的流程图；
图4是本申请实施例提供的又一种服饰搭配推荐方法的流程图；
图5是本申请实施例提供的一种服饰搭配推荐装置；
图6是本申请实施例提供的一种移动终端的结构框图；
图7是本申请实施例提供的一种智能手机的结构框图。
具体实施方式
下面结合附图和实施例对本申请进行说明。此处所描述的实施例仅仅用于解释本申请，而非对本申请的限定。为了便于描述，附图中仅示出了与本申请相关的部分而非全部结构。
在更加详细地讨论示例性实施例之前应当提到的是，一些示例性实施例被描述成作为流程图描绘的处理或方法。虽然流程图将多个步骤描述成顺序的处理，但是其中的许多步骤可以被并行地、并发地或者同时实施。此外，一个或多个步骤的顺序可以被重新安排。当其操作完成时所述处理可以被终止，但是还可以具有未包括在附图中的附加步骤。所述处理可以对应于方法、函数、规程、子例程、子程序等等。
相关技术中，研究者们已经提出了诸多的实现服装自动分类的算法。传统 的分类算法一般采用两步方法，第一步从输入图像中计算出人为设置的特征，第二步根据算出的特征去训练一个分类器，该分类器设置为对测试数据进行分类。由于人工设计的特征的局限性，传统方法效果的好坏很大程度上取决于人为选择的特征是否合理，具有很大的盲目性，普遍存在分类准确率低的问题。为解决上述问题，本申请提出一种服饰搭配推荐方案，能够快速且准确的识别服饰类型，并实时提供服饰搭配建议。
图1为本申请实施例提供的一种服饰搭配推荐方法的流程图，该方法可以由服饰搭配推荐装置来执行，其中，该装置可由软件和/或硬件实现，一般可集成在移动终端中，如具有摄像头的移动终端。如图1所示，该方法包括：
步骤110、获取以服饰单品为拍摄对象的拍摄图片。
其中，服饰单品可以是装饰人体的物品的总称，包括但不限于上衣、裤子、裙子、鞋子、帽子、围巾、领带、提包、皮带、首饰等。
拍摄图片可以包括用户已选定的服饰单品为拍摄对象，按下拍照按钮，通过摄像头捕获的画面。其中，拍摄图片可以是对用户已有的服饰单品进行拍照得到的图片，还可以是对用户正在试穿的服饰单品进行拍照得到的图片，或者对专卖店中悬挂或平铺的服饰单品进行拍照得到的图片、或者行人穿着的服饰单品进行拍照得到的图片等。例如，用户在逛街时购买了一件上衣，为了得知与其搭配的服饰，可以对该上衣进行拍照，得到该上衣的拍摄图片。
在一实施例中，拍摄图片还可以是用户在拍摄界面下，点击图片库(或相册)获取的历史拍摄图片。例如，用户在拍摄界面下，点击位于拍照按钮附近的图片，进入图片库或相册，进而，选择包含服饰单品的历史拍摄图片(即，以服饰单品为拍摄对象的历史拍摄图片)。
在一实施例中，所述拍摄图片包括用户当前拍摄的图片或从图片库获取的历史拍摄图片。
步骤120、将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息。
其中，该分类模型为根据不同类别的服饰图片样本训练的深度学习模型。在一实施例中，深度学习模型可以选择卷积神经网络模型。在服务器中构建、训练并优化上述分类模型。本申请实施例中对该卷积神经网络模型的层数、神经元的数量、卷积核和/或权重等网络参数不作限定。
通过采用不同类别的服饰图片样本对该卷积神经网络模型进行训练，得到 分类模型。其中，用于训练该卷积神经网络模型的服饰图片样本包括服装单品及饰品单品图片，包括但不限于上衣图片、裤子图片、裙子图片、鞋子图片、帽子图片、围巾图片、领带图片、提包图片、皮带图片、首饰图片。
示例性的，可以是通过网络爬虫从网络平台下载服饰单品图片，其中，网络平台中的服饰单品图片的来源可以是购物网站、电子版的时尚杂志或用户上传的服饰单品图片等，并将所下载的服饰单品图片进行分类。分类依据可以是基于服饰性别(即受众为男士还是女士)、服饰风格(即为校园、复古、办公还是休闲等)、服饰款式(包括长袖、短袖、长裤、短裤等结构、流行元素及服饰质地或材质等)及功能信息(包括常规服饰还是功能性服饰)。例如，将上衣图片按照女士T恤、男士T恤、女士衬衣、男士衬衣、女士西装上衣、男士西装上衣、女士风衣、男士风衣、女士大衣、男士大衣、豹纹、碎花、校园及办公等进行分类。将裤子图片按照女士长裤、男士长裤、女士短裤、男士短裤、校园、复古及质地等进行分类。可以根据实际需求对裙子图片、鞋子图片、首饰图片等服饰图片进行分类，分类方式与上述上衣图片及裤子图片的分类方式类似，此处不再一一列举。然后，将分类好得服饰图片与其分类信息关联存储为服饰图片样本。
为了便于理解，下面对基于服饰图片样本训练卷积神经网络模型的过程进行说明。
首先，进行模型构建，即预先设置隐藏层的数目及输入层、隐藏层和输出层中每层的节点数，以及初始化卷积神经网络的第一参数，其中，第一参数包括每层的偏置值及每个边的权重，初步得到卷积神经网络模型的框架。
