Commodity information display method and electronic equipment

Abstract

The embodiment of the invention discloses a commodity information display method and electronic equipment, and the method comprises the steps: providing a commodity information page, and enabling a displayed commodity main image to be generated in a manner of splicing commodity images of a plurality of visual angles, the commodity graph of at least part of the visual angles is generated by a first AI generation model according to commodity information provided by a commodity publisher user; after a request for checking the upper body effect of a user for a target commodity is received through the operation option, generating a user upper body effect picture of the target commodity under multiple visual angles according to the commodity pictures of the multiple visual angles corresponding to the target commodity and the three-dimensional model of the user image; and regenerating a commodity main picture according to the user upper body effect pictures of the target commodity in the multiple viewing angles. Through the embodiment of the invention, the views of the commodity in more views can be complemented, the user can visually see the own upper body effect of the clothing commodity, the decision cost is reduced, and the shopping experience and the conversion rate are improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of information processing, in particular to a commodity information display method and an electronic device. BACKGROUND

[0002] In a commodity information service system, a commodity picture is very important information for a user, especially for clothing commodities in a cross-border scenario. An overseas user needs to understand the color and style of a commodity through a commodity picture, imagine the effect of the commodity on the user's body, and although the user cannot make a purchase decision directly through the commodity picture, the commodity picture is usually the main way to attract the user to further click and the like. For example, in a commodity search scenario, a commodity search result page displays commodity cards of multiple commodities, and each commodity card displays a commodity main picture. In most cases, a user is attracted by the commodity main picture to click the commodity card to enter a commodity detail page.

[0003] Generally, a commodity picture depends on shooting and uploading by a merchant. However, since the shooting cost of a commodity picture is usually high, the system side usually does not have a standard requirement for uploading of a clothing picture, and therefore, the commodity picture of many clothing commodities is of low quality or cannot completely present the overall appearance of the clothing. For a commodity information service system, there is also a problem of low commodity shopping efficiency. SUMMARY

[0004] The present application provides a commodity information display method and an electronic device, which can complete the view of a commodity from more perspectives, enable a user to directly see the effect of a clothing commodity on the user's body, reduce decision-making cost, improve shopping experience and conversion rate, reduce the rate of returned goods, and improve commodity shopping efficiency.

[0005] The present application provides the following solutions: A commodity information display method comprises the following steps: providing a commodity information page, the commodity information page being used to display commodity information of at least one commodity, the commodity including a clothing commodity, the commodity information page including a commodity main picture display area, wherein the displayed commodity main picture is generated by stitching commodity pictures from multiple perspectives, and at least part of the commodity pictures from the multiple perspectives are generated by a first AI generation model according to commodity information provided by a commodity publisher user; the commodity information page further includes an operation option for viewing a user body effect of the commodity; after receiving a request for viewing a user body effect of a target commodity through the operation option, generating a user body effect picture of the target commodity from multiple perspectives according to commodity pictures from multiple perspectives corresponding to the target commodity and a three-dimensional model of a user image, wherein the three-dimensional model of the user image is generated by a second AI generation model according to a user image photo and / or body shape data uploaded by the user. According to the user's upper body effect map of the target commodity in multiple perspectives, the commodity main map is regenerated for display in the commodity main map display area of the target commodity in the commodity information page.

[0006] The method further comprises: determining a to-be-displayed commodity in a commodity information page provided by a first commodity information service system; the first commodity information service system comprises a system providing country-specific commodity information service for buyers in multiple countries / regions; obtaining commodity detail information uploaded by a publisher user associated with the to-be-displayed commodity in a second commodity information service system for the to-be-displayed commodity, the commodity detail information at least comprising a commodity map; the second commodity information service system is a system providing commodity information service for buyers in a single country / region; processing the commodity detail information using the first AI generation model to generate commodity maps in multiple perspectives for the to-be-displayed commodity, so as to generate a commodity main map of the to-be-displayed commodity by stitching commodity maps in multiple perspectives, and display the commodity main map in the commodity information page.

[0007] The method further comprises: If the commodity information uploaded by the publisher user of the to-be-displayed commodity does not include a model figure's upper body effect map of the commodity, after generating commodity maps in at least part of the perspectives by the first AI generation model, a virtual model figure image is matched for the to-be-displayed commodity, and upper body effect maps of the to-be-displayed commodity in multiple perspectives on the virtual model figure image are generated, so as to stitch the upper body effect maps of the virtual model figure image in multiple perspectives into a commodity main map.

[0008] The virtual model figure image is generated by a third AI generation model according to a plurality of commodity styles and a country-specific model feature database; When matching a virtual model figure image for a target commodity, a matching virtual model figure image is determined according to clothing metadata information of the target commodity and country-specific information of the target user.

[0009] The first AI generation model is trained based on training data to obtain the ability to generate commodity maps in multiple perspectives, the training data comprising commodity maps and corresponding label information of a plurality of commodities in a commodity information service system, and clothing metadata; wherein the label information of the commodity map comprises shooting perspective information corresponding to the commodity map.

[0010] The first AI generation model is trained by inputting product pictures of a plurality of products and corresponding generation instructions into the first AI generation model, to instruct the first AI generation model to generate a plurality of perspective product pictures according to the input product pictures, and to perform multi-round iterative optimization on the generation results by using the annotation information.

[0011] The method further comprises: When generating the user upper body effect picture of the target product in a plurality of perspectives, the fitting posture information of the three-dimensional model of the user image is determined according to the clothing metadata information of the target product, so as to generate a plurality of perspective user upper body effect pictures of the user image trying on the target product in corresponding fitting postures.

[0012] A product information display method comprises: A product information page is displayed, the product information page being used to display product information of at least one product, the product including a clothing product, the product information page including a product main picture display area, wherein the displayed product main picture is generated by stitching a plurality of perspective product pictures, at least part of the perspective product pictures being generated by a first AI generation model according to product information provided by a product publisher user; the product information page further includes an operation option for viewing a user upper body effect; After receiving an operation request of a user through the operation option, the request is submitted to a server, so as to generate a user upper body effect picture of the target product in a plurality of perspectives according to a plurality of perspective product pictures corresponding to the target product and a three-dimensional model of a user image, and return the product main picture regenerated according to the user upper body effect picture of the target product in a plurality of perspectives to the client; the three-dimensional model of the user image is generated by a second AI generation model according to a user image photo and / or body shape data uploaded by the user; The regenerated product main picture is displayed in the main picture display area of the product information page.

[0013] A product search result display method comprises: After receiving a search request about a clothing product, a product search result page is provided, the product search result page including a plurality of resource positions for displaying product information cards of a plurality of products, the product information card including a product main picture display area, wherein the displayed product main picture is generated by stitching a plurality of perspective product pictures, at least part of the perspective product pictures being generated by a first AI generation model according to product information provided by a product publisher user; the product search result page further includes an operation option for viewing a user upper body effect of a product provided in a resource position unit; After receiving a request for viewing the user-on-body effect of a target commodity through the operation option, the request is submitted to a server, so that the server generates a user-on-body effect drawing of the target commodity in multiple perspectives according to a plurality of perspective commodity drawings corresponding to the target commodity and a three-dimensional model of a user image, and returns a regenerated main commodity drawing to the client according to the user-on-body effect drawing of the target commodity in multiple perspectives; the three-dimensional model of the user image is generated by a second AI generation model according to a user image photo and / or body shape data uploaded by a user; The regenerated main commodity drawing is displayed in the main drawing display area in the commodity information card of the target commodity.

