Wearing and building recommendation method, electronic equipment, storage medium and product

Abstract

The invention discloses a wearing and building recommendation method, electronic equipment, a storage medium and a product, and relates to the technical field of artificial intelligence, and the disclosed wearing and building recommendation method comprises the following steps: determining alternative shops from a preset shop database according to geographic position data; according to the commodity inventory data of the alternative shops, establishing an alternative wearing and building library; and through a preset wearing and building recommendation large model, determining recommended wearing and building from the alternative wearing and building library according to the user feature data and / or the wearing and building information data. According to the method and the device, the offline clothing consumption process can be optimized, so that the shopping experience of the user during offline clothing purchase is improved.

Description

Technical Field

[0001] The present application relates to the field of artificial intelligence technology, and in particular to outfit recommendation methods, electronic devices, storage media, and products. Background Art

[0002] In today's consumer market, physical clothing stores can provide consumers with an intuitive fitting experience, the feel of fabrics, and instant shopping satisfaction.

[0003] However, users also face many problems when purchasing clothing in offline physical stores, such as limited style selection, cumbersome and time-consuming shopping process, and lack of matching guidance, which seriously affect the user's shopping experience. Summary of the Invention

[0004] The main purpose of this application is to provide a clothing recommendation method, electronic device, storage medium and product, aiming to optimize the offline clothing consumption process to improve the user's shopping experience when purchasing clothing offline.

[0005] To achieve the above objectives, this application proposes a method for recommending outfits, which includes:

[0006] Determine alternative stores from a preset store database based on geographic location data;

[0007] Establishing a library of alternative outfits based on the product inventory data of the alternative stores;

[0008] Through the preset outfit recommendation model, the recommended outfit is determined from the alternative outfit library according to user feature data and / or outfit information data.

[0009] In one embodiment, before the step of determining a recommended outfit from the library of candidate outfits based on user feature data and / or outfit information data using a preset outfit recommendation model, the step further includes:

[0010] Acquire a multi-source data set, wherein the multi-source data set includes clothing attribute data, clothing knowledge data, and user behavior data;

[0011] Constructing a clothing feature vector based on the clothing attribute data, constructing a style knowledge graph based on the clothing knowledge data, and constructing an implicit feedback matrix based on the user behavior data;

[0012] Extracting feature data from the clothing feature vector, the style knowledge graph, and the implicit feedback matrix;

[0013] The neural network model to be trained is trained based on the feature data to obtain an outfit recommendation model.

[0014] In one embodiment, the step of determining candidate stores from a preset store database based on geographic location data includes:

[0015] Calculating a dynamic search range based on geographic location data, wherein the geographic location data includes current positioning coordinates and movement speed, and the size of the dynamic search range is proportional to the movement speed;

[0016] A set of stores within the dynamic search range of the current positioning coordinates is determined from a preset store database, and alternative stores are determined from the store set according to preset alternative store screening rules, wherein the alternative store screening rules include screening rules for store business status, store type and / or store product price range.

[0017] In one embodiment, the step of determining a recommended outfit from the library of candidate outfits based on user feature data using a preset outfit recommendation model includes:

[0018] Acquiring user feature data and generating a user feature vector based on the user feature data, wherein the user feature data includes body parameters, style preference characteristics and / or scene adaptation requirements;

[0019] Generate a feature vector for each alternative outfit based on the item label of each item in the alternative outfit library;

[0020] Calculating the matching degree between the user feature vector and each of the candidate outfit feature vectors through the target outfit recommendation model, so as to screen a set of highly matching outfit candidates from the candidate outfit library;

[0021] Generate recommended outfits based on the high-matching outfit candidate set.

[0022] In one embodiment, the step of determining a recommended outfit from the library of candidate outfits based on the outfit information data using a preset outfit recommendation model includes:

[0023] Acquire outfit information data and generate an outfit feature vector based on the outfit information data, wherein the outfit information data is input by scanning, image uploading, and / or text retrieval, and the outfit feature vector includes type, color, version, material, and / or style vectors;

[0024] Generate a feature vector for each alternative outfit based on the item label of each item in the alternative outfit library;

[0025] Calculating the matching degree between the outfit feature vector and each of the candidate outfit feature vectors through the target outfit recommendation model, so as to screen a set of highly matching outfit candidates from the candidate outfit library;

[0026] Generate recommended outfits based on the high-matching outfit candidate set.

[0027] In one embodiment, the step of establishing a library of alternative outfits based on the product inventory data of the candidate stores includes:

[0028] Obtaining product inventory data in real time from the inventory management system of the candidate store;

[0029] The alternative items in the alternative stores are classified according to the product inventory data and the selected classification rules to obtain an alternative outfit library.

[0030] In one embodiment, after the step of determining a recommended outfit from the library of candidate outfits based on user feature data and / or outfit information data using a preset outfit recommendation model, the method further includes:

[0031] Combining each recommended item corresponding to the recommended outfit with a preset user image to obtain a visual matching image;

[0032] When a purchase instruction for the recommended outfit is received, purchase guidance information including the store location and inventory status of each recommended item is generated.

[0033] In addition, to achieve the above-mentioned purpose, the present application also proposes an electronic device, which includes: a memory, a processor, and a computer program stored in the memory and runnable on the processor, and the computer program is configured to implement the steps of the outfit recommendation method as described above.