然后，利用服饰图片样本对该预先设置的卷积神经网络模型进行前向传播和后向传播两个阶段的训练。在后向传播训练计算得到的误差达到期望误差值时，训练结束，并得到分类模型。在一实施例中，可以是根据服饰图片样本，采用前向传播算法及后向传播算法训练该卷积神经网络模型，学习出神经网络模型的第二参数，其中，第二参数是计算服饰图片样本的实际输出与期望输出的偏差，根据该偏差采用后向传播算法计算得到的用于调整第一参数的修正值。采用第二参数更新第一参数，继续执行上述训练过程。然后，计算模型误差，其中，模型误差可以根据服饰图片样本的实际输出与期望输出的偏差确定，在所述模型误差达到期望误差值时，训练结束，得到分类模型。
在分类模型训练完成后，响应移动终端的下载请求，将该分类模型移植到 移动终端内。在一实施例中，由于服务器与移动终端的运算能力存在较大差异，在将分类模型移植到移动终端之前，还要对分类模型进行优化。示例性的，采用预设的优化策略对所述卷积神经网络模型进行优化，其中，对所述卷积神经网络模型的优化包括内部网络结构优化、卷积层的实现方式优化、池化层的实现方式优化中的至少一项。例如，增加残差块构建残差神经网络模型，或者调整残差块的结构。又如，对于卷积层的实现方式的优化可以是减少输出通道和输入通道的连接数量，即输出通道不再和所有输入通道有关，只和相邻的输入通道相关。又如，在卷积层的实现上增加基层，将卷积分为两个步骤：首先，输入的每一通道单独运算，在同样尺寸卷积核的作用下，每一通道得到中间计算结果，将中间计算结果的每一个通道称为一个基层；然后，将多个通道进行合并，得到卷积层的输出结果。又如，通过图像压缩系数设计池化层中用于图像压缩的矩阵。
移动终端若获取到用户在拍摄界面下输入的服饰搭配功能的启动操作，启用服饰搭配功能，以便于根据用户指定的拍摄图片提供与该拍摄图片内的服饰单品的搭配建议。在服饰搭配功能启用后，将已获取的拍摄图片输入分类模型，可以快速并准确的识别该拍摄图片中服饰单品的类别，得到其类别信息。其中，类别信息包括服饰单品对应的性别、风格、款式及功能信息等。由于通过分类模型隐式地对图像的特征进行提取，相比于相关技术中人工设计的特征提取更加快速和准确。
步骤130、根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议。
其中，服饰搭配建议包括提供与拍摄照片中的服饰单品匹配的套装或饰品等的建议信息，包括但不限于与拍摄图片中的服饰单品匹配的搭配服饰的文字描述、与拍摄图片中的服饰单品匹配的搭配服饰图片。
示例性的，根据拍摄图片中服饰单品的类别信息查询预先配置于移动终端内的搭配数据库，确定与该类别信息相符的搭配建议。其中，搭配建议包括一条条的搭配记录。例如，针对运动型女士上衣的搭配记录包括搭配一条棉质的运动半身裙；以及，针对运动型女士上衣的搭配记录还可以是搭配一条棉质哈伦裤；或者，针对运动型女士上衣搭配一条功能型速干裤等。针对女士西装的搭配记录包括搭配一条工装西裤；以及，针对女士西装的搭配记录还可以包括搭配一条西装面料的短裙；或者针对女士西装的搭配记录还可以包括搭配一条西装短裤等。在一实施例中，上述搭配记录是通过对网络平台中的服饰搭配图 片进行统计、分析处理，抽象出的服饰搭配策略。在一实施例中，通过服务器预先构建搭配数据库，并将构建好的搭配数据库下发给移动终端。例如，服务器通过网络爬虫从购物网站、时尚杂志或服装发布会等网络平台下载服饰搭配图片。采用大数据分析的手段分析所下载的服饰搭配图片，抽象出服装及饰品的搭配策略，以搭配记录的形式关联存储服饰单品的类别信息与搭配服饰的特征信息，并将搭配记录存储于搭配数据库中。
在生成服饰搭配建议后，可以通过弹出对话框的方式提示用户与该服饰单品搭配的服装或饰品的特征信息，以便于用户根据该特征信息找寻与之匹配的商品。图2是本申请提供的一种服饰搭配推荐界面示意图。如图2所示，在用户启用了服饰搭配功能后，按下拍照按钮220对悬挂的女士圆领长袖纯色棉衬衣进行拍照，得到一张拍摄图片210。移动终端在获取到该拍摄图片210后，自动将该拍摄图片210输入预置的分类模型中，通过分类模型识别出该上衣的类别信息包括：女士、圆领、长袖衬衣、棉质及纯色。在获取该上衣的类别信息后，调用预设的搭配数据库，根据该类别信息查询到对应的搭配建议包括但不限于a、搭配一件纯色V领毛衣；b、搭配一条裤子A；c、搭配一件裙子B；d、搭配一顶帽子C；e、搭配一个单肩挎包D；f、搭配一支手表E。以弹出对话框230的形式展示上述搭配建议。
在一实施例中，移动终端将所述搭配建议发送至云端服务器，所述搭配建议用于指示所述云端服务器查询与所述搭配建议匹配的商品的位置信息；获取所述云端服务器返回的与所述搭配建议匹配的商品的位置信息，并在电子地图上显示所述位置信息。在一实施例中，若用户当前位于商场中，可以通过移动终端连接云端服务器(例如商场服务器或第三方服务器)。移动终端将该弹出对话框中的搭配服饰的特征信息发送至云端服务器，以通过云端服务器查询本商场内或者临近的其它商场内是否提供与该特征信息匹配的商品信息。在查询到附近存在与该特征信息匹配的商品信息时，将商家位置及店铺名称推送至移动终端。示例性的，移动终端在获取到云端服务器发送的与弹出对话框中的搭配服饰的特征信息匹配的商品的位置信息时，调用地图应用程序，在地图中标记该商品的位置信息及店铺名称。