[0014] A commodity information display method, comprising: Providing a commodity information page, the commodity information page being used to display commodity information of at least one commodity, the commodity including a clothing commodity, the commodity information page including a main commodity drawing display area; the commodity information page further including an operation option for viewing a user-on-body effect of a commodity; After receiving a request for viewing the user-on-body effect of a target commodity through the operation option, a user-on-body effect drawing of the target commodity is generated according to a main commodity drawing of the target commodity and a three-dimensional model of a user image; the three-dimensional model of the user image is generated by an AI generation model according to a user image photo and / or body shape data uploaded by a user; A main commodity drawing is regenerated according to the user-on-body effect drawing of the target commodity, for replacing and displaying in a main commodity drawing display area of the target commodity in the commodity information page.

[0015] A commodity drawing information processing system, comprising: A model module for providing an AI generation model; A data module for collecting data to generate a training data set, and training the AI generation model, so that the AI generation model obtains the ability to generate commodity drawings in multiple perspectives; the training data includes commodity drawings and corresponding label information of a plurality of commodities in a commodity information service system, and clothing metadata; wherein the label information of the commodity drawing includes shooting perspective information corresponding to the commodity drawing; An image module for determining commodity drawings in multiple perspectives for clothing commodities, wherein the commodity drawings in at least some perspectives are generated by the AI generation model according to commodity information provided by a commodity publisher user associated with a target commodity; The guide module is configured to determine a target product to be displayed when a product information page needs to be displayed, and obtain a plurality of perspective product images corresponding to the target product, generate a main product image of the target product by splicing the plurality of perspective product images, and display the main product image in a main product image display area of the product information page; provide an operation option for viewing a user upper body effect of the product in the product information page; after receiving a request for viewing the user upper body effect of the target product through the operation option, generate a user upper body effect image of the target product in multiple perspectives through the image module according to the plurality of perspective product images corresponding to the target product and a three-dimensional model of a user image, wherein the three-dimensional model of the user image is generated by a second AI generation model according to a user image photo and / or body shape data uploaded by a user; and regenerate a main product image according to the user upper body effect image of the target product in multiple perspectives, and display the main product image in the main product image display area of the target product in the product information page.

[0016] A computer readable storage medium having stored thereon a computer program, which, when executed by a processor, implements the steps of the method of any preceding method.

[0017] An electronic device comprising: one or more processors; and a memory associated with the one or more processors, the memory configured to store program instructions that, when executed by the one or more processors, perform the steps of the method of any preceding method.

[0018] A computer program product comprising computer program / computer executable instructions to implement the steps of the method of any preceding method when executed by a processor in an electronic device.

[0019] According to the specific embodiments provided in the present application, the following technical effects are disclosed: Through the embodiments of the present application, first, a commodity information page can be provided, the commodity displayed on the commodity information page includes a clothing commodity, which includes a commodity main picture display area, the commodity main picture displayed is generated by splicing commodity pictures of multiple perspectives, wherein at least part of the commodity pictures of multiple perspectives are generated by a first AI generation model according to commodity information provided by a commodity publisher user. In addition, the commodity information page can also include an operation option for viewing the user's upper body effect. After receiving a request for viewing the user's upper body effect of a target commodity through the operation option, a user's upper body effect picture of the target commodity under multiple perspectives can be generated according to the multiple perspective commodity pictures corresponding to the target commodity and a three-dimensional model of the user's image, wherein the three-dimensional model of the user's image can be generated by a second AI generation model according to a user's image photo and / or body shape data uploaded by the user. Then, the commodity main picture can be regenerated according to the user's upper body effect picture of the target commodity under multiple perspectives, so as to be used to replace the commodity main picture display area of the target commodity on the commodity information page for display. In this way, the traditional single commodity picture shopping guide scheme can be evolved into a multi-perspective view shopping guide scheme, and the user's upper body effect under multiple perspectives can also be generated and directly replaced into the commodity main picture display area for display. In this way, not only the view of the commodity under more perspectives can be completed, but also the user can directly see the upper body effect of the clothing commodity, which reduces the decision cost, improves the shopping experience and conversion rate, and improves the commodity shopping guide efficiency of the commodity information service system.

[0020] Of course, implementing any product of the present application does not necessarily require all the advantages described above. BRIEF DESCRIPTION OF DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings needed in the embodiments will be briefly introduced as follows. Obviously, the drawings in the following description only constitute some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0022] Figure 1 is a schematic diagram of the system architecture provided by the embodiments of the present application; Figure 2 is a flowchart of the first method provided by the embodiments of the present application; Figure 3 is a schematic diagram of the first interface provided by the embodiments of the present application; Figure 4 is a schematic diagram of the second interface provided by the embodiments of the present application; Figure 5 is a flowchart of the second method provided by the embodiments of the present application; Figure 6 is a flowchart of a third method provided by an embodiment of the present application; Figure 7 is a flowchart of a fourth method provided by an embodiment of the present application; Figure 8 is a schematic diagram of a system provided by an embodiment of the present application; Figure 9 is a schematic diagram of an electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION

[0023] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art belong to the scope of protection of the present application.

[0024] First of all, it needs to be pointed out that the main product image generally refers to the first picture displayed in the product information page. Whether it is a product detail page or a product display page (including a product search result page, a product recommendation page, etc.), such a main product image is usually the first picture presented after opening the page, and therefore, it usually has the most representative features of the product, which can intuitively show the appearance, color, style, etc. of the product. In the prior art, especially in the non-cross-border scenario dominated by domestic users, the main product image is usually a single image or a short video for a small number of products. In the case of a single product image as the main image, the picture is usually taken from a certain perspective of the product, and most of them are from the front perspective. However, in the product information service system, there are some merchants who have published some products in the non-cross-border scenario first, and then opened an export plan, that is, they sell some or all of the products in the domestic market for a period of time, and then convert them to "export", that is, sell them to foreign users in the cross-border scenario. In the cross-border scenario, overseas users will rely more on the views of multiple perspectives such as the front, side and back of the product to understand the overall appearance of the product, and then make a purchase decision or a decision on the next step. Obviously, the product information uploaded by the merchant in the non-cross-border scenario cannot meet the needs of the above-mentioned cross-border users in most cases.

[0025] On the other hand, if the merchant has already uploaded the product information in a non-cross-border scenario, the system can help the merchant complete the translation of the product information and the publication of the product information in the cross-border product information system in an automated manner, thereby saving the operation cost of the merchant. However, if the product image information uploaded by the merchant in the non-cross-border scenario lacks some views of different perspectives, in the prior art, the merchant usually needs to re-shoot and supplement the views of other perspectives, which obviously increases the cost of the merchant and is very inefficient, and also has a negative impact on the automation of the publication of the product in the process of “internal conversion to external”.

[0026] To address the above issues, in the embodiments of the present application, the AI (Artificial Intelligence) generation technology is first applied to the scenario of product main image display. Specifically, based on the product image uploaded by the merchant, the AI generation model is used to complete the product image of at least part of the perspectives. For example, if the merchant only uploads the view of the front perspective of a product, the AI generation model can be used to generate the views of the side and back perspectives. Then, the product main image can be generated by splicing the views of different perspectives. In this way, when the product information is displayed on the product detail page or the product display page, the image generated by splicing the views of different perspectives can be used as the product main image for display, so that the user can obtain the overall features of the product from the image first presented in the page, thereby helping the user to make a purchase decision or guiding the user to perform the next step of clicking into the detail page.

[0027] In addition, based on the generation of multiple perspective views of the product by the AI generation model, the embodiments of the present application also provide a user upper body effect image of a specific product, that is, multiple images corresponding to the user upper body effect images displayed from different perspectives can be generated, and then the user upper body effect images of different perspectives are spliced into a product main image and displayed in the product main image display area of the product information page.

[0028] Of course, for the scenarios of product search and product recommendation, since multiple product information needs to be displayed in the page, the user can select which product to display the user upper body effect image as needed. Specifically, in the default state, the product main image generated by splicing the multiple perspective views can be displayed, and an operation option can be provided in the product information page. The operation option can be provided in units of products, that is, different products can correspond to different operation options, and if the user needs to view the user upper body effect image of a product, the user can initiate a request through the operation option.