[0034] In addition, to achieve the above-mentioned purpose, the present application also proposes a storage medium, which is a computer-readable storage medium. A computer program is stored on the storage medium. When the computer program is executed by the processor, the steps of the outfit recommendation method as described above are implemented.

[0035] In addition, to achieve the above-mentioned purpose, the present application also provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements the steps of the outfit recommendation method as described above.

[0036] This application proposes a method for recommending outfits, which determines alternative stores from a preset store database based on geographic location data; establishes an alternative outfit library based on the product inventory data of the alternative stores; and determines recommended outfits from the alternative outfit library based on user feature data and / or outfit information data through a preset outfit recommendation model.

[0037] In summary, in this application, alternative stores are determined based on geographic location data, and an alternative outfit library is established based on the product inventory data of the alternative stores, so that users can quickly obtain clothing resource information from multiple nearby stores, providing users with a wider range of style choices, and avoiding the process of users blindly searching for clothing in the store, saving shopping time. At the same time, through the outfit recommendation model, recommended outfits are quickly provided based on user characteristics and / or outfit information data, which makes up for the problem of insufficient matching guidance in offline physical stores, reduces the time cost of users' self-matching and selection, makes the shopping process more efficient and convenient, and improves the user's shopping experience when purchasing clothing offline. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0039] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0040] Figure 1 A flowchart of the first embodiment of the outfit recommendation method of this application is provided;

[0041] Figure 2 Another flowchart provided for the second embodiment of the outfit recommendation method of this application;

[0042] Figure 3 A schematic diagram of the outfit recommendation process provided in Example 2 of the outfit recommendation method of this application;

[0043] Figure 4 This is a schematic diagram of the structure of the outfit recommendation device according to an embodiment of the present application;

[0044] Figure 5 This is a schematic diagram of the device structure of the hardware operating environment involved in the outfit recommendation method in the embodiment of the present application.

[0045] The purpose, features and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0046] It should be understood that the specific embodiments described herein are merely used to explain the technical solutions of the present application and are not intended to limit the present application.

[0047] In order to better understand the technical solution of the present application, a detailed description will be given below in conjunction with the accompanying drawings and specific implementation methods.

[0048] In today's consumer market, physical clothing stores can provide users with an intuitive fitting experience, the feel of fabrics, and instant shopping satisfaction.

[0049] However, users also face many problems when purchasing clothing in offline physical stores, such as limited style selection, cumbersome and time-consuming shopping process, and lack of matching guidance, which seriously affect the user's shopping experience.

[0050] In summary, how to optimize the offline clothing consumption process to improve users' shopping experience when purchasing clothing offline has become a technical problem that urgently needs to be solved in this field.

[0051] The main solutions of the embodiments of the present application are: determining alternative stores from a preset store database based on geographic location data; establishing an alternative outfit library based on the product inventory data of the alternative stores; and determining recommended outfits from the alternative outfit library based on user feature data and / or outfit information data through a preset outfit recommendation model.

[0052] Therefore, in this embodiment, alternative stores are determined based on geographic location data, and an alternative outfit library is established based on the product inventory data of the alternative stores, so that users can quickly obtain clothing resource information from multiple nearby stores, providing users with a wider range of style choices, and avoiding the process of users blindly searching for clothing in the store, saving shopping time. At the same time, the outfit recommendation model quickly provides recommended outfits based on user characteristics and / or outfit information data, making up for the problem of insufficient matching guidance in offline physical stores, reducing the time cost of users' self-matching and selection, making the shopping process more efficient and convenient, and improving the user's shopping experience when purchasing clothing offline.

[0053] It should be noted that the execution subject of this embodiment can be a computer service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, mobile phone, etc., or an electronic device capable of performing the above functions. The following uses electronic devices as an example to illustrate this embodiment and the following embodiments.

[0054] Based on this, the embodiment of the present application provides a method for recommending outfits. Figure 1 , Figure 1 This is a flow chart of the first embodiment of the outfit recommendation method of this application.

[0055] In this embodiment, the outfit recommendation method includes steps S10 to S30:

[0056] Step S10, determining candidate stores from a preset store database based on the geographic location data;

[0057] Based on the user's current geographic location information, a set of alternative stores that meet the distance conditions or specific geographic range are accurately filtered from the preset store database.

[0058] Among them, the store database contains detailed information of the store, such as store address, business hours, inventory status, etc. In the process of screening alternative stores, stores that are currently not open can be excluded, and a preliminary judgment can be made on the store inventory to screen out stores with a certain inventory as alternative stores.

[0059] In a feasible embodiment, step S10 may include steps S101 to S102:

[0060] Step S101, calculating a dynamic search range based on geographic location data, wherein the geographic location data includes current positioning coordinates and movement speed, and the size of the dynamic search range is proportional to the movement speed;

[0061] The user's current location coordinates are obtained through the hybrid positioning technology of GPS (Global Positioning System) + base station + WiFi (Wireless Fidelity). At the same time, the user's movement speed is monitored. Specifically, the data of mobile terminal sensors (such as mobile phone accelerometers and gyroscopes) can be used to calculate whether the user is stationary or moving, as well as the speed of movement, and the search range can be dynamically adjusted according to the speed of movement.