本实施例的技术方案，通过在拍摄界面中启用服饰搭配功能，将获取到的以服饰单品为对象的拍摄图片输入预先配置的分类模型，以识别该拍摄图片中该服饰单品的类别信息；根据该类别信息调用预设的搭配数据库，生成服饰搭 配建议，并展示该服饰搭配建议。采用上述技术方案，根据分类模型对拍摄图片进行快速且准确的识别，确定拍摄图片中服饰单品的类别，再根据该类别确定与之匹配的搭配建议，实现基于用户输入的服饰单品，及时地提供专业的搭配建议，缩短用户在购物搭配上花费的时间。
图3是本申请实施例提供的另一种服饰搭配推荐方法的流程图。如图3所示，该方法包括：
步骤310、获取以服饰单品为拍摄对象的拍摄图片。
步骤320、获取服饰搭配功能的状态信息。
其中，服饰搭配功能用于提供服饰搭配建议，该功能可以被预先配置于移动终端内，并根据用户输入的操作指示启用或禁用。例如，在移动终端的相机应用中集成该服饰搭配功能，若用户打开相机应用，则可以通过点击设定界面开关启用或禁用该服饰搭配功能。
其中，状态信息包括启用状态及禁用状态，并且在设定界面开关处于打开状态的情况下，确定该服饰搭配功能的状态信息为启用状态，在设定界面开关处关闭状态的情况下，确定该服饰搭配功能的状态信息为禁用状态。
步骤330、根据所述状态信息判断所述服饰搭配功能是否被启用，若所述服饰搭配功能没有被启用，则执行步骤340，若所述服饰搭配功能已被启用，则执行步骤350。
根据设定界面开关的状态判断服饰搭配功能是否被启用。在检测到用户打开该设定界面开关的情况下，确定服饰搭配功能被启用，执行步骤350。在检测到用户关闭该设定界面开关的情况下，确定服饰搭配功能被禁用，执行步骤340。
步骤340、将拍摄图片存储于图片库。
步骤350、将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息。
步骤360、根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的搭配建议。
其中，类别信息包括性别、风格、款式及功能信息等。示例性的，根据该服饰单品的性别、风格、款式及功能信息查询预设的搭配数据库，确定与所述服饰单品对应的搭配建议。由于预设的搭配数据库中关联存储有与类别信息对应的搭配记录，因此，根据拍摄图片中服饰单品的类别信息查询该预设的搭配数据库可以得到与该服饰单品匹配的搭配服饰的特征信息，该特征信息构成了 搭配建议。例如，通过分类模型确定拍摄图片中服饰单品的类别信息包括女士、雪纺连衣裙、圆领、碎花等。基于该类别信息查询预设的搭配数据库确定与之匹配的搭配建议包括但不限于：a、搭配一件纯色长款毛衣S；b、搭配一件黑色皮衣；c、搭配一件黑色靴子；d、搭配一件纯色外套R；e、搭配一条格子围巾；f、搭配一个红色手提包。上述搭配建议以文字方式存储于预设搭配数据库中采用文字记录方式可以缩小预设的搭配数据库的数据量，避免过度占用移动终端的存储空间。
在一实施例中，在移动终端根据类别信息查询预设的搭配数据库的情况下，调用预设的搭配数据库，确定与该类别信息匹配的搭配建议，采用默认的排序以文字的形式将该搭配建议显示在展示窗口中。
在一实施例中，为了更直观的展示该搭配建议，还可以在确定与该类别信息匹配的搭配建议之后，获取网络平台中符合该搭配建议的商品图片，并以展示窗口的形式显示所述商品图片。例如，以拍摄图片中服饰单品的类别信息包括女士、雪纺连衣裙、圆领、碎花为例，可以基于上述搭配建议查询网络平台，确定符合上述搭配建议的搭配服饰的商品图片，采用展示窗口的形式滚动显示该商品图片。在一实施例中，还可以检测用户的点击操作，响应该点击操作提供被选中的商品图片的详细信息。
步骤370、将所述搭配建议发送至云端服务器。
移动终端与云端服务器建立通信连接，并将搭配建议发送至云端服务器。该云端服务器设置为根据该搭配建议确定与该搭配建议匹配的商品的位置信息，并在电子地图上标注该位置信息，并设计在每个商品位置上显示该商品的缩略图。
步骤380、获取所述云端服务器返回的与所述搭配建议匹配的商品的位置信息。
由云端服务器获取与搭配建议匹配的商品的位置信息，并调用电子地图应用程序显示该位置信息，并在该位置信息对应的位置显示该商品的缩略图。
步骤390、获取用户针对所述搭配建议的反馈信息。
其中，反馈信息包括用户针对搭配建议的采纳情况，包括但不限于用户输入的是否采纳搭配建议的采纳信息，例如，若用户选中某一商品图片，以继续显示该商品图片的详细信息，则认为用户采纳了该条搭配建议。又如，在电子地图上显示与搭配建议匹配的商品的位置信息的情况下，若检测到用户输入路 径规划指示，该路径规划指示的目的地是该商品对应的位置，则认为用户采纳了该条搭配建议。
在一实施例中，获取用户针对所述搭配建议的反馈信息，包括：获取用户输入的针对搭配建议的采纳信息，作为反馈信息。其中，所述采纳信息包括用户选中的任一以展示窗口的形式显示的所述商品图片，以及用户输入指示商品位置信息的路径规划。
步骤3100、根据所述反馈信息确定用户的搭配喜好，并根据所述搭配喜好更新所述搭配数据库中服饰搭配建议的推荐优先级。