[0029] In another aspect, in order to generate the user-on-body effect picture, the user can also upload a photo and / or body shape data of himself / herself, and the AI generation model can generate a three-dimensional model of the user image according to the user image photo and / or body shape data uploaded by the user, which can be saved on the server. Wherein, when the user first initiates a request for generating a user-on-body effect picture, it can be judged whether the user has uploaded a user image photo and / or body shape data. If yes, the user-on-body effect picture can be generated directly by using the three-dimensional model of the user image saved on the server. Otherwise, the user can be prompted to upload a user image photo and / or body shape data first, and then generate a three-dimensional model of the user image, and then generate a user-on-body effect picture of a specific product.

[0030] In the case of generating a three-dimensional model of the user image, after receiving a request for generating a user-on-body effect picture initiated by the user for a target product, the user-on-body effect picture of the target product in multiple perspectives can be generated according to the product pictures in multiple perspectives corresponding to the target product and the three-dimensional model of the user image. Then, the user-on-body effect pictures in multiple perspectives can be spliced into a main product picture and displayed in the main product picture display area of the target product. In this way, the user can directly view the user-on-body effect picture of the specific product on his / her body in the main product picture display area, and the multiple perspective views make the user more intuitive to understand the overall user-on-body effect of the specific product on his / her body.

[0031] From the perspective of system architecture, referring to Figure 1 The embodiments of the present application can provide AI generation of multi-perspective views of products and splicing of multi-perspective views of products into main product pictures for display in product information pages in business scenarios such as product search, product recommendation, product details, etc. in a product information service system. Specifically, the system can include multiple layers such as a model layer, a data layer, an image layer, and a shopping guide layer. In the model layer, AI models can be provided, including the AI generation model in the embodiments of the present application, and other algorithm models for functions such as clothing fitting, posture fitting, and cultural fitting. In the data layer, data collection, data cleaning and labeling, etc. can be provided, and the amount of training data can be enriched through data augmentation and synthesis, which can be used to train the AI generation model and other algorithm models. The trained models can be used to generate multi-perspective views of specific products, and can also be used for image processing, saving, etc. The generated multi-perspective views can be displayed in the shopping guide layer, which includes product search or personalized product recommendation, etc. In addition, the specific AI generation model can also generate a three-dimensional model of the user image, and generate multi-perspective user-on-body effect pictures according to the multi-perspective views of the product and the three-dimensional model of the user image, and then splice the multi-perspective user-on-body effect pictures into a new main product picture.

[0032] From the perspective of application scenarios, the scheme provided by the embodiments of the present application can be used in ordinary commodity information service systems, or, in a preferred manner, AI generation of multi-view views can be performed based on commodity pictures uploaded by domestic merchants in a cross-border scenario, so as to provide more rich commodity information for overseas users. For example, in a commodity search scenario, an overseas user inputs the keyword "lady dress" for search, and the system side can first perform commodity recall. For the recalled commodities, a commodity main picture can be generated by splicing commodity pictures in multiple views, wherein the views in some views can be generated by an AI generation model according to information such as commodity pictures uploaded by merchants in a domestic scenario; in the process of displaying specific search results, the user can view the upper body effect picture of a target commodity on himself / herself, at this time, the commodity main picture area in the information card of the commodity in the commodity search result page will display the commodity main picture spliced by the user's upper body effect picture of the specific commodity in multiple views.

[0033] The specific implementation scheme provided by the embodiments of the present application will be described in detail below.

[0034] Embodiment one First, the embodiment one provides a commodity information display method from the perspective of the server, see Figure 2 The method can specifically include: S201: providing a commodity information page, the commodity information page is used to display commodity information of at least one commodity, the commodity includes a clothing commodity, and the commodity information page includes a commodity main picture display area, wherein the displayed commodity main picture is generated by splicing commodity pictures in multiple views, and at least part of the commodity pictures in the multiple views are generated by a first AI generation model according to commodity information provided by a commodity publisher user.

[0035] Among them, the commodity information page can have multiple types, for example, in one case, the commodity information page can include a commodity display page used to display commodity information of multiple commodities, wherein the specific commodity display page can include a commodity search result display page or a commodity recommendation information display page. The commodity display page can include multiple resource positions, and the resource position includes a resource position used to display a commodity information card of a single commodity. That is, in the commodity display page, information of multiple commodities can be displayed, and information of each commodity can be carried in the form of a commodity information card, and the commodity main picture generated by AI generation of multi-view views and splicing can be displayed in the main picture display area of the commodity information card.

[0036] Alternatively, the specific commodity information page can also be a commodity detail page for displaying details of a single commodity, and the commodity detail page also includes a commodity main image display area. The AI-generated multi-view image and the spliced commodity main image can also be displayed in the main image display area of the commodity detail page.

[0037] In the embodiments of the present application, the commodity displayed in the specific commodity information page can be a clothing commodity, specifically a garment, an accessory, a hat, and the like. Wherein, which commodities are displayed in the commodity information page can be determined in multiple ways. For example, if the commodity information page is a commodity search result display page, the commodities displayed specifically can be commodities that meet the search conditions obtained by searching according to the search keywords and the like input by the user. For example, assuming that the search keywords input by the user are “vacation-style dress”, the commodities displayed on the page can be commodities that belong to the “dress” category and have the “vacation-style” label, and the like, retrieved from the commodity information library. Alternatively, if the commodity information page is a commodity recommendation information display page, the commodities displayed specifically in the page can be commodities that the user can be interested in, determined according to the historical behavior data, preference information, recent behavior data, and the like of the current viewer user, which can include clothing commodities. Alternatively, if the commodity information page is a commodity detail page, the commodity displayed specifically in the page is the commodity itself associated with the specific commodity detail page. For example, assuming that the user clicks on a detail page link of a commodity in a commodity display page, the commodity detail page of the commodity can be displayed, and accordingly, the commodity can be the commodity displayed in the commodity information page in the embodiments of the present application.

[0038] Wherein, in the embodiments of the present application, whether the commodity display page or the commodity detail page, the specific commodity main image is no longer a single commodity image, but is spliced from multi-view commodity images. Wherein, at least part of the multi-view commodity images can be AI-generated. In specific implementation, the process of multi-view image generation and splicing for a specific commodity can be completed in advance in an offline manner, or can also be generated online, that is, when a commodity needs to be displayed in a specific page, AI generation and splicing of multi-view images are performed for the commodity to obtain a commodity main image including multiple view images. Alternatively, an online + offline combination can also be used to achieve the AI generation of multi-view images, for example, for some high-frequency commodities (commodities with high frequency of recall in search or recommendation scenarios), the AI generation of multi-view images can be completed in advance by offline processing, and for non-high-frequency commodities, online generation can be used, so that a balance between response rate, real-time calculation cost and system storage data volume can be achieved.

[0039] The AI generation model specifically refers to a deep learning model containing a large number of parameters. Such an AI generation model can store and process a large amount of information due to its large parameter size, thereby achieving higher performance in content generation tasks. The specific AI generation model can include a single-modal generation model or a multi-modal generation model. The single-modal generation model can only generate content of a certain modality, for example, a "text-to-text" model (generating text based on text), a "text-to-image" model (generating images based on text), and the like. The multi-modal generation model refers to an AI generation model that can generate or process multiple different modalities of content, i.e., the same AI generation model can be used to complete content generation tasks of multiple different modalities such as text, images, and videos, or generate content of a certain modality based on multiple modalities of content, for example, a "(text+image)-to-image" model (i.e., generating another image based on input text and images), and the like. In the embodiments of the present application, since the AI generation model is mainly used to generate a specific product image of a product from a certain perspective, an AI generation model with image generation capability can be used to achieve this, for example, the aforementioned "(text+image)-to-image" model, and the like.

