CN117332105A - Method and electronic device for providing product matching information - Google Patents

Abstract

Embodiments of the present application disclose a method and an electronic device for providing product matching information. The method includes: determining a first image to be matched with clothing; the first image includes: the first clothing and its related information when the human body is wearing it. The first image content of the first partial human body part; from the preset material library, determine at least one second image that can be matched; the second image includes: the second clothing and its related clothing in the human body wearing state The second image content of the second partial human body part; splicing the first image content and the second image content; using the artificial intelligence AI large-scale parameter model to redraw the spliced image , to generate the target image. Through the embodiments of the present application, more flexible clothing matching recommendation information can be provided, and the output information matching results can serve as a reference more effectively.

Description

Method and electronic device for providing product matching information

Technical field

This application relates to the field of information processing technology, and in particular to methods and electronic devices for providing product matching information.

Background technique

In the product information service system, clothing is an important product category. When users purchase clothing products, finding a suitable match based on their favorite clothing is also a common need in users' life scenarios. For example, if a user is interested in a top, they usually also want to match the top with a bottom, and so on.

In order to meet users' needs for clothing matching purchases, some systems can provide users with clothing information. However, this clothing information is usually pre-matched by merchants based on the clothing in their own stores, and then displayed on the product details page. Recommended in-store clothing matching solutions can be provided on other pages. Alternatively, a special outfit channel may be provided to recommend cross-store matching of clothing. However, in the existing technology, it is usually still necessary to manually match the clothes offline in advance and then publish them. Consumers can Browse specific matching plans in this channel, etc. In short, the above solutions are not flexible enough.

Contents of the invention

This application provides a method and electronic device for providing product matching information, which can provide more flexible clothing matching recommendation information, and can make the output information matching results serve as a reference more effectively.

This application provides the following solutions:

A method of providing product matching information, including:

Determine the first image to be matched with clothing; the first image includes: the first image content of the first clothing and its related first partial human body part in the human body wearing state;

Determine at least one second image that can be matched from the preset material library; the second image includes: the second image content of the second garment and its related second partial human body part in the human body wearing state;

performing splicing processing on the first image content and the second image content;

The artificial intelligence AI large-scale parameter model is used to redraw the spliced image to generate a target image. The target image is used to simulate the wearing state of the first garment and the second garment on the same human body. Matching effect.

Wherein, the images stored in the material library correspond to the commodities provided in the associated information service system;

Determining the first image to be matched with clothing includes:

Provide product search results based on product search requests submitted by users;

In response to a clothing matching request initiated for a designated product in the product search results, a target image associated with the designated product in the material library is determined as the first image.

Wherein, the product search request is: a product search request submitted based on clothing-related images uploaded by the user, and the product search results include products of the same style as the clothing body included in the image uploaded by the user.

Wherein, the splicing process of the first image content and the second image content includes:

Identify clothing key points and human body key point information contained in the first image content and the second image content;

According to the identified clothing key points and human body key point information, the first image content and the second image content are spliced.

Wherein, when the target image is displayed through the client, an operation option for searching for same-style products for the second clothing in the target image is also provided, so that after receiving the same-style product search request, the same-style product is provided. search results.

Wherein, when the target image is displayed through the client, operation options for designing the clothing image in the target image are also provided;

The method also includes:

After receiving a request for custom design of the clothing image in the target image, providing optional design elements to add the target design element to the clothing image;

In response to the request for image generation of the design results, the AI large-scale parameter model is used to redraw the designed target image in order to integrate the clothing image in the target image with the target design elements to generate a realistic image. Costume effect image.

Among them, it also includes:

In response to a request to perform a same-style product search on the designed clothing image, the same-style product search results are provided.

An image processing method including:

After receiving a request to design the clothing image in the target image, provide candidate design elements for designing the clothing image, so as to add the target design element to the clothing image;

In response to the request for image generation of the design results, the AI large-scale parameter model is used to redraw the designed target image in order to integrate the clothing image in the target image with the target design elements to generate a realistic image. Costume effect image.