移动终端对一定时间段内的用户的反馈信息进行分析，归纳处理，确定用户的搭配喜好，根据该搭配喜欢调整搭配数据库中搭配建议的推荐优先级。例如，基于对一个月内的用户反馈信息的分析、归纳，确定用户喜好休闲类服饰，则根据该喜好为搭配数据库内相匹配的搭配建议设置较高的推荐优先级。还可以是，基于对一个月内的用户反馈信息的分析、归纳，确定用户喜欢上衣搭配半身裙，则可以为搭配数据库内的半身裙配置较高的推荐优先级。在一实施例中，还可以根据搭配建议的历史采纳记录，确定用户喜好，根据用户喜好进一步为搭配建议排序，从而，根据排序结果为搭配建议分配不同优先级。
本实施例的技术方案，通过获取以服饰单品为对象的拍摄图片；并将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息；根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的搭配建议；获取网络平台中符合所述搭配建议的商品图片，并以展示窗口的形式按照用户的搭配喜好显示所述商品图片，即与用户搭配喜好越接近的商品，排序越靠前，实现直观、形象的展示搭配建议，便于用户基于该搭配建议进行购物，节省了用户搭配服饰的时间。另外，将所述搭配建议发送至云端服务器，从而可以通过电子地图规划前往该商品的位置的导航路径，从而节省了用户的购物时间。
图4是本申请实施例提供的又一种服饰搭配推荐方法的流程图，该方法包括：
步骤410、获取以服饰单品为拍摄对象的拍摄图片。
步骤420、获取服饰搭配功能的状态信息。
步骤430、根据所述状态信息判断所述服饰搭配功能是否被启用，若所述服 饰搭配功能没有被启用，则执行步骤440，若所述服饰搭配功能已被启用，则执行步骤450。
步骤440、将拍摄图片存储于图片库。
步骤450、将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息。
步骤460、根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的服饰搭配图片。
其中，预设的搭配数据库中存储服饰搭配图片。
通过服务器预先构建搭配数据库，并将构建好的搭配数据库下发给移动终端。例如，服务器通过网络爬虫从购物网站、时尚杂志或服装发布会等网络平台下载服饰搭配图片。对服饰搭配图片进行一系列图像处理后存储于预设的搭配数据库中，并对应标注每个服饰搭配图片包括的服饰单品的类别信息。其中，图像处理包括但不限于对于仅颜色不同的服饰搭配图片进行去重处理，对于区分度不高的图片，进行保留清晰度较高的服饰搭配图片的处理等。
由于搭配数据库中类别信息与服饰搭配图片是关联存储的，在识别出拍摄图片中服饰单品的类别信息后，可以根据类别信息查询搭配数据库，确定与该类别信息匹配的服饰搭配图片。
步骤470、根据所述服饰搭配图片中与所述服饰单品搭配的服装和/或饰品生成服饰搭配建议，并以展示窗口的形式展示所述服饰搭配建议。
示例性的，如果拍摄图片中的服饰单品为女士雪纺碎花连衣裙，则根据该服饰搭配图片中除该女士雪纺碎花连衣裙之外的服装及饰品生成服饰搭配建议，并以展示窗口的形式展示该服饰搭配建议。在一实施例中，还可以根据用户点击操作，显示一个或多个服饰单品的详细信息。
步骤480、将所述搭配建议发送至云端服务器。
在一实施例中，当检测到用户针对所显示的服饰搭配建议的点击操作时，将点击操作对应的搭配建议发送至云端服务器。示例性的，若用户点击黑色皮衣，则将黑色皮衣的图片发送至云端服务器，以使云端服务器根据该图片搜寻该商品或相似商品的位置。
步骤490、获取所述云端服务器返回的与所述搭配建议匹配的商品的位置信息。
本实施例的技术方案，通过获取以服饰单品为对象的拍摄图片；并将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息；根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的服饰搭配图片；根据所述服饰搭配图片中与所述服饰单品搭配的服装和/或饰品生成服饰搭配建议，并以展示窗口的形式展示所述服饰搭配建议，实现在移动终端的搭配数据库中查询与服饰单品匹配的搭配服饰的图片，不依赖于网络，可以迅速的显示服饰搭配图片。另外，将所述搭配建议发送至云端服务器，从而可以通过电子地图规划前往该商品的位置的导航路径，从而节省了用户的购物时间。
图5是本申请实施例提供的一种服饰搭配推荐装置，该装置可由软件和/或硬件实现，一般可被集成在移动终端内，用于执行本申请实施例提供的服饰搭配推荐方法。如图5所示，该装置包括：
图片拍摄模块510，设置为获取以服饰单品为拍摄对象的拍摄图片；
服饰识别模块520，设置为将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息，其中，所述分类模型为根据不同类别的服饰图片样本训练的深度学习模型；
建议生成模块530，设置为根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议。