[0040] Of course, in order to make the specific AI generation model more suitable for the multi-perspective product view generation scenario in the embodiments of the present application and improve the generation quality, the AI generation model can also be trained using dedicated training data based on the open-source AI generation model. The specific training data used can also be collected and labeled from the product information service system. For example, the product information service system can collect a large amount of product information such as product images (including product main images, scene images, etc.), clothing metadata (category, style, color, style, etc.), and the like of a large number of products, and label the categories and perspectives of specific product images, for example, the product image categories can include main images, scene images, and the like, and the perspectives can include front, side, back, and the like. Then, the AI generation model can be trained using these training data and their labeled information. Specifically, the AI generation model can be input with product images of multiple products in the training data and corresponding generation instructions to instruct the AI generation model to generate product images of multiple perspectives based on the input product images, and then the labeled information can be used to optimize the generation results. After multiple iterations, the training of the AI generation model can be completed. The trained AI generation model can have the ability to generate product images of specified perspectives based on product images, clothing metadata, and the like of target products.

[0041] Specifically, when training the above-mentioned first AI generation model, you can collect clothing product images uploaded by merchants from the product information service system, and then extract clothing metadata from them. After that, you can perform data cleaning (including removing some low-quality data, etc.) and labeling. Then, these product images, clothing metadata, and labeling information can be collected as training data to be used for the pre-trained AI generation model to obtain the first AI generation model in the embodiment of this application, and enable it to have the ability to generate views of the product from other perspectives based on the input content such as the input product image and product text description information. In the subsequent process of using the first AI generation model to generate and display multi-perspective views, user behavior data can also be collected, including data such as user clicks and dwell time on the spliced ​​product main image, so as to continuously optimize the first AI generation model through these behavior data.

[0042] In a specific implementation, when the first AI generation model generates multi-perspective views for a product, the generated multi-perspective views may only include the product itself. For example, if a product is a short-sleeved top, the generated multi-perspective views may include the front, side, and back views of the top. Alternatively, because the model's upper body rendering is also very important for clothing products, in a preferred implementation, the first AI generation model may also generate an upper body rendering with the model's image based on the generated multi-perspective views.

[0043] If the product information provided by the product publisher (e.g., a merchant) (which can be published product information and / or product information uploaded to the server but not yet published) includes images of a model wearing the product, the first AI generation model can generate images of the model wearing the product from other perspectives when generating product images from a limited number of perspectives. In other words, if the merchant captures a model wearing the product from the upper body when capturing product images, the first AI generation model can continue to use the same model image when generating views from other perspectives, ensuring consistency in the model's appearance across different perspectives.

[0044] If the product information uploaded by the product publisher user does not include a model's upper body rendering of the product, after the first AI generation model generates a product image from at least a partial perspective, it is also possible to match a virtual model image for the target product and generate upper body renderings of the product image from multiple perspectives on the virtual model image, so that the upper body renderings of the virtual model image corresponding to the multiple perspectives can be spliced ​​into the main product image.

[0045] Among them, regarding the virtual model character image, it can be randomly generated, or since an important application scenario of the embodiment of the present application is a cross-border e-commerce scene, specific buyer users can be distributed in multiple different countries / regions, and models in different countries / regions can be different in body shape, skin color, etc., and at the same time, models in the same country / region can also have some commonalities in body shape, skin color, etc. Therefore, in the preferred mode, the virtual model character image can also be generated by a third AI generation model according to the clothing metadata of the commodity and the nationalized model feature database (including the body shape, skin color, etc. of the model character). That is, different clothing metadata and different countries / regions can correspond to different virtual model character images. In this way, when matching the virtual model character image for the target commodity, the appropriate virtual model character image can be selected according to the clothing metadata of the target commodity and the country / region information of the current viewer user. Then, the multi-perspective commodity image generated by the first AI generation model is matched to the virtual model character image, and a multi-perspective virtual model character image is obtained.

[0046] As described above, the embodiments of the present application can be mainly applied to cross-border scenarios, that is, in the process of displaying a commodity information page to overseas users, the specific commodity to be displayed is determined, if the commodity is published by the merchant in the commodity information service system for domestic users, the multi-view view of the commodity can be generated based on the commodity information (including commodity pictures, text description information, etc.) published by the merchant in the domestic scenario through the AI generation model, and the commodity main picture is spliced and displayed to the overseas users in the commodity information page. That is, for the step S201, the commodity to be displayed in the commodity information page provided by the first commodity information service system can be determined first; the first commodity information service system includes a system providing country-specific commodity information service to buyers in multiple countries / regions; the so-called country-specific commodity information service can include translation and other processing of commodity information, that is, for the same commodity information page, when displayed to users in different countries / regions, the commodity information can be translated into the language of the corresponding country, in addition, the layout of the commodity information on the page and the like can also be processed differently for different countries / regions. When displaying the commodity information page in the above-mentioned first commodity information service system to a user in a certain country / region, for the target commodity to be displayed, the commodity detail information uploaded by the publisher user associated with the target commodity in the second commodity information service system can be obtained, and the commodity detail information can at least include commodity pictures. The specific second commodity information service system can be a system providing commodity information service to buyers in a single country / region. Since the commodity pictures published in such a second commodity information service system can only have views in some perspectives, the first AI generation model can be used to generate views in more perspectives for the specific target commodity, so as to generate a commodity main picture through multi-perspective view splicing and display it in the specific commodity information page of the first commodity information service system. In this way, the overseas user can obtain more detailed appearance information about the specific commodity from the commodity main picture, which helps him make more efficient decisions on the next step of behavior such as commodity purchase.

[0047] Of course, in a non-cross-border scenario, the AI generation model can also be used to generate commodity pictures in some perspectives for a specific commodity, and a commodity main picture can also be generated through multi-perspective view splicing and displayed in a commodity information page, which can also enable domestic users to obtain information about the commodity in multiple perspectives through the commodity main picture, and is also conducive to helping users make purchase decisions or decisions on the next action point.

[0048] Specifically, when displaying the generated product main view in the embodiments of the present application, if it is a product display page, since the page includes multiple resource positions, each of which is used to display a product information card of a product, and the product information card includes a product main view display area, the product main view obtained by splicing can be displayed in the main view display area in the product information card. If it is a product detail page, there is also a main view display area in the page, and thus the product main view obtained by splicing can also be displayed in the main view display area in the product detail page.

[0049] For the case of displaying in a product display page, adaptive improvements can also be made to the arrangement and layout of resource positions and the like. Specifically, in the traditional way, for a product search result display page or a product recommendation information display page, the product information cards in the page are usually arranged in a "one row two" manner, that is, two columns are arranged vertically, and two product information cards are displayed in each row, which makes the width of the main view display area in the product information card at most half of the screen of the terminal device. In the case of using a single image as a product main view, the area of the main view display area in the product information card is sufficient, but in the embodiments of the present application, since the product main view is spliced from multiple perspective product images, if the product information cards are still arranged in a "one row two" manner, the size of each perspective product image will become relatively small, which may not effectively present the characteristics of the product.

[0050] Therefore, in the embodiments of the present application, the resource positions in the product display page can be arranged in a "one row one" manner, that is, each resource position occupies the full screen horizontally, and different resource positions are arranged in a single column vertically, so that only one product information card is displayed in each row, and the size of the main view display area in each product information card is large enough in the horizontal direction. When splicing multiple perspective views, the multiple perspective product images can be arranged in a single row and multiple columns from left to right and spliced into a product main view.

[0051] In the preferred implementation, the multiple perspective views in the embodiments of the present application can specifically be three-perspective views, that is, they can include three views of the front, side and back of the clothing, so that the overall appearance of the clothing product can be presented through fewer views. Of course, one perspective can correspond to multiple product images, for example, the side perspective can include views of the left side and right side perspective. In addition, more or fewer views in different perspectives can also be generated, as long as the overall appearance of the clothing product can be displayed relatively completely.