[0062] For example, if the user is stationary, the search range is set to a smaller value, such as within 1 kilometer; if the user is moving, the search range is gradually expanded according to the preset proportional relationship based on the speed of movement. For example, for each increase in speed value, the search range is expanded by a certain proportion to ensure that the search range can cover the areas that the user may reach, but will not be too large to reduce the accuracy of the recommendation.

[0063] Step S102: determine a store set within the dynamic search range of the current positioning coordinates from a preset store database, and determine alternative stores from the store set according to preset alternative store screening rules, wherein the alternative store screening rules include screening rules for store business status, store type and / or store product price range.

[0064] After determining the dynamic search range, all stores within the dynamic search range of the current positioning coordinates are retrieved from the preset store database to form a preliminary store set. Then, the store set is filtered according to the preset alternative store screening rules to determine the alternative stores.

[0065] For example, the store status of each store in the store collection is checked, and those that are currently closed are excluded. Secondly, based on the user's needs or preferences, specific types of stores are filtered out. For example, if the user needs to buy clothing, non-clothing stores are excluded. Finally, the store's price range is considered for filtering. If the user has a clear price preference, such as preferring mid- to high-end prices, stores that primarily sell low-priced items are excluded. Using these filtering rules, eligible candidate stores are identified from the initial store collection.

[0066] Step S20, creating a library of alternative outfits based on the product inventory data of the candidate stores;

[0067] By connecting with the inventory management system of the alternative store, the product inventory data of the alternative store can be obtained in real time. The product inventory data may include detailed information such as the type, material, color, size, quantity, etc. of each item sold in the store, and the inventory data is updated in real time. Then, the product information is sorted out from the obtained product inventory data and an alternative outfit library is generated. Each alternative item in the alternative outfit library is a clothing item that is currently actually available for sale in stores within a certain range of the user's current geographical location, ensuring the feasibility and immediacy of the recommendation results.

[0068] It is worth mentioning that the alternative outfit library can be in the form of categorizing and displaying the alternative items, or generating outfit combinations for the alternative items according to certain matching rules (such as style matching, color matching, occasion matching, season matching, etc.) and then displaying them.

[0069] In a feasible embodiment, step S20 may include steps S201 to S202:

[0070] Step S201, obtaining commodity inventory data from the inventory management system of the candidate store in real time;

[0071] By connecting with the inventory management systems of alternative stores in real time, the latest product inventory data can be obtained from the inventory management systems of each alternative store. At the same time, the obtained inventory data is preliminarily sorted out to remove invalid or duplicate information to ensure the accuracy and completeness of the product inventory data.

[0072] Step S202: Classify the candidate items in the candidate stores according to the product inventory data and the selected classification rules to obtain a library of alternative outfits.

[0073] Based on the acquired inventory data, each candidate item in the candidate store is classified according to preset classification rules. The classification rules include but are not limited to the style type (such as hats, shoes, tops, bottoms, coats, etc.), style (such as casual style, business style, sports style, etc.), applicable season (such as spring, summer, autumn, winter), material (such as cotton, linen, wool, etc.), and price range. Multiple classification rules can also be combined and executed according to actual application needs. For example, each candidate unit can be classified by type into hats, clothes, accessories, and shoes, and then further subdivided according to the applicable season to filter out items suitable for the current season.

[0074] Through these classification rules, alternative items can be organized into a structured alternative outfit library. In this alternative outfit library, each item is marked with a clear classification label, which makes it convenient for the subsequent outfit recommendation model to quickly match the appropriate outfit combination according to user needs.

[0075] Step S30: Determine the recommended outfit from the library of alternative outfits based on the preset outfit recommendation model and user feature data and / or outfit information data.

[0076] Based on a preset outfit recommendation model trained with a large amount of clothing matching data, it comprehensively analyzes user feature data (such as body parameters, style preference characteristics, scene adaptation requirements, etc.) and / or outfit information data (such as pictures of clothing items already owned by the user, etc.), intelligently matches and determines the most suitable recommended outfit plan for the user from the alternative outfit library, and provides users with personalized and professional outfit suggestions.

[0077] Furthermore, in a feasible implementation scheme, the user behavior data, such as the length of stay in the fitting room, the matching comparison operation track, the historical purchase record, etc., are integrated simultaneously through the outfit recommendation big model, and the clothing characteristics (material, version, color HSB value, etc.), style characteristics (casual style, business style, etc.) of each outfit and the matching information with the user feature data and / or outfit information data are extracted from the alternative outfit library to comprehensively calculate the matching degree of each outfit with the user. Through the multi-objective optimization unit in the outfit recommendation big model, the aesthetic score, inventory availability, price range and other factors are comprehensively considered to screen out outfits that meet the user's needs. At the same time, the outfit recommendation big model can also be used to further optimize the recommendation results according to the user's current scenario needs (such as commuting, dating, banquets, etc.). For new users, the cold start processing unit provides preliminary recommendations based on the user's basic characteristics and general outfit rules, and gradually optimizes the recommendation results as the user behavior data accumulates.