本实施例的技术方案提供一种服饰搭配推荐装置，根据分类模型对拍摄图片进行快速且准确的识别，确定拍摄图片中服饰单品的类别，再根据该类别确定与之匹配的搭配建议，实现基于用户输入的服饰单品，及时地提供专业的搭配建议，缩短用户在购物搭配上花费的时间。
在一实施例中，该装置还包括：
状态获取模块，设置为在将所述拍摄图片输入预先配置的分类模型之前，获取服饰搭配功能的状态信息，其中，所述服饰搭配功能用于提供服饰搭配建议；
其中，所述服饰识别模块520是设置为：
若根据所述状态信息确定所述服饰搭配功能已启用，则执行将所述拍摄图片输入预先配置的分类模型的操作。
在一实施例中，该装置可以包括功能判定模块，设置为根据所述状态信息判断所述服饰搭配功能是否被启用；
若所述服饰搭配功能已启用，则服饰识别模块520设置为执行将所述拍摄图片输入预先配置的分类模型的操作。
在一实施例中，建议生成模块530是设置为：
根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的搭配建议；
获取网络平台中符合所述搭配建议的商品图片，并以展示窗口的形式显示所述商品图片。
在一实施例中，建议生成模块530是设置为：
根据所述服饰单品的性别、风格、款式及功能信息查询预设的搭配数据库，确定与所述服饰单品对应的搭配建议。
在一实施例中，该装置还包括：
反馈信息获取模块，设置为在以展示窗口的形式显示所述商品图片之后，获取用户针对所述搭配建议的反馈信息；
优先级调整模块，设置为根据所述反馈信息确定所述用户的搭配喜好，并根据所述搭配喜好更新所述搭配数据库中服饰搭配建议的推荐优先级。
在一实施例中，建议生成模块530是设置为：
根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的服饰搭配图片；
根据所述服饰搭配图片中与所述服饰单品搭配的服装和/或饰品生成服饰搭配建议，并以展示窗口的形式展示所述服饰搭配建议。
在一实施例中，该装置还包括：
建议发送模块，设置为在展示所述服饰搭配建议之后，将所述搭配建议发送至云端服务器，所述搭配建议用于指示所述云端服务器查询与所述搭配建议匹配的商品的位置信息；
位置获取模块，设置为获取所述云端服务器返回的与所述搭配建议匹配的商品的位置信息，并显示所述位置信息。
在一实施例中，所述反馈信息获取模块是设置为：
获取用户输入的针对搭配建议的采纳信息，作为反馈信息；
其中，所述采纳信息包括用户选中的任一以展示窗口的形式显示的所述商品图片，以及用户输入指示商品位置信息的路径规划。
在一实施例中，所述拍摄图片包括用户当前拍摄的图片或从图片库获取的 历史拍摄图片。
本申请实施例还提供一种包含计算机可执行指令的存储介质，所述计算机可执行指令在由计算机处理器执行时用于执行一种服饰搭配推荐方法，该方法包括：
获取以服饰单品为拍摄对象的拍摄图片；
将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息，其中，所述分类模型为根据不同类别的服饰图片样本训练的深度学习模型；
根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议。
存储介质——任何类型的存储器设备或存储设备。术语“存储介质”旨在包括：安装介质，例如光盘只读存储器(Compact Disc Read-Only Memory，CD-ROM)、软盘或磁带装置；计算机系统存储器或随机存取存储器，诸如动态随机存取存储器(Dynamic Random Access Memory，DRAM)、(Double Data Rate Random Access Memory，DDR RAM)、静态随机存取存储器(Static Random-Access Memory，SRAM)、扩展数据输出随机存取存储器(Extended Data Output Random Access Memory，EDO RAM)，兰巴斯随机存取存储器(Rambus Random Access Memory，Rambus RAM)等；非易失性存储器，诸如闪存、磁介质(例如硬盘或光存储)；寄存器或其它相似类型的存储器元件等。存储介质可以还包括其它类型的存储器或其组合。另外，存储介质可以位于程序在其中被执行的第一计算机系统中，或者可以位于不同的第二计算机系统中，第二计算机系统通过网络(诸如因特网)连接到第一计算机系统。第二计算机系统可以提供程序指令给第一计算机用于执行。术语“存储介质”可以包括可以驻留在不同位置中(例如在通过网络连接的不同计算机系统中)的两个或更多存储介质。存储介质可以存储可由一个或多个处理器执行的程序指令(例如计算机程序)。
当然，本申请实施例所提供的一种包含计算机可执行指令的存储介质，其计算机可执行指令不限于如上所述的服饰搭配推荐操作，还可以执行本申请任意实施例所提供的在服饰搭配推荐方法中的相关操作。