Among them, it also includes:

In response to a request to perform a same-style product search on the designed clothing image, the same-style product search results are provided.

A method of providing clothing matching information, including:

Receive a clothing matching request, which includes the required target clothing style and/or scene information, as well as text and/or image information used to describe the target person's human characteristics;

Determine a clothing matching scheme that meets the needs of the target dressing style and/or scene, and use an AI large-scale parameter model to generate a target image, the target image being used to express the clothing matching scheme when worn by the target person matching effect.

Among them, it also includes:

In response to a request to search for the same type of products for the clothing in the clothing matching scheme, search results for the same type of products are provided.

A device that provides product matching information, including:

The first image determination unit is used to determine the first image to be matched with clothing; the first image includes: the first image content of the first clothing and its related first partial human body part in the human body wearing state;

The second image determination unit is used to determine at least one second image that can be matched from the preset material library; the second image includes: the second clothing and its related second partial human body when the human body is wearing it. The second image content of the part;

A splicing processing unit configured to splice the first image content and the second image content;

The AI redrawing unit is used to use the artificial intelligence AI large-scale parameter model to redraw the spliced image to generate a target image. The target image is used to simulate the first clothing and the second clothing in the same The matching effect is carried out on the human body when worn.

An image processing device, including:

A design element providing unit, configured to provide candidate design elements for designing the clothing image after receiving a request to design the clothing image in the target image, so as to add the target design element to the clothing image;

The AI redrawing unit is configured to respond to a request for image generation of the design result and use the AI large-scale parameter model to redraw the designed target image, so as to match the clothing image in the target image with the target design. Elements are merged to generate images with realistic clothing effects.

A device that provides clothing matching information, including:

A request receiving unit, configured to receive a clothing matching request, which includes the required target dressing style and/or scene information, as well as text and/or image information used to describe the human body characteristics of the target person;

An AI generation unit is used to determine a clothing matching scheme that meets the needs of the target dressing style and/or scene, and uses an AI large-scale parameter model to generate a target image, where the target image is used to express the clothing matching scheme based on the requirements of the target clothing style and/or scene. Describe the matching effect when the target person is wearing it.

A computer-readable storage medium on which a computer program is stored, which implements the steps of any of the foregoing methods when executed by a processor.

An electronic device including:

one or more processors; and

A memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform any of the foregoing methods. A step of.

According to the specific embodiments provided in this application, this application discloses the following technical effects:

Through the embodiments of the present application, a second image that can be matched with other first clothing can be determined for the first image specified by the user, and both the first image and the second image can include image content of specific clothing. and the image content of relevant local human body parts in the human body wearing state. In this way, when the matching results are returned, the above image content can be spliced and redrawn through the AI large-scale parameter model to generate a model for simulating the matching of the first garment and the second garment when worn on the same human body. The target image for the effect. In this way, it can not only recommend matching objects for the user, but also provide the matching effect of the first clothing and the second clothing in the same matching scheme when worn on the same body. Moreover, this matching effect will be more real and natural, and can be more Effectively serve as a reference.

Of course, implementing any product of this application does not necessarily require achieving all the above-mentioned advantages at the same time.

Description of drawings

In order to more clearly explain the embodiments of the present application or the technical solutions in the prior art, the drawings needed to be used in the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some of the drawings of the present application. Embodiments, for those of ordinary skill in the art, other drawings can also be obtained based on these drawings without exerting creative efforts.

Figure 1 is a schematic diagram of the system architecture provided by the embodiment of the present application;

Figure 2 is a flow chart of the first method provided by the embodiment of the present application;

Figure 3 is a schematic diagram of the AI matching interface provided by the embodiment of the present application;

Figure 4 is a schematic diagram of the process of generating matching images by AI provided by the embodiment of the present application;

Figure 5 is a schematic diagram of an advanced matching interface provided by an embodiment of the present application;

Figure 6 is a schematic diagram of the "AI Brush" functional interface provided by the embodiment of the present application;

Figure 7 is a schematic diagram of another interface of the "AI brush" function provided by the embodiment of the present application;

Figure 8 is a flow chart of the second method provided by the embodiment of the present application;

Figure 9 is a flow chart of the third method provided by the embodiment of the present application;

Figure 10 is a schematic diagram of the scene matching function interface provided by the embodiment of the present application;

Figure 11 is a schematic diagram of the community platform interface provided by the embodiment of the present application;

Figure 12 is a schematic diagram of an electronic device provided by an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art fall within the scope of protection of this application.