[0078] Thus, in an embodiment of the present application, alternative stores are determined based on geographic location data, and an alternative outfit library is established based on the product inventory data of the alternative stores, so that users can quickly obtain clothing resource information from multiple nearby stores, providing users with a wider range of style choices, and avoiding the process of users blindly searching for clothing in the store, saving shopping time. At the same time, through the outfit recommendation model, recommended outfits are quickly provided based on user characteristics and / or outfit information data, which makes up for the problem of insufficient matching guidance in offline physical stores, reduces the time cost of users' self-matching and selection, makes the shopping process more efficient and convenient, and improves the user's shopping experience when purchasing clothing offline.

[0079] Based on the first embodiment of the present application, in the second embodiment of the present application, the same or similar contents as those in the above-mentioned embodiment 1 can be referred to the above introduction and will not be described in detail later. Figure 2 As shown, before step S30, steps A10 to A40 may also be included:

[0080] Step A10, obtaining a multi-source data set, wherein the multi-source data set includes clothing attribute data, clothing knowledge data, and user behavior data;

[0081] Data is collected from multiple sources to form a multi-source dataset. Specifically, clothing attribute data can be obtained based on the management system of each store, the e-commerce platform API (Application Programming Interface), and RFID (Radio Frequency Identification) tag scanning. Clothing attribute data includes detailed information such as the material, version, color HSB value, process complexity, and seasonal applicability index of the clothing; clothing knowledge data can be derived from clothing modeling data extracted from high-definition videos of various fashion weeks, fashion magazines, and designers' professional knowledge bases; user behavior data is obtained through user behavior records on the platform, such as the length of time spent in the fitting room, matching comparison operation trajectories, and historical purchase records.

[0082] Step A20: constructing a clothing feature vector based on clothing attribute data, constructing a style knowledge graph based on clothing knowledge data, and constructing an implicit feedback matrix based on user behavior data;

[0083] Clothing attribute data is used to construct a clothing feature vector, which can cover dimensions such as material, style, color HSB (Hue, Saturation, Brightness) value, process complexity and seasonal applicability index; based on clothing knowledge data, the CLIP model (Contrastive Language-Image Pre-training, a multimodal pre-training model) is used to extract image-text matching features and construct a style knowledge graph to reflect the relationship between different clothing styles; based on user behavior data, an implicit feedback matrix is constructed, and the user's preference for different clothing is calculated through a weighted algorithm. For example, the constructed implicit feedback matrix is: F(u,i) = α*click+β*collection+γ*cross-store purchase, which means that different weights are assigned to click, collection and cross-store purchase behaviors respectively.

[0084] Step A30, extracting feature data from the clothing feature vector, style knowledge graph, and implicit feedback matrix;

[0085] The apparel feature vector, style knowledge graph, and implicit feedback matrix are integrated, and a feature extraction algorithm is used to extract features useful for outfit recommendations. For example, material and pattern features are extracted from the apparel feature vector, style features related to fashion trends are extracted from the style knowledge graph, and historical user preference features are extracted from the implicit feedback matrix. These features complement each other to form a comprehensive description of clothing combinations.

[0086] Step A40: Train the neural network model to be trained based on the feature data to obtain an outfit recommendation model.

[0087] The extracted feature data is used as input to train the neural network model. Through deep learning algorithms, the model learns the complex relationships between clothing features, style characteristics, and user preferences. For example, it learns which materials and styles users of a certain style are more likely to choose. After multiple rounds of training and optimization, the resulting model is able to generate personalized outfit recommendations based on user characteristics and clothing information, providing users with precise outfit suggestions.

[0088] In a feasible embodiment, step S30 may include steps S301 to S304:

[0089] Step S301: acquiring user feature data and generating a user feature vector based on the user feature data, wherein the user feature data includes body parameters, style preference characteristics and / or scene adaptation requirements;

[0090] The user feature data input by the user is obtained from the preset user interaction module. The user feature data may include the body parameters (such as height, weight, shoulder width, etc.), style preferences (such as favorite dressing style, color preferences, etc.) and scene adaptation requirements (such as commuting, dating, business, etc.) entered by the user. After integrating these data, a user feature vector is generated through encoding. This vector can comprehensively reflect the user's personalized characteristics in terms of body shape, style and scene requirements.

[0091] It is worth mentioning that user feature data can be information data directly input by the user in the user interaction module, and can also be analyzed by parsing photos in the user's mobile terminal album to determine the user's dressing style, and can also be determined by parsing the schedule data (such as business meetings, appointment schedules) in the user's mobile terminal to determine the scene adaptation requirements.

[0092] Step S302: Generate a feature vector for each candidate outfit based on the item label of each item in the candidate outfit library;

[0093] The label information of each item is extracted from the alternative outfit library. These labels include the type of item (such as hats, shoes, tops, bottoms, jackets, etc.), color, version, material and style. Based on this label information, the alternative outfit feature vector of each item is generated. This vector can accurately describe the feature label of each item.

[0094] Step S303: Calculate the matching degree between the user feature vector and the feature vectors of each candidate outfit using the target outfit recommendation model to select a set of highly matching outfit candidates from the candidate outfit library;

[0095] The user feature vector and the feature vectors of the alternative outfits for each item in the alternative outfit library are input into the target outfit recommendation model. The model calculates the degree of matching between the two through a deep learning algorithm. The calculation of the matching degree takes into account the similarity between the user feature vector and the feature vectors of each alternative outfit in multiple dimensions, such as style consistency, color coordination, and version adaptability. Based on the degree of matching, a candidate set of outfits with a higher degree of matching is screened out from the alternative outfit library.