本申请实施例提供了一种移动终端，该移动终端内具有操作系统，该移动终端中可集成本申请实施例提供的服饰搭配推荐装置。其中，移动终端可以为智能手机、平板电脑(Portable Device，PAD)、智能可穿戴设备等。图6是本申请实施例提供的一种移动终端的结构框图。如图6所示，该移动终端可以包括 存储器610和处理器620。所述存储器610，设置为存储计算机程序、搭配数据库及分类模型等；所述处理器620读取并执行所述存储器610中存储的计算机程序。所述处理器620在执行所述计算机程序时实现以下步骤：获取以服饰单品为拍摄对象的拍摄图片；将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息，其中，所述分类模型为根据不同类别的服饰图片样本训练的深度学习模型；根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议。
上述示例中列举的存储器及处理器均为移动终端的部分元器件，所述移动终端还可以包括其它元器件。以智能手机为例，说明上述移动终端可能的结构。图7是本申请实施例提供的一种智能手机的结构框图。如图7所示，该智能手机可以包括：存储器701、中央处理器(Central Processing Unit，CPU)702(又称处理器)、外设接口703、射频(Radio Frequency，RF)电路705、音频电路706、扬声器711、电源管理芯片708、输入/输出(Input/Output，I/O)子系统709、其他输入/控制设备710以及外部端口704，这些部件通过一个或多个通信总线或信号线707来通信。
图示智能手机700仅仅是移动终端的一个范例，并且智能手机700可以具有比图中所示出的更多的或者更少的部件，可以组合两个或更多的部件，或者可以具有不同的部件配置。图中所示出的每种部件可以在包括一个或多个信号处理和/或专用集成电路在内的硬件、软件、或硬件和软件的组合中实现。
下面就本实施例提供的集成有服饰搭配推荐装置的智能手机进行详细的描述。
存储器701，所述存储器701可以被CPU702、外设接口703等访问，所述存储器701可以包括高速随机存取存储器，还可以包括非易失性存储器，例如一个或多个磁盘存储器件、闪存器件、或其他易失性固态存储器件。在存储器701中存储计算机程序，还可以搭配数据库及分类模型等。
外设接口703，所述外设接口703可以将设备的输入和输出外设连接到CPU702和存储器701。
I/O子系统709，所述I/O子系统709可以将设备上的输入输出外设，例如触摸屏712和其他输入/控制设备710，连接到外设接口703。I/O子系统709可以包括显示控制器7071和用于控制其他输入/控制设备710的一个或多个输入控制器7092。其中，一个或多个输入控制器7092从其他输入/控制设备710接收 电信号或者向其他输入/控制设备710发送电信号，其他输入/控制设备710可以包括物理按钮(按压按钮、摇臂按钮等)、拨号盘、滑动开关、操纵杆、点击滚轮。值得说明的是，输入控制器7092可以与以下任一个连接：键盘、红外端口、通用串行总线(Universal Serial Bus，USB)接口以及诸如鼠标的指示设备。
触摸屏712，所述触摸屏712是用户终端与用户之间的输入接口和输出接口，将可视输出显示给用户，可视输出可以包括图形、文本、图标、视频等。
I/O子系统709中的显示控制器7071从触摸屏712接收电信号或者向触摸屏712发送电信号。触摸屏712检测触摸屏上的接触，显示控制器7071将检测到的接触转换为与显示在触摸屏712上的用户界面对象的交互，即实现人机交互，显示在触摸屏712上的用户界面对象可以是运行游戏的图标、联网到相应网络的图标等。在一实施例中，设备还可以包括光鼠，光鼠是不显示可视输出的触摸敏感表面，或者是由触摸屏形成的触摸敏感表面的延伸。
RF电路705，设置为建立手机与无线网络(即网络侧)的通信，实现手机与无线网络的数据接收和发送。例如收发短信息、电子邮件等。在一实施例中，RF电路705接收并发送RF信号，RF信号也称为电磁信号，RF电路705将电信号转换为电磁信号或将电磁信号转换为电信号，并且通过该电磁信号与通信网络以及其他设备进行通信。RF电路705可以包括用于执行这些功能的已知电路，其包括但不限于天线系统、RF收发机、一个或多个放大器、调谐器、一个或多个振荡器、数字信号处理器、编译码器(COder-DECoder，CODEC)芯片组、用户标识模块(Subscriber Identity Module，SIM)等等。
音频电路706，设置为从外设接口703接收音频数据，将该音频数据转换为电信号，并且将该电信号发送给扬声器711。
扬声器711，设置为将手机通过RF电路705从无线网络接收的语音信号，还原为声音并向用户播放该声音。
电源管理芯片708，设置为为CPU702、I/O子系统及外设接口所连接的硬件进行供电及电源管理。
本申请实施例提供的移动终端，可以基于用户输入的服饰单品，及时地提供专业的搭配建议，缩短用户在购物搭配上花费的时间。
上述实施例中提供的服饰搭配推荐装置、存储介质及移动终端可执行本申请任意实施例所提供的服饰搭配推荐方法，具备执行该方去相应的功能模块和效果。未在上述实施例中详尽描述的技术细节，可参见本申请任意实施例所提供的服饰搭配推荐方法。