[0052] For example, as shown in FIG. 6, the product main view 601 is generated by splicing multiple perspective product images, and the product main view 601 is displayed in the main view display area 602 in the product information card 603. Figure 3As shown in the search result display page in an example provided by the embodiment of the application, a commodity information card is shown at 31, and a main image display area in the commodity information card is shown at 32. From Figure 3 As can be seen, the main image display area displays a commodity main image including commodity images in three perspectives of front, side and back (wherein the side perspective includes two commodity images of left side and right side), and each commodity information card can occupy the full screen in the horizontal direction.

[0053] Regarding the commodity detail page, since the area of the main image display area in the commodity detail page is relatively large, the size thereof need not be adjusted, but when splicing the multi-perspective commodity images, in order to avoid the situation of proportion imbalance, a multi-row and multi-column arrangement can be adopted, for example, assuming that there are four commodity images in total, they can be arranged in two rows and two columns, and the like.

[0054] S202: After receiving a request for viewing a user upper body effect of a target commodity through an operation option provided in the commodity information page, generating a user upper body effect drawing of the target commodity in multiple perspectives according to the multiple-perspective commodity images corresponding to the target commodity and the three-dimensional model of the user image; the three-dimensional model of the user image is generated by a second AI generation model according to a user image photo and / or body shape data uploaded by the user.

[0055] The foregoing step S201 is a basic step of the specific scheme provided by the embodiment of the application, that is, in the default state, when displaying the commodity information page, an AI generation model can be used to generate multiple-perspective commodity images for a specific commodity and splice them into a commodity main image for display. During the display process, an operation option for viewing a user upper body effect drawing can be provided at the commodity granularity, and if a user needs to view a user upper body effect drawing of a target commodity, the user can initiate a request through this operation option. This operation option can be fixedly provided at a certain position in the commodity information page, for example, at the lower right corner of a specific resource bit, or the like. Alternatively, in order to enhance interactivity, the operation option can also be provided in the form of a pop-up layer, for example, as shown in Figure 4 (A), which is a search result page. Figure 3 As shown in the search result page, an operation option for initiating a request for viewing a user upper body effect is provided in the form of a pop-up layer (specifically, the operation option is displayed when a user stays at a certain information card for more than a certain threshold, and the like), and the user can initiate a specific request for viewing a user upper body effect by clicking a “try it now” button or the like.

[0056] After receiving the above request, the user's upper body effect picture of the target commodity in multiple perspectives can be generated according to the commodity pictures of multiple perspectives corresponding to the target commodity and the three-dimensional model of the user image. As described above, the three-dimensional model of the user image can be generated by the second AI generation model according to the user image photo and / or body shape data uploaded by the user. Specifically, after the user clicks the "try it now" option and the like shown in (A), if the user uses the function for the first time and has not uploaded the user photo or body shape data, the user can be prompted to upload. After uploading, the second AI generation model can generate a specific user image three-dimensional model for the user. If the user image three-dimensional model has been generated for the user before, the three-dimensional model saved on the server can be used to directly generate the user's upper body effect picture of the target commodity in multiple perspectives. Figure 4

[0057] In specific implementation, the information such as material, texture, color, and version of the clothing can be extracted from the commodity pictures in multiple perspectives, the clothing and the background can be separated by the image segmentation model, the clothing can be "worn" on the virtual image of the user, the user's upper body effect picture in multiple perspectives such as front, side, and back can be generated, and the background image such as background virtualization or standardization can be added.

[0058] S203: Regenerate the commodity main picture of the target commodity according to the user's upper body effect picture of the target commodity in multiple perspectives, and replace the commodity main picture of the target commodity in the commodity main picture display area of the commodity information page for display.

[0059] After generating the user's upper body effect picture of the target commodity in multiple perspectives, the commodity main picture of the target commodity can be regenerated and replaced for display in the commodity main picture display area corresponding to the target commodity. When generating the user's upper body effect picture of the target commodity in multiple perspectives, the try-on posture information of the three-dimensional model of the user image can also be determined according to the clothing metadata information of the target commodity, so as to generate the user's upper body effect picture of the target commodity in multiple perspectives in the corresponding try-on posture. For example, the effect after replacement can be as shown in (B). Figure 4 (B) (the different commodities can be ignored in the figure mainly for presenting the replacement from the model image to the user image).

[0060] ​In summary, through the embodiments of the present application, first, a commodity information page can be provided, and the commodities displayed on the commodity information page include clothing commodities, which include a commodity main picture display area. The displayed commodity main picture is generated by splicing commodity pictures of multiple perspectives, wherein at least part of the commodity pictures of multiple perspectives are generated by a first AI generation model according to commodity information provided by a commodity publisher user. In addition, the commodity information page can also include an operation option for viewing user upper body effects. After receiving a request for viewing user upper body effects of a target commodity through the operation option, a user upper body effect picture of the target commodity in multiple perspectives can be generated according to multiple commodity pictures of multiple perspectives corresponding to the target commodity and a three-dimensional model of a user image. The three-dimensional model of the user image can be generated by a second AI generation model according to a user image photo and / or body shape data uploaded by the user. Then, a commodity main picture can be regenerated according to the user upper body effect picture of the target commodity in multiple perspectives, so as to replace the commodity main picture display area of the target commodity in the commodity information page for display. In this way, the traditional single commodity picture shopping guide scheme can be evolved into a multi-perspective view shopping guide scheme, and multi-perspective user upper body effects can also be generated and directly displayed in the commodity main picture display area. In this way, not only can the views of commodities in more perspectives be completed, but also users can directly view the upper body effects of clothing commodities, reduce decision-making costs, improve shopping experience and conversion rate, reduce return rates, and improve the commodity shopping guide efficiency of the commodity information service system.

[0061] Embodiment Two This embodiment two is corresponding to embodiment one, and from the perspective of the client, a commodity information display method is provided, which is described with reference to Figure 5 The method can include: S501: display a commodity information page, wherein the commodity information page is used to display commodity information of at least one commodity, the commodity includes clothing commodities, and the commodity information page includes a commodity main picture display area. The displayed commodity main picture is generated by splicing commodity pictures of multiple perspectives, wherein at least part of the commodity pictures of multiple perspectives are generated by a first AI generation model according to commodity information provided by a commodity publisher user. The commodity information page also includes an operation option for viewing user upper body effects. S502: After receiving the operation request of the user through the operation option, submit the request to the server to generate the user's upper body effect drawing of the target commodity under multiple perspectives according to the commodity drawings of multiple perspectives corresponding to the target commodity and the three-dimensional model of the user's image, and return the regenerated commodity main drawing to the client according to the user's upper body effect drawing of the target commodity under multiple perspectives; the three-dimensional model of the user's image is generated by the second AI generation model according to the user's image photo and / or body shape data uploaded by the user; S503: Replace the regenerated commodity main drawing to the main drawing display area of the commodity information page for display.

[0062] Embodiment three This embodiment three is for the application of specific scheme in the commodity search scene, and provides a commodity search result display method, which is described with reference to Figure 6 The method can include: S601: After receiving the search request of the clothing commodity, provide a commodity search result page, which includes multiple resource positions for displaying commodity information cards of multiple commodities, and the commodity information card includes a commodity main drawing display area, wherein the displayed commodity main drawing is generated by pasting multiple perspective commodity drawings, and at least part of the perspective commodity drawings are generated by the first AI generation model according to the commodity information provided by the commodity publisher user; the commodity search result page also includes operation options for viewing the user's upper body effect of the commodity in the form of resource positions; S602: After receiving the request for viewing the user's upper body effect of the target commodity through the operation option, submit the request to the server to generate the user's upper body effect drawing of the target commodity under multiple perspectives according to the commodity drawings of multiple perspectives corresponding to the target commodity and the three-dimensional model of the user's image, and return the regenerated commodity main drawing to the client according to the user's upper body effect drawing of the target commodity under multiple perspectives; the three-dimensional model of the user's image is generated by the second AI generation model according to the user's image photo and / or body shape data uploaded by the user; S603: Replace the regenerated commodity main drawing to the main drawing display area in the commodity information card of the target commodity for display.