[0096] Step S304: Generate recommended outfits based on the high-matching outfit candidate set.

[0097] From the set of high-matching outfit candidates, other factors (such as inventory status, price range, user historical preferences, etc.) are comprehensively considered to finally determine the recommended outfit. The recommendation results can be displayed to users in the form of pictures and texts, including the overall effect of the outfit, detailed information of the items, and reasons for the recommendation (such as style matching, color matching, etc.), to help users quickly understand the highlights and adaptability of the recommended outfit.

[0098] In a feasible embodiment, step S30 may further include steps S305 to S308:

[0099] Step S305: Acquire outfit information data and generate an outfit feature vector based on the outfit information data. The outfit information data may be input by scanning, image uploading, and / or text retrieval. The outfit feature vector includes type, color, pattern, material, and / or style vectors.

[0100] The dressing information data input by the user is obtained from the preset user interaction module. The input method can be scanning and recognition (such as scanning the product barcode or QR code), picture uploading (users upload pictures of single items) or text retrieval (users enter descriptions of single items), etc. Based on these input data, the feature information such as type, color, version, material and style of the single items that the user is interested in is extracted to generate a dressing feature vector, which can accurately describe the features of the single item that the user is currently interested in.

[0101] Step S306: Generate a feature vector for each candidate outfit based on the item label of each item in the candidate outfit library;

[0102] The label information of each item is extracted from the alternative outfit library. These labels include the type of item (such as hats, shoes, tops, bottoms, jackets, etc.), color, version, material and style. Based on this label information, the alternative outfit feature vector of each item is generated. This vector can accurately describe the feature label of each item.

[0103] Step S307: Calculate the matching degree between the outfit feature vector and the feature vectors of each candidate outfit using the target outfit recommendation model to select a set of highly matching outfit candidates from the candidate outfit library;

[0104] The outfit feature vector and the alternative outfit feature vectors of each item in the alternative outfit library are input into the target outfit recommendation model. The model calculates the matching degree between the two through a deep learning algorithm. The matching degree calculation takes into account the similarity between the outfit feature vector and the feature vectors of each alternative outfit in multiple dimensions, such as style consistency, color coordination, version adaptability, etc. Based on the level of matching, a set of outfit candidates with higher matching degree is screened out from the alternative outfit library.

[0105] Step S308: Generate recommended outfits based on the high-matching outfit candidate set.

[0106] From the set of high-matching outfit candidates, other factors (such as inventory status, price range, user historical preferences, etc.) are comprehensively considered to finally determine the recommended outfit. The recommendation results can be displayed to users in the form of pictures and texts, including the overall effect of the outfit, detailed information of the items, and reasons for the recommendation (such as style matching, color matching, etc.), to help users quickly understand the highlights and adaptability of the recommended outfit.

[0107] Similarly, it can be seen that the specific implementation steps of determining the recommended outfits from the alternative outfit library based on user feature data and outfit information data through the preset outfit recommendation model are: obtaining user feature data and outfit information data, and generating a user feature vector based on the user feature data, and generating an outfit feature vector based on the outfit information data; generating each alternative outfit feature vector based on the item label of each item in the alternative outfit library; calculating the matching degree of the user feature vector, the outfit feature vector and each alternative outfit feature vector through the target outfit recommendation model to screen a high-matching outfit candidate set from the alternative outfit library; and generating a recommended outfit based on the high-matching outfit candidate set.

[0108] In a feasible embodiment, step S30 may further include steps S40 to S50:

[0109] Step S40: combining the recommended items corresponding to the recommended outfit with the preset user image to obtain a visual matching image;

[0110] The recommended items corresponding to the determined outfit are intelligently combined with the preset user image, so that each recommended item is "worn" on the user image, resulting in an intuitive visual matching image. This allows users to see how the recommended outfit will look on them in advance without actually trying it on, providing users with a more realistic and vivid reference and helping them to more accurately judge whether they like the recommended outfit. The user image is a representative image provided by the user in advance or generated by the system based on the user's basic information.

[0111] Exemplarily, first, a user image is obtained, which is usually a frontal half-body photo uploaded by the user. Then, the user image is preprocessed, including body contour extraction, key point detection and body parameter correction. A lightweight U-Net (convolutional neural network) model is used to separate the background, locate key points such as the shoulder line, waistline, and neckline, and infer parameters such as the waist-to-hip ratio that are not provided based on basic body parameters such as height and weight entered by the user. At the same time, the images of recommended items are standardized, uniformly converted into white background front views, and the pattern features are extracted. Finally, through image synthesis technology, the recommended items are naturally integrated with the user image, and edge feathering and texture fidelity technology are used to optimize the synthesis effect, generate a visual matching image, and intuitively display the overall effect of the recommended outfit.

[0112] It is worth mentioning that after generating a visual matching image, users can adjust the matching items and save one or more matching combinations they like.

[0113] S50: When a purchase instruction for a recommended outfit is received, purchase guidance information including the store location and inventory status of each recommended item is generated.

[0114] When a user issues a purchase instruction for a recommended outfit, the app first confirms the store source of each item in the recommended outfit and obtains detailed information about the store from the store database, including the store's location, business hours, and current inventory status. Then, based on this information, a purchase guide is generated, clearly informing the user of the store location of each item and indicating inventory status, such as availability and full size coverage. If the items selected by the user come from different stores, a cross-store checkout function is also provided, allowing users to add items from different stores to a unified shopping cart and complete a combined payment. Navigation instructions to the store are also provided to ensure that users can successfully purchase their desired items.