Claims (20)

  1. 一种服饰搭配推荐方法，包括：
    获取以服饰单品为拍摄对象的拍摄图片；
    将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息，其中，所述分类模型为根据不同类别的服饰图片样本训练的深度学习模型；
    根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议。
  2. 根据权利要求1所述的方法，在将所述拍摄图片输入预先配置的分类模型之前，还包括：
    获取服饰搭配功能的状态信息，其中，所述服饰搭配功能用于提供服饰搭配建议；
    其中，所述将所述拍摄图片输入预先配置的分类模型，包括：
    若根据所述状态信息确定所述服饰搭配功能已启用，则执行将所述拍摄图片输入预先配置的分类模型的操作。
  3. 根据权利要求1所述的方法，其中，根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议，包括：
    根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的搭配建议；
    获取网络平台中符合所述搭配建议的商品图片，并以展示窗口的形式显示所述商品图片。
  4. 根据权利要求3所述的方法，其中，根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的搭配建议，包括：
    根据所述服饰单品的性别、风格、款式及功能信息查询预设的搭配数据库，确定与所述服饰单品对应的搭配建议。
  5. 根据权利要求3所述的方法，在以展示窗口的形式显示所述商品图片之后，还包括：
    获取用户针对所述搭配建议的反馈信息；
    根据所述反馈信息确定所述用户的搭配喜好，并根据所述搭配喜好更新所述搭配数据库中服饰搭配建议的推荐优先级。
  6. 根据权利要求1所述的方法，其中，根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议，包括：
    根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的服饰搭配图片；
    根据所述服饰搭配图片中与所述服饰单品搭配的服装和饰品中的至少一种生成服饰搭配建议，并以展示窗口的形式展示所述服饰搭配建议。
  7. 根据权利要求1至6中任一项所述的方法，在展示所述服饰搭配建议之后，还包括：
    将所述搭配建议发送至云端服务器，其中，所述搭配建议用于指示所述云端服务器查询与所述搭配建议匹配的商品的位置信息；
    获取所述云端服务器返回的与所述搭配建议匹配的商品的位置信息，并显示所述位置信息。
  8. 根据权利要求5所述的方法，其中，获取用户针对所述搭配建议的反馈信息，包括：
    获取用户输入的针对搭配建议的采纳信息，作为反馈信息；
    其中，所述采纳信息包括用户选中的任一以展示窗口的形式显示的所述商品图片，以及用户输入指示商品位置信息的路径规划。
  9. 根据权利要求1-8中任一项所述的方法，其中，所述拍摄图片包括用户当前拍摄的图片或从图片库获取的历史拍摄图片。
  10. 一种服饰搭配推荐装置，包括：
    图片拍摄模块，设置为获取以服饰单品为拍摄对象的拍摄图片；
    服饰识别模块，设置为将所述拍摄图片输入预先配置的分类模型，以识别所述拍摄图片中所述服饰单品的类别信息，其中，所述分类模型为根据不同类别的服饰图片样本训练的深度学习模型；
    建议生成模块，设置为根据所述类别信息生成服饰搭配建议，展示所述服饰搭配建议。
  11. 根据权利要求10所述的装置，还包括：
    状态获取模块，设置为获取服饰搭配功能的状态信息，其中，所述服饰搭配功能用于提供服饰搭配建议；
    其中，所述服饰识别模块是设置为：
    若根据所述状态信息确定所述服饰搭配功能已启用，则执行将所述拍摄图片输入预先配置的分类模型的操作。
  12. 根据权利要求10所述的装置，其中，所述建议生成模块是设置为：
    根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的搭配建议；
    获取网络平台中符合所述搭配建议的商品图片，并以展示窗口的形式显示所述商品图片。
  13. 根据权利要求12所述的装置，其中，所述建议生成模块是设置为：
    根据所述服饰单品的性别、风格、款式及功能信息查询预设的搭配数据库，确定与所述服饰单品对应的搭配建议。
  14. 根据权利要求12所述的装置，还包括：
    反馈信息获取模块，设置为在以展示窗口的形式显示所述商品图片之后，获取用户针对所述搭配建议的反馈信息；
    优先级调整模块，设置为根据所述反馈信息确定所述用户的搭配喜好，并根据所述搭配喜好更新所述搭配数据库中服饰搭配建议的推荐优先级。
  15. 根据权利要求10所述的装置，其中，所述建议生成模块是设置为：
    根据所述类别信息查询预设的搭配数据库，确定与所述类别信息匹配的服饰搭配图片；
    根据所述服饰搭配图片中与所述服饰单品搭配的服装和饰品中的至少一种生成服饰搭配建议，并以展示窗口的形式展示所述服饰搭配建议。
  16. 根据权利要求10-15中任一项所述的装置，还包括：
    建议发送模块，设置为将所述搭配建议发送至云端服务器，其中，所述搭配建议用于指示所述云端服务器查询与所述搭配建议匹配的商品的位置信息；
    位置获取模块，设置为获取所述云端服务器返回的与所述搭配建议匹配的商品的位置信息，并显示所述位置信息。
  17. 根据权利要求14所述的装置，其中，所述反馈信息获取模块是设置为：
    获取用户输入的针对搭配建议的采纳信息，作为反馈信息；
    其中，所述采纳信息包括用户选中的任一以展示窗口的形式显示的所述商品图片，以及用户输入指示商品位置信息的路径规划。
  18. 根据权利要求10-17中任一项所述的装置，其中，所述拍摄图片包括用户当前拍摄的图片或从图片库获取的历史拍摄图片。
  19. 一种计算机可读存储介质，其上存储有计算机程序，该程序被处理器执行时实现如权利要求1-9中任一项所述的服饰搭配推荐方法。
  20. 一种移动终端，包括存储器，处理器及存储在存储器上并可在处理器运行的计算机程序，所述处理器执行所述计算机程序时实现如权利要求1-9中任一项所述的服饰搭配推荐方法。
PCT/CN2018/116799 2017-12-20 2018-11-21 服饰搭配推荐方法、装置、存储介质及移动终端 WO2019120031A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201711382746.3A CN109949116A (zh) 2017-12-20 2017-12-20 服饰搭配推荐方法、装置、存储介质及移动终端
CN201711382746.3 2017-12-20