[0063] Embodiment four In the foregoing embodiments, mainly on the basis of the AI generating the multi-view view of the commodity and splicing into the commodity main view, the multi-view user upper body effect picture can be generated, and after splicing into a new commodity main view, it is replaced and displayed in the commodity main view display area. In this embodiment four, the generation of the user upper body effect picture and the replacement and display in the main view display area can also not depend on the generation of the multi-view view. Specifically, this embodiment four provides a commodity information display method, see Figure 7 The method can specifically include: S701: providing a commodity information page, the commodity information page is used to display commodity information of at least one commodity, the commodity includes a clothing commodity, and the commodity information page includes a commodity main view display area; The commodity information page also includes an operation option for viewing the user upper body effect of the commodity; S702: after receiving a request for viewing the user upper body effect of a target commodity through the operation option, generating a user upper body effect picture of the target commodity according to the commodity main view of the target commodity and a three-dimensional model of a user image; The three-dimensional model of the user image is generated by an AI generation model according to a user image photo and / or body shape data uploaded by the user; S703: regenerating a commodity main view according to the user upper body effect picture of the target commodity, for replacing and displaying in the commodity main view display area of the target commodity in the commodity information page.

[0064] That is to say, in this embodiment four, the commodity main view display mode in the default state is not limited, as long as the user upper body effect picture of the target commodity can be generated according to the commodity main view of the target commodity and the three-dimensional model of the user image, and further generating a new commodity main view, replacing and displaying in the commodity main view display area, which is within the protection scope of this embodiment four.

[0065] Embodiment five This embodiment five provides a commodity picture information processing system, see Figure 8 The device can include: The model module 801 is used to provide an AI generation model; The data module 802 is used to collect data to generate a training data set, and train the AI generation model, so that the AI generation model obtains the ability to generate a plurality of view commodity pictures, the training data includes commodity information service system commodity picture and corresponding label information, clothing metadata; wherein the label information of the commodity picture includes the shooting angle information corresponding to the commodity picture; The image module 803 is configured to determine a plurality of view angle product images for a clothing product, wherein the product images in at least some view angles are generated by the AI generation model based on product information provided by a product publisher user associated with the target product; The guide module 804 is configured to determine a target product to be displayed when a product information page needs to be displayed, and obtain the plurality of view angle product images corresponding to the target product, generate a main product image of the target product by splicing the plurality of view angle product images, and display the main product image in a main product image display area of the product information page; provide an operation option for viewing a user body effect of the product in the product information page; after receiving a request for viewing the user body effect of the target product through the operation option, generate a user body effect image of the target product in a plurality of view angles based on the plurality of view angle product images corresponding to the target product and a three-dimensional model of a user image by the image module; the three-dimensional model of the user image is generated by a second AI generation model based on a user image photo and / or body shape data uploaded by a user; and regenerate a main product image based on the user body effect image of the target product in a plurality of view angles, and display the main product image in the main product image display area of the target product in the product information page.

[0066] For the parts not described in the above embodiments two to five, please refer to the description in the embodiment one and other parts of the specification, which will not be repeated here. It should be noted that the embodiments of the present application can involve the use of user data. In actual application, user-specific personal data can be used in the schemes described herein, as long as the applicable laws and regulations in the country are met (for example, the user's explicit consent, the user's actual notification, etc.), and the use of user-specific personal data is within the scope allowed by the applicable laws and regulations.

[0067] Corresponding to the embodiment one, the present application also provides a product information display device, which can include: A page providing unit is configured to provide a product information page, the product information page is configured to display product information of at least one product, the product includes a clothing product, and the product information page includes a main product image display area, wherein the main product image displayed in the main product image display area is generated by splicing a plurality of view angle product images, and the product images in at least some view angles are generated by a first AI generation model based on product information provided by a product publisher user; the product information page also includes an operation option for viewing a user body effect of the product; The user upper body effect picture generation unit is configured to, after receiving a request for viewing a user upper body effect of a target commodity through the operation option, generate user upper body effect pictures of the target commodity in multiple perspectives according to the multiple commodity pictures in the multiple perspectives corresponding to the target commodity and the three-dimensional model of the user image; and the three-dimensional model of the user image is generated by a second AI generation model according to a user image photo and / or body shape data uploaded by a user. The commodity main picture replacement unit is configured to regenerate a commodity main picture according to the user upper body effect pictures of the target commodity in the multiple perspectives, and replace the commodity main picture of the target commodity in a commodity main picture display area of the commodity information page for display.

[0068] Specifically, the page providing unit can be specifically configured to: determine a to-be-displayed commodity in a commodity information page provided by a first commodity information service system; the first commodity information service system includes a system providing country-specific commodity information services for multiple country / region buyer users; obtain commodity detail information uploaded by a publisher user associated with the to-be-displayed commodity in a second commodity information service system for the to-be-displayed commodity, the commodity detail information at least including commodity pictures; the second commodity information service system is a system providing commodity information services for a single country / region buyer user; process the commodity detail information by using the first AI generation model, and generate commodity pictures in multiple perspectives for the to-be-displayed commodity, so as to generate a commodity main picture of the to-be-displayed commodity by pasting the commodity pictures in the multiple perspectives, and display the commodity main picture in the commodity information page.

[0069] Alternatively, the page providing unit can be specifically configured to: if the commodity information uploaded by the publisher user of the to-be-displayed commodity does not include an upper body effect picture of a model character on a commodity, after generating commodity pictures in at least part of the perspectives by using the first AI generation model, match a virtual model character image for the to-be-displayed commodity, and generate upper body effect pictures of the commodity pictures in the multiple perspectives on the virtual model character image, so as to splice the upper body effect pictures of the virtual model character images in the multiple perspectives into a commodity main picture.

[0070] Specifically, the virtual model character image is generated by a third AI generation model according to multiple commodity styles and a country-specific model characteristic database; when matching a virtual model character image for a target commodity, a matching virtual model character image is determined according to clothing metadata information of the target commodity and country-specific information of the target user.

[0071] The first AI generation model is trained in advance based on training data, and has the ability to generate multiple perspective product images. The training data includes product images and corresponding label information of multiple products in a product information service system, and clothing metadata. The label information of the product image includes shooting perspective information corresponding to the product image.

[0072] In an implementation, the first AI generation model can be trained by inputting product images of multiple products and corresponding generation instructions to the first AI generation model, to instruct the first AI generation model to generate multiple perspective product images according to the input product images, and to perform multiple rounds of iterative optimization on the generation results using the label information.

[0073] In addition, the device can further include: The try-on posture information determination unit is configured to determine try-on posture information of a three-dimensional model of a user image according to clothing metadata information of the target product when generating the user upper body effect image of the target product under multiple perspectives, so as to generate multiple perspective user upper body effect images of the user image trying on the target product under corresponding try-on postures.

[0074] Corresponding to Embodiment Two, the present application also provides a product information display device, which can include: The page display unit is configured to display a product information page, the product information page being configured to display product information of at least one product, the product including a clothing product, and the product information page including a product main image display area, wherein the displayed product main image is generated by stitching multiple perspective product images, and at least part of the perspective product images are generated by the first AI generation model according to product information provided by a product publisher user; the product information page further includes an operation option for viewing a user upper body effect; The request submission unit is configured to submit an operation request of a user to a server through the operation option, to generate a user upper body effect image of the target product under multiple perspectives according to multiple perspective product images corresponding to the target product and a three-dimensional model of a user image, and to return the product main image regenerated according to the user upper body effect image of the target product under multiple perspectives to the client; the three-dimensional model of the user image is generated by the second AI generation model according to a user image photo and / or body shape data uploaded by the user; The replacement display unit is configured to display the regenerated product main image in the main image display area of the product information page.