In the embodiments of this application, in order to provide users with more flexible clothing matching information, a corresponding solution is provided. In this solution, for the clothing image specified by the user, the clothing can be provided with a wider range of information. Recommended combinations, not limited to within the same store. It should be noted here that in some solutions in the existing technology, there may be solutions for smart matching, but these are usually based on a list of recommended matching clothing for the user to choose. For example, the user initiates smart matching based on a certain top. After requesting, you can return to a product list page, which displays information about pants, skirts and other clothing recommended by the algorithm that can be matched with the top. However, the specific matching effect of this top with the specific pants, skirts and other clothing in the list depends on the user's own imagination. Obviously, this will not be conducive to the user's further decision-making.

In order to help users make decisions, some solutions can also provide recommended matching solutions while also providing the upper body effects of specific matching solutions on the human body. For example, if a top is matched with a pair of pants, the top and pants can be displayed. The effect when worn on models and other characters can make the specific matching effect more intuitive. In this solution, material pictures need to be prepared in advance for specific clothing products. This material picture is usually a picture of the clothing with a three-dimensional display effect (for example, the clothing displayed in the picture can be a photo taken in a propped-up state. , instead of shooting in a flat state such as laying flat, etc.), after determining that a certain garment can be matched with another garment, the three-dimensional pictures corresponding to the two garments can be matched to a pre-generated mannequin model. This simulates the specific upper body matching effect.

Although the above solution can provide a more intuitive matching effect, in this implementation of matching three-dimensional pictures of different clothing to a mannequin, since the mannequin is virtual, during the matching process, the clothing is usually just If the three-dimensional picture is moved to a certain position on the human model, the generated image will appear stiff, not realistic and natural enough, and cannot show the state of specific clothing when worn by the real human body. For example, some fabrics naturally droop with gravity. etc.

Based on the above situation, in the embodiment of the present application, an improved solution is provided. In this solution, intelligent matching recommendations can still be provided for any product specified by the user, and matching renderings of different garments in the specific matching scheme when worn by the human body can be provided. However, it does not match the three-dimensional images of the garments. to the pre-generated virtual human model, but cut out, crop, and splice the real model images corresponding to the specific clothing, and then use the ability of AI (Artificial Intelligence, artificial intelligence) large-scale parameter model to edit the spliced images. Redraw to make the image more realistic and natural.

Specifically, in the embodiment of the present application, a material library can be established in advance. The material library can be mainly established based on the upper body pictures of real models of specific clothing products. Of course, the upper body pictures of real models here can be the current product information service system. The image corresponding to the specific product in the image, or it can also be an image collected through other means. That is, the clothing in the image may not be the product currently on sale in the product information service system. However, as long as the clothing can be collected The real-life upper body picture can be used to generate the material picture in the embodiment of this application. For example, when establishing this material library, you can collect photos of multiple garments taken when worn by real models, etc., and then cut them out to extract images of specific garments and their images when worn by real models. Relevant human body parts, for example, in a photo, model A is wearing a top C1 and a pair of pants T1. At this time, you can extract the image of the top and the model's upper body. Of course, you can also extract it from the image. Take out the image of the pants and the model's lower body. In this way, the following material pictures can be added to the material library: the image of top C1 on model A. The image includes model A's head, hands, etc. and the top. Relevant body parts, excluding legs, feet and other body parts that have nothing to do with wearing tops; in addition, it can also include images of pants T1 worn on model B. The image includes legs, feet, etc. of model A that are related to wearing pants. body parts, excluding head, arms and other body parts that have nothing to do with wearing pants, etc. In this way, multiple material pictures can be saved in the material library, and each material picture can include images of specific clothing when worn by real models, as well as partial body parts of the specific real models involved. Or, in addition to generating material pictures through the above-mentioned processing of photos of real clothing, virtual material pictures can also be generated through AI large-scale parameter models, and so on.