Publications (1)

Publication Number Publication Date
WO2019120031A1 true WO2019120031A1 (zh) 2019-06-27

Family

ID=66992475

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2018/116799 WO2019120031A1 (zh) 2017-12-20 2018-11-21 服饰搭配推荐方法、装置、存储介质及移动终端

Country Status (2)

Country Link
CN (1) CN109949116A (zh)
WO (1) WO2019120031A1 (zh)

Families Citing this family (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111161022A (zh) * 2019-12-26 2020-05-15 汉口北进出口服务有限公司 一种商品推荐方法及装置
CN113159876B (zh) * 2020-01-21 2023-08-22 海信集团有限公司 服装搭配推荐装置、方法及存储介质
CN112069358B (zh) * 2020-08-18 2022-03-25 北京达佳互联信息技术有限公司 信息推荐方法、装置及电子设备
CN112307242A (zh) * 2020-11-11 2021-02-02 百度在线网络技术（北京）有限公司 服饰搭配方法及装置、计算设备和介质
CN113435941A (zh) * 2021-07-08 2021-09-24 武汉纺织大学 一种服装搭配推荐方法及装置、可读存储介质
CN113468353A (zh) * 2021-07-20 2021-10-01 柒久园艺科技(北京)有限公司 一种基于图形的游客互动方法、装置、电子设备及介质
CN114820135A (zh) * 2022-05-16 2022-07-29 温州鞋革产业研究院 一种服饰智能搭配系统和方法

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104331417A (zh) * 2014-10-09 2015-02-04 深圳码隆科技有限公司 一种用户个人服饰的搭配方法
CN104951966A (zh) * 2015-07-13 2015-09-30 百度在线网络技术（北京）有限公司 推荐服饰商品的方法及装置
CN106504064A (zh) * 2016-10-25 2017-03-15 清华大学 基于深度卷积神经网络的服装分类与搭配推荐方法及系统

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN105787574A (zh) * 2014-12-26 2016-07-20 阿里巴巴集团控股有限公司 提供座位信息的方法及装置
CN105279679A (zh) * 2015-11-17 2016-01-27 小米科技有限责任公司 服饰搭配方法及装置
US10346893B1 (en) * 2016-03-21 2019-07-09 A9.Com, Inc. Virtual dressing room
CN107123033A (zh) * 2017-05-04 2017-09-01 北京科技大学 一种基于深度卷积神经网络的服装搭配方法

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104331417A (zh) * 2014-10-09 2015-02-04 深圳码隆科技有限公司 一种用户个人服饰的搭配方法
CN104951966A (zh) * 2015-07-13 2015-09-30 百度在线网络技术（北京）有限公司 推荐服饰商品的方法及装置
CN106504064A (zh) * 2016-10-25 2017-03-15 清华大学 基于深度卷积神经网络的服装分类与搭配推荐方法及系统

Also Published As

Publication number Publication date
CN109949116A (zh) 2019-06-28

Similar Documents

Publication Publication Date Title
WO2019120031A1 (zh) 服饰搭配推荐方法、装置、存储介质及移动终端
WO2019134560A1 (zh) 搭配模型构建方法、服饰推荐方法、装置、介质及终端
US10963939B1 (en) Computer vision based style profiles
CN106462979B (zh) 时尚偏好分析
US10109051B1 (en) Item recommendation based on feature match
US10346893B1 (en) Virtual dressing room
US10083521B1 (en) Content recommendation based on color match
US20200342320A1 (en) Non-binary gender filter
CN107665238B (zh) 图片处理方法和装置、用于图片处理的装置
US11132734B2 (en) System and method for social style mapping
US11836963B2 (en) Computer vision assisted item search
CN111295669A (zh) 图像处理系统
US10776417B1 (en) Parts-based visual similarity search
US11586666B2 (en) Feature-based search
CN108363750B (zh) 服装推荐方法及相关产品
WO2021102655A1 (zh) 网络模型训练方法、图像属性识别方法、装置及电子设备
US20190287163A1 (en) Method and device for generating object coordination information
CN112905889A (zh) 服饰搜索方法及装置、电子设备和介质
US11972466B2 (en) Computer storage media, method, and system for exploring and recommending matching products across categories
KR20210130953A (ko) 딥러닝 기반 가상 이미지 생성방법 및 시스템
US10771544B2 (en) Online fashion community system and method
WO2019134501A1 (zh) 模拟用户试装的方法、装置、存储介质及移动终端
KR102476884B1 (ko) 크리에이터 매칭 서비스를 통해 의류 정보를 추천하는 서버의 제어 방법
KR102344818B1 (ko) 가상 옷장 구축 시스템 및 그 제어 방법
US20230059006A1 (en) Fashion product recommendation method, apparatus, and system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18891949

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18891949

Country of ref document: EP

Kind code of ref document: A1