[0075] Corresponding to Embodiment Three, the present application also provides a product search result display device, which can include: The commodity search result page providing unit is configured to provide a commodity search result page after receiving a search request about a clothing commodity, the commodity search result page including a plurality of resource positions for displaying a plurality of commodity information cards of commodities, the commodity information cards including a commodity main image display area, wherein the commodity main image displayed is generated by stitching a plurality of view commodity images, and at least part of the view commodity images are generated by a first AI generation model based on commodity information provided by a commodity publisher user; the commodity search result page further includes an operation option for viewing a user body effect of a commodity provided in a resource position unit; The user body effect image generation unit is configured to submit a request to a server after receiving a request for viewing a user body effect of a target commodity through the operation option, so that the server generates a user body effect image of the target commodity in a plurality of views based on a plurality of view commodity images corresponding to the target commodity and a three-dimensional model of a user image, and returns the target commodity to the client after regenerating a commodity main image based on the user body effect image of the target commodity in a plurality of views; the three-dimensional model of the user image is generated by a second AI generation model based on a user image photo and / or body shape data uploaded by a user; The commodity main image replacement unit is configured to replace the regenerated commodity main image to the main image display area in the commodity information card of the target commodity for display.

[0076] Corresponding to the fourth embodiment, the application also provides a commodity information display device, which can include: The commodity information page providing unit is configured to provide a commodity information page for displaying commodity information of at least one commodity, the commodity including a clothing commodity, the commodity information page including a commodity main image display area; the commodity information page further includes an operation option for viewing a user body effect of a commodity; The user body effect image generation unit is configured to generate a user body effect image of a target commodity based on a commodity main image of the target commodity and a three-dimensional model of a user image after receiving a request for viewing a user body effect of the target commodity through the operation option; the three-dimensional model of the user image is generated by an AI generation model based on a user image photo and / or body shape data uploaded by a user; The commodity main image replacement display unit is configured to regenerate a commodity main image based on the user body effect image of the target commodity, so as to replace the commodity main image display area of the target commodity in the commodity information page for display.

[0077] In addition, the embodiment of the present application further provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to implement the steps of the method in any one of the foregoing method embodiments.

[0078] and an electronic device comprising: one or more processors; and a memory associated with the one or more processors, the memory configured to store program instructions that, when executed by the one or more processors, perform the steps of the method in any one of the foregoing method embodiments.

[0079] A computer program product comprising computer program / computer executable instructions that, when executed by a processor in an electronic device, implement the steps of the method in the foregoing method embodiments.

[0080] wherein, Figure 9 An exemplary shows the architecture of an electronic device, for example, the device 900 can be a mobile phone, a computer, a digital broadcast terminal, a messaging device, a game console, a tablet device, a medical device, a fitness device, a personal digital assistant, an aircraft, etc.

[0081] Referring to Figure 9 , the device 900 can include one or more of the following components: a processing component 902, a memory 904, a power supply component 906, a multimedia component 908, an audio component 910, an input / output (I / O) interface 912, a sensor component 914, and a communication component 916.

[0082] The processing component 902 usually controls the overall operation of the device 900, such as operations associated with displaying, making phone calls, data communications, camera operations, and recording operations. The processing component 902 can include one or more processors 920 to execute instructions to complete all or part of the steps of the methods provided by the technical solutions of the present disclosure. In addition, the processing component 902 can include one or more modules to facilitate the interaction between the processing component 902 and other components. For example, the processing component 902 can include a multimedia module to facilitate the interaction between the multimedia component 908 and the processing component 902.

[0083] The memory 904 is configured to store various types of data to support operations of the device 900. Examples of such data include instructions for any application or methods operating on the device 900, contact data, phonebook data, messages, pictures, videos, and so on. The memory 904 can be implemented by any type of volatile or nonvolatile storage devices or a combination thereof such as static random access memory (SRAM), electrically erasable programmable read only memory (EEPROM), erasable programmable read only memory (EPROM), programmable read only memory (PROM), read only memory (ROM), magnetic memory, flash memory, magnetic disk, or optical disk.

[0084] The power component 906 provides power to the various components of the device 900. The power component 906 can include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power for the device 900.

[0085] The multimedia component 908 includes a screen providing an output interface between the device 900 and a user. In some embodiments, the screen includes a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touch, swiping, and gestures on the touch panel. The touch sensors can not only sense a boundary of a touch or swiping action, but also detect duration and pressure related to the touch or swiping action. In some embodiments, the multimedia component 908 includes a front camera and / or a rear camera. The front and / or rear camera can receive external multimedia data when the device 900 is in an operating mode, such as a shooting mode or a video mode. Each of the front and rear camera can be a fixed optical lens system or have a focal length and optical zoom capability.

[0086] The audio component 910 is configured to output and / or input audio signals. For example, the audio component 910 includes a microphone (MIC) configured to receive external audio signals when the device 900 is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode. The received audio signals can be further stored in the memory 904 or transmitted via the communication component 916. In some embodiments, the audio component 910 also includes a speaker for outputting audio signals.

[0087] The I / O interface 912 provides an interface between the processing component 902 and peripheral interface modules, which can be a keyboard, a click wheel, a button, and so on. The buttons can include, but are not limited to, a home button, a volume button, a start button, and a lock button.

[0088] The sensor component 914 includes one or more sensors for providing status assessments for various aspects of the device 900. For example, the sensor component 914 can detect an open / closed position of the device 900, relative positioning of components of the device 900, such as a display and keypad of the device 900, changes in position of the device 900 or a component of the device 900, presence or absence of user contact with the device 900, orientation or acceleration / deceleration of the device 900, and temperature changes of the device 900. The sensor component 914 can include proximity sensor(s) configured to detect presence of nearby objects without any physical contact. The sensor component 914 can also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 914 can also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

[0089] The communication component 916 is configured to facilitate wired or wireless communication between the device 900 and another device. The device 900 can access a wireless network based on a communication standard, such as WiFi, or a mobile communication network standard, such as Global System for Mobile Communication (GSM), Code Division Multiple Access (CDMA), Wideband Code Division Multiple Access (WCDMA), Long Term Evolution (LTE), and others. In an exemplary embodiment, the communication component 916 receives a broadcast signal or broadcast related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component 916 further includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) technology, infrared data association (IrDA) technology, Ultra-Wide Band (UWB) technology, Bluetooth (BT) technology and other technologies.

[0090] In an exemplary embodiment, the device 900 can be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, micro-controllers, microprocessors or other electronic elements, to perform the methods described above.

[0091] In an exemplary embodiment, a non-transitory computer readable storage medium, such as the memory 904 including instructions, is also provided, which can be executed by the processor 920 of the device 900 to complete the methods provided by the techniques of this disclosure. For example, the non-transitory computer readable storage medium can be a ROM, a random access memory (RAM), a CD-ROM, a magnetic tape, a floppy disc, and an optical data storage device, etc.

[0092] Those skilled in the art can clearly understand the application by the description of the above embodiments that the application can be implemented by means of software and the necessary universal hardware platforms. Based on such an understanding, the technical solutions of the application can be embodied in the form of a software product, and the computer software product can be stored in a storage medium, such as a ROM / RAM, a magnetic disk, an optical disk, and the like, and includes a plurality of instructions to make a computer device (which can be a personal computer, a server, or a network device, and the like) execute the methods described in each embodiment or some parts of the embodiments of the application.

[0093] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each embodiment focuses on the difference from other embodiments. In particular, for the system or the system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the related parts can be referred to the part of the method embodiment. The above-described system and system embodiment are merely illustrative, and the units described as separate components can be or can not be physically separated, and the components displayed as units can be or can not be physical units, that is, they can be located in one place, or can be distributed on multiple network units. According to the actual needs, some or all of the modules can be selected to achieve the purpose of the embodiment. Those skilled in the art can understand and implement without creative labor.