[0115] In summary, in this embodiment, by integrating multi-source data sets, including clothing attribute data, clothing knowledge data, and user behavior data, clothing feature vectors, style knowledge graphs, and implicit feedback matrices are constructed, and then a large model for clothing recommendation that can accurately match user shopping needs is trained. In the clothing recommendation process, the user's body parameters, style preferences, scene adaptation requirements, and input clothing information are taken into consideration, and personalized clothing recommendations are dynamically generated in combination with geographic location information and store inventory status. Through virtual fitting technology, users can intuitively see the overall effect of the recommended outfit. Finally, when the user decides to buy, purchase guidance information including the store location and inventory status of the recommended items is quickly generated, realizing a one-stop service from recommendation to purchase. As a result, the personalization, accuracy, and convenience of clothing recommendations are improved, providing users with efficient, intuitive, and practical clothing solutions, reducing the time cost of users' self-matching and selection, making the shopping process more efficient and convenient, and improving the user's shopping experience when purchasing clothing offline.

[0116] For example, in order to help understand the implementation process of the outfit recommendation method obtained by combining this embodiment with the above embodiments, please refer to Figure 3 , Figure 3 A brief flowchart of an outfit recommendation method is provided, specifically:

[0117] After entering the mall, users scan the barcode with their mobile devices. The system then reads the database information and performs geographic location analysis. The system then uses a large model for outfit recommendations to analyze the aesthetics of the outfits based on the data and the user's location, recommending suitable outfit combinations. After selecting their preferred outfit, the system searches for other clothing stores and sends them information about these stores. These stores are usually located near the user's actual location, making it convenient for them to purchase. Users then proceed to the corresponding stores based on the information they receive to complete their purchase. This entire process creates a closed-loop service from the moment the user enters the mall to the moment they complete their purchase. Through the system's intelligent recommendations and guidance, the system improves the user's shopping experience and efficiency.

[0118] It should be noted that the above examples are only used to understand the present application and do not constitute a limitation on the dressing recommendation method of the present application. More simple transformations based on this technical concept are all within the scope of protection of the present application.

[0119] The present application also provides a device for recommending outfits. Figure 4 , the recommended outfits include:

[0120] A store determination module 10 is used to determine candidate stores from a preset store database based on geographic location data;

[0121] The outfit library building module 20 is used to build an alternative outfit library based on the product inventory data of the alternative stores;

[0122] The outfit recommendation module 30 is used to determine recommended outfits from the alternative outfit library based on user feature data and / or outfit information data using a preset outfit recommendation model.

[0123] Optionally, the outfit recommendation device further includes a large model training module (not shown), which is used to:

[0124] Acquire a multi-source data set, wherein the multi-source data set includes clothing attribute data, clothing knowledge data, and user behavior data;

[0125] Construct clothing feature vectors based on clothing attribute data, construct style knowledge graphs based on clothing knowledge data, and construct implicit feedback matrices based on user behavior data;

[0126] Extract feature data from clothing feature vectors, style knowledge graphs, and implicit feedback matrices;

[0127] The neural network model to be trained is trained based on the feature data to obtain the outfit recommendation model.

[0128] Optionally, the store determination module 10 is further configured to:

[0129] Calculating a dynamic search range based on geographic location data, where the geographic location data includes current positioning coordinates and movement speed, and the size of the dynamic search range is proportional to the movement speed;

[0130] A set of stores within the dynamic search range of the current positioning coordinates is determined from a preset store database, and alternative stores are determined from the store set according to preset alternative store screening rules, wherein the alternative store screening rules include screening rules for store business status, store type and / or store product price range.

[0131] Optionally, the outfit recommendation module 30 is further configured to:

[0132] Acquiring user feature data and generating a user feature vector based on the user feature data, wherein the user feature data includes body parameters, style preference characteristics and / or scene adaptation requirements;

[0133] Generate feature vectors for each alternative outfit based on the item labels of each item in the alternative outfit library;

[0134] The target outfit recommendation model calculates the matching degree between the user's feature vector and the feature vectors of each candidate outfit, thereby screening a set of highly matching outfit candidates from the candidate outfit library.

[0135] Generate recommended outfits based on a set of highly matching outfit candidates.

[0136] Optionally, the outfit recommendation module 30 is further configured to:

[0137] Acquire outfit information data and generate an outfit feature vector based on the outfit information data, wherein the outfit information data is input through scanning, image uploading, and / or text retrieval, and the outfit feature vector includes type, color, version, material, and / or style vectors;

[0138] Generate feature vectors for each alternative outfit based on the item labels of each item in the alternative outfit library;

[0139] The target outfit recommendation model is used to calculate the matching degree between the outfit feature vector and the feature vectors of each candidate outfit, so as to select a set of highly matching outfit candidates from the candidate outfit library.

[0140] Generate recommended outfits based on a set of highly matching outfit candidates.

[0141] Optionally, the outfit library establishment module 20 is further configured to:

[0142] Obtain product inventory data in real time from the inventory management system of the candidate store;

[0143] According to the product inventory data and the selected classification rules, the alternative items in the alternative stores are classified to obtain an alternative outfit library.