[0094] The above describes the product information display method and the electronic device provided by the application in detail, and the principle and implementation of the application are described by applying specific examples. The above embodiment is only used to help understand the method and the core idea of the application; meanwhile, for those skilled in the art, according to the idea of the application, the specific implementation and application range can be changed. In conclusion, the content of the specification should not be understood as a limitation of the application.

Claims

1. A method for displaying product information, characterized in that: include: Providing a product information page for displaying product information of at least one product, wherein the product includes apparel. The product information page includes a main product image display area, wherein the displayed main product image is generated by piecing together product images from multiple perspectives, wherein at least some of the product images from some perspectives are generated by a first AI generation model based on product information provided by the product publisher. The product information page also includes an option for viewing how the product would look on a user. After receiving a request to view the user's body image for the target product through the operation option, generating images of the target product's body image from multiple perspectives based on the product images corresponding to the target product and the three-dimensional model of the user's image; the three-dimensional model of the user's image is generated by a second AI generation model based on the user's image photo and / or body shape data uploaded by the user; The main image of the product is regenerated based on the user's upper body effect images of the target product under multiple viewing angles, so as to be replaced in the product main image display area of ​​the target product in the product information page for display.

2. The method according to claim 1, characterized in that The product information page provided includes: Determining products to be displayed on a product information page provided by a first product information service system; the first product information service system includes: a system that provides country-specific product information services to buyers in multiple countries / regions; Obtaining product details uploaded by a publisher associated with the product to be displayed in a second product information service system, wherein the product details include at least a product image; the second product information service system is a system that provides product information services to buyers in a single country / region; The product details information is processed using the first AI generation model to generate product images from multiple perspectives for the product to be displayed, so as to generate the main product image of the product to be displayed by piecing together the product images from multiple perspectives and display it on the product information page.

3. The method according to claim 2, characterized in that The product information page further includes: If the product information uploaded by the publisher user of the product to be displayed does not include a model's upper body rendering of the product, then after the first AI generation model generates a product image from at least a partial perspective, a virtual model image is also matched for the product to be displayed, and upper body renderings of the product images from multiple perspectives on the virtual model image are generated, so that the upper body renderings of the virtual model image corresponding to the multiple perspectives are spliced ​​into the main product image.

4. The method according to claim 3, characterized in that The virtual model character image is generated by a third AI generation model based on a database of model characteristics of multiple product styles and countries; When matching a virtual model character image for a target product, the matching virtual model character image is determined based on the clothing metadata information of the target product and the nationalization information of the target user.

5. The method according to claim 1, wherein The first AI generation model is pre-trained based on training data to obtain the ability to generate product images from multiple perspectives. The training data includes product images of multiple products in the product information service system and corresponding annotation information and clothing metadata; wherein the annotation information of the product images includes the shooting perspective information corresponding to the product images.

6. The method according to claim 5, characterized in that The first AI generation model is trained in the following manner: multiple product images and corresponding generation instructions are input into the first AI generation model to instruct the first AI generation model to generate multiple perspective product images based on the input product images, and the generation results are optimized for multiple rounds of iterations using the annotation information.

7. The method according to claim 1, characterized in that Also includes: When generating renderings of the target product on the user's upper body from multiple perspectives, the fitting posture information of the three-dimensional model of the user image is determined based on the clothing metadata information of the target product, so as to generate renderings of the user's upper body from multiple perspectives in which the user image tries on the target product in the corresponding fitting postures.

8. A method for displaying product information, characterized in that: include: Displaying a product information page, the product information page is used to display product information of at least one product, the product including apparel, the product information page including a main product image display area, wherein the displayed main product image is generated by piecing together product images from multiple perspectives, wherein at least some of the product images from some perspectives are generated by a first AI generation model based on product information provided by the product publisher; the product information page also includes an operation option for viewing the user's body effect; After receiving the user's operation request through the operation option, the request is submitted to the server to generate user upper body renderings of the target product from multiple perspectives based on the product images of the target product from multiple perspectives and the three-dimensional model of the user's image, and regenerate the product main image based on the user upper body renderings of the target product from multiple perspectives and return it to the client; the three-dimensional model of the user's image is generated by the second AI generation model based on the user's image photo and / or body shape data uploaded by the user; The regenerated main image of the product is replaced in the main image display area of ​​the product information page for display.

9. A method for displaying product search results, characterized in that: include: Upon receiving a search request for apparel products, a product search results page is provided. The product search results page includes multiple resource slots for displaying product information cards for multiple products. The product information cards include a main product image display area, where the displayed main product image is generated by piecing together product images from multiple perspectives, wherein at least some of the product images from some perspectives are generated by a first AI generation model based on product information provided by the product publisher user. The product search results page also includes an operation option, provided in resource slots, for viewing how the product would look on a user. After receiving a request for viewing the user's upper body effect for the target product through the operation option, the request is submitted to the server, so that the server generates the user's upper body effect images of the target product from multiple perspectives based on the product images of the target product from multiple perspectives and the three-dimensional model of the user's image, and regenerates the product main image based on the user's upper body effect images of the target product from multiple perspectives and returns it to the client; the three-dimensional model of the user's image is generated by a second AI generation model based on the user's image photo and / or body shape data uploaded by the user; The regenerated product main image is replaced in the main image display area of ​​the product information card of the target product for display.

10. A method for displaying product information, characterized in that: include: Providing a product information page for displaying product information of at least one product, wherein the product includes clothing products, and the product information page includes a main product image display area; the product information page also includes an operation option for viewing the user's body effect of the product; After receiving a request to view how the target product would look on a user through the operation option, generating a user-worn image of the target product based on the target product's main image and a three-dimensional model of the user's image; the three-dimensional model of the user's image is generated by an AI generation model based on the user's image photo and / or body shape data uploaded by the user; The main product image is regenerated based on the user's body effect image of the target product, so as to be replaced in the product main image display area of ​​the target product in the product information page for display.

11. A product image information processing system, characterized in that: include: Model module, used to provide AI generation model; A data module is configured to collect data to generate a training data set and train the AI ​​generation model to enable the AI ​​generation model to generate product images from multiple perspectives, wherein the training data includes product images of multiple products in the product information service system and corresponding annotation information and apparel metadata; wherein the annotation information of the product images includes the shooting perspective information corresponding to the product images; An image module, configured to determine product images from multiple perspectives for a clothing product, wherein the product images from at least some perspectives are generated by the AI ​​generation model based on product information provided by a product publisher user associated with the target product; A shopping guide module is used to determine the target product to be displayed when it is necessary to display a product information page, and obtain the multiple perspective product images corresponding to the target product, and generate a main product image of the target product by splicing the product images of the multiple perspectives, for display in the product main image display area of ​​the product information page; an operation option for viewing the user's upper body effect of the product is provided in the product information page; after receiving a request to view the user's upper body effect for the target product through the operation option, the image module generates the user's upper body effect image of the target product under multiple perspectives based on the product images of the multiple perspectives corresponding to the target product and the three-dimensional model of the user image; the three-dimensional model of the user image is generated by a second AI generation model based on the user image photos and / or body shape data uploaded by the user; the product main image is regenerated based on the user's upper body effect image of the target product under multiple perspectives, for replacement in the product main image display area of ​​the target product in the product information page for display.

12. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method according to any one of claims 1 to 10 are implemented.

13. An electronic device, characterized in that: include: one or more processors; as well as A memory associated with the one or more processors, the memory being used to store program instructions, wherein when the program instructions are read and executed by the one or more processors, the steps of the method according to any one of claims 1 to 10 are performed.

14. A computer program product comprising a computer program / computer executable instructions, characterized in that When the computer program / computer executable instructions are executed by a processor in an electronic device, the steps of the method according to any one of claims 1 to 10 are implemented.

Images