[0144] Optionally, the outfit recommendation device further includes a purchase guidance module (not shown), which is used to:

[0145] Combine the recommended items corresponding to the recommended outfit with the preset user image to obtain a visual matching image;

[0146] When a purchase instruction for a recommended outfit is received, purchase guidance information is generated including the store location and inventory status of each recommended item.

[0147] The outfit recommendation device provided in the embodiments of this application utilizes the outfit recommendation method described in the above embodiments to optimize the offline clothing consumption process and enhance the user's offline shopping experience. Compared to the prior art, the beneficial effects of the outfit recommendation device provided in the embodiments of this application are the same as those of the outfit recommendation method described in the above embodiments. Other technical features of the outfit recommendation device are the same as those disclosed in the outfit recommendation method described in the above embodiments and are not further elaborated here.

[0148] An embodiment of the present application provides an electronic device, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can execute the outfit recommendation method in the above-mentioned embodiment one.

[0149] Reference below Figure 5 , which shows a schematic diagram of the structure of an electronic device suitable for implementing the embodiments of the present application. The electronic device in the embodiments of the present application may include but is not limited to complete devices such as multimedia interactive all-in-one devices, touch all-in-one devices, etc. Figure 5 The electronic device shown is merely an example and should not limit the functions and scope of use of the embodiments of the present application.

[0150] like Figure 5 As shown, the electronic device may include a processing device 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes based on programs stored in a read-only memory 1002 or programs loaded from a storage device 1003 into a random access memory 1004. The random access memory 1004 also stores various programs and data required for the operation of the electronic device. The processing device 1001, the read-only memory 1002, and the random access memory 1004 are connected to each other via a bus 1005. An input / output interface 1006 is also connected to the bus. Typically, the following systems may be connected to the input / output interface 1006: an input device 1007 including, for example, a touch screen, a touchpad, a keyboard, a mouse, an image sensor, a microphone, an accelerometer, a gyroscope, etc.; an output device 1008 including, for example, a liquid crystal display (LCD), a speaker, a vibrator, etc.; a storage device 1003 including, for example, a magnetic tape or a hard disk; and a communication device 1009. The communication device 1009 may allow the electronic device to communicate with other devices wirelessly or wired to exchange data. Although the figures show electronic devices with various systems, it should be understood that it is not required to implement or have all of the systems shown. More or fewer systems may be implemented or have instead.

[0151] In particular, according to the embodiments disclosed in the present application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, the embodiments disclosed in the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program comprising program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device 1003, or installed from a read-only memory 1002. When the computer program is executed by the processing device 1001, the above-mentioned functions defined in the method of the embodiment disclosed in the present application are performed.

[0152] The electronic device provided in the embodiment of the present application, using the outfit recommendation method in the above embodiment, can optimize the offline clothing consumption process to improve the user's shopping experience when purchasing clothing offline. Compared with the prior art, the beneficial effects of the electronic device provided in the embodiment of the present application are the same as the beneficial effects of the outfit recommendation method provided in the above embodiment, and the other technical features of the electronic device are the same as the features disclosed in the outfit recommendation method in the previous embodiment, and are not further described here.

[0153] It should be understood that the various parts disclosed in the embodiments of the present application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any appropriate manner in any one or more embodiments or examples.

[0154] The above are only specific embodiments of the present application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

[0155] An embodiment of the present application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, and the computer-readable program instructions are used to execute the outfit recommendation method in the above embodiment.

[0156] The computer-readable storage medium provided in the embodiment of the present application can be, for example, a USB flash drive, but is not limited to electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems or devices, or any combination thereof. More specific examples of computer-readable storage media can include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM: Random Access Memory), a read-only memory (ROM: Read Only Memory), an erasable programmable read-only memory (EPROM: Erasable Programmable Read Only Memory or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM: CD-Read Only Memory), an optical storage device, a magnetic storage device, or any suitable combination thereof. In this embodiment, the computer-readable storage medium can be any tangible medium that contains or stores a program that can be used by or in combination with an instruction execution system or device. The program code contained on the computer-readable storage medium can be transmitted using any appropriate medium, including but not limited to: wires, optical cables, RF (Radio Frequency: radio frequency), etc., or any suitable combination thereof.

[0157] The computer-readable storage medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.

[0158] The above-mentioned computer-readable storage medium carries one or more programs. When the above-mentioned one or more programs are executed by an electronic device, the electronic device is enabled to: determine alternative stores from a preset store database based on geographic location data; establish an alternative outfit library based on the product inventory data of the alternative store; and determine recommended outfits from the alternative outfit library based on user feature data and / or outfit information data through a preset outfit recommendation model.

[0159] The computer program code for performing the operations of the embodiments of the present application can be written in one or more programming languages or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, C++, and conventional procedural programming languages such as "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as an independent software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer can be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computer (for example, using an Internet service provider to connect through the Internet).

[0160] The flow charts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present application. In this regard, each box in the flow chart or block diagram can represent a module, program segment or a part of code, and the module, program segment or a part of code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order than that marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flow chart, and the combination of the boxes in the block diagram and / or flow chart can be implemented by a dedicated hardware-based system that performs the specified function or operation, or can be implemented by a combination of dedicated hardware and computer instructions.

[0161] The modules described in the embodiments of the present application may be implemented in software or hardware, wherein the name of a module does not necessarily limit the unit itself.

[0162] The readable storage medium provided in the embodiments of the present application is a computer-readable storage medium, which stores computer-readable program instructions (i.e., a computer program) for executing the above-mentioned outfit recommendation method. This computer-readable storage medium can optimize the offline clothing consumption process to improve the user's shopping experience when purchasing clothing offline. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in the embodiments of the present application are the same as the beneficial effects of the outfit recommendation method provided in the above-mentioned embodiments, and will not be repeated here.

[0163] An embodiment of the present application also provides a computer program product, including a computer program, which implements the steps of the above-mentioned outfit recommendation method when executed by a processor.

[0164] The computer program product provided in the embodiments of this application is capable of mining effective information from data generated by information technology systems. Compared with the prior art, the beneficial effects of the computer program product provided in the embodiments of this application are the same as those of the outfit recommendation method provided in the above embodiments, and will not be elaborated here.

[0165] The above are only some embodiments of the present application and are not intended to limit the patent scope of the present application. All equivalent structural transformations made using the contents of the present application specification and drawings under the technical concept of the present application, or direct / indirect application in other related technical fields are included in the patent protection scope of the present application.

Claims

1. A method for recommending outfits, characterized in that: The outfit recommendation method includes: Determine alternative stores from a preset store database based on geographic location data; Establishing a library of alternative outfits based on the product inventory data of the alternative stores; Through the preset outfit recommendation model, the recommended outfit is determined from the alternative outfit library according to user feature data and / or outfit information data.

2. The outfit recommendation method according to claim 1, wherein: Before the step of determining a recommended outfit from the candidate outfit library based on user feature data and / or outfit information data using a preset outfit recommendation model, the method further includes: Acquire a multi-source data set, wherein the multi-source data set includes clothing attribute data, clothing knowledge data, and user behavior data; Constructing a clothing feature vector based on the clothing attribute data, constructing a style knowledge graph based on the clothing knowledge data, and constructing an implicit feedback matrix based on the user behavior data; Extracting feature data from the clothing feature vector, the style knowledge graph, and the implicit feedback matrix; The neural network model to be trained is trained based on the feature data to obtain an outfit recommendation model.

3. The outfit recommendation method according to claim 1, wherein: The step of determining candidate stores from a preset store database based on geographic location data includes: Calculating a dynamic search range based on geographic location data, wherein the geographic location data includes current positioning coordinates and movement speed, and the size of the dynamic search range is proportional to the movement speed; A set of stores within the dynamic search range of the current positioning coordinates is determined from a preset store database, and alternative stores are determined from the store set according to preset alternative store screening rules, wherein the alternative store screening rules include screening rules for store business status, store type and / or store product price range.

4. The outfit recommendation method according to claim 1, wherein: The step of determining a recommended outfit from the library of candidate outfits based on user feature data using a preset outfit recommendation model includes: Acquiring user feature data and generating a user feature vector based on the user feature data, wherein the user feature data includes body parameters, style preference characteristics and / or scene adaptation requirements; Generate a feature vector for each alternative outfit based on the item label of each item in the alternative outfit library; Calculating the matching degree between the user feature vector and each of the candidate outfit feature vectors through the target outfit recommendation model, so as to screen a set of highly matching outfit candidates from the candidate outfit library; Generate recommended outfits based on the high-matching outfit candidate set.

5. The outfit recommendation method according to claim 1, wherein: The step of determining a recommended outfit from the candidate outfit library based on the outfit information data using a preset outfit recommendation model includes: Acquire outfit information data and generate an outfit feature vector based on the outfit information data, wherein the outfit information data is input by scanning, image uploading, and / or text retrieval, and the outfit feature vector includes type, color, version, material, and / or style vectors; Generate a feature vector for each alternative outfit based on the item label of each item in the alternative outfit library; Calculating the matching degree between the outfit feature vector and each of the candidate outfit feature vectors through the target outfit recommendation model, so as to screen a set of highly matching outfit candidates from the candidate outfit library; Generate recommended outfits based on the high-matching outfit candidate set.

6. The outfit recommendation method according to claim 1, wherein: The step of establishing a library of alternative outfits based on the product inventory data of the alternative stores includes: Obtaining product inventory data in real time from the inventory management system of the candidate store; The alternative items in the alternative stores are classified according to the product inventory data and the selected classification rules to obtain an alternative outfit library.

7. The outfit recommendation method according to claim 1, wherein: After the step of determining a recommended outfit from the candidate outfit library based on user feature data and / or outfit information data using a preset outfit recommendation model, the method further includes: Combining each recommended item corresponding to the recommended outfit with a preset user image to obtain a visual matching image; When a purchase instruction for the recommended outfit is received, purchase guidance information including the store location and inventory status of each recommended item is generated.

8. An electronic device, characterized in that: The electronic device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program is configured to implement the steps of the outfit recommendation method according to any one of claims 1 to 7.

9. A storage medium, characterized in that: The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, the steps of the outfit recommendation method as described in any one of claims 1 to 7 are implemented.

10. A computer program product, characterized in that The computer program product includes a computer program, which, when executed by a processor, implements the steps of the outfit recommendation method as described in any one of 1 to 7.

Images