In this way, when making matching recommendations based on a certain clothing product specified by the user, firstly, relevant algorithms can be used to determine other clothing products that can be matched with the clothing product, and then the material images corresponding to these clothing products can be found from the material library; or , you can also use the ability of AI large-scale parameter models to determine the specific material pictures that can be matched by understanding the material pictures in the material library, etc. After that, one or more sets of matching plans can be generated. For each set of matching plans, the material pictures corresponding to different clothing can be spliced together. For example, if the user specifies the recommended matching clothing for a certain top, then multiple matching items can be determined. After matching the pants, skirts, etc., the material pictures corresponding to the top can be spliced together with the material pictures corresponding to the pants, skirts, etc. Since the material pictures corresponding to the tops include the upper body of the real model and other body parts, the corresponding material pictures of the pants, skirts, etc. The material picture includes the lower body of the real model and other body parts. Therefore, by splicing the two material pictures together, a complete human body can be formed, and the matching effect of specific clothing when the human body is worn is presented.

In this way, since the specific material pictures are all images of clothing in the state of being worn by real models, compared to the upper body effect that is synthesized through three-dimensional pictures and virtual characters, it can better reflect the upper body state of specific clothing. sense of reality. Among them, in order to obtain a better splicing effect, you can also identify key points of clothing and human body key points in different material images during the splicing process. Then, based on this key point information, image simulation can be performed during image splicing. changes in reflection to make the spliced image look more realistic and natural.

In addition, due to the splicing process, the material pictures corresponding to different clothing in the same matching scheme may be different models. Even the same model may be shot at different times and locations. Therefore, even if The affine change ability of the splicing algorithm is achieved, but it may also cause the spliced image to be somewhat unnatural. For example, the brightness of the upper and lower clothes, the color of the model's skin, etc. are not uniform, making the overall image look uncoordinated, or some body parts of the model (for example, hands, etc.) may have been modified during the cutout process. The extraction is not complete enough, etc. Therefore, in the embodiment of the present application, the drawing capability of the AI large-scale parameter model can be further used to redraw on the basis of the above-mentioned spliced images, thereby making the specific images more beautiful and realistic, and better simulating the The matching effects of different clothing on the same human body, etc.

That is to say, in the embodiment of the present application, on the basis of the clothing matching ability, by splicing the upper body pictures of real models of clothing, and with the help of the drawing ability of the AI large-scale parameter model, users can be provided with a more realistic and natural Clothes are matched with upper body renderings to help users make decisions more effectively.

Among them, in actual applications, there can be a variety of specific application scenarios. For example, in the scenario of "searching for products by pictures", the user can upload a photo with an image of clothing, etc., and initiate a search. At this time, you can Return to the list of similar products of this clothing. When users browse these products of the same type, they may be interested in one of the products, but want to know what clothing can match the product. Therefore, operation portals such as "AI matching" can be provided for each product in the product list. Users can initiate matching requests through this operation portal. Correspondingly, through the splicing of the aforementioned material images, AI redrawing, etc., the user can The user generates multiple recommended matching plans, and the user can view the matching effects of different clothing in the specific matching plan when the human body is wearing it, and so on.

Of course, the material pictures in the material library can be material pictures corresponding to some representative clothing. After generating the images corresponding to the above matching scheme, if the user is interested in the matching object in one of the images, he can continue to initiate a target For requests such as "find the same style" of the matching object, corresponding pages such as a list of products of the same style for the matching object can be displayed, thereby helping the user to more quickly find other clothing that can match the clothing of current interest.

In addition, in an optional way, the "AI matching" function can also be provided to users, allowing users to design based on specific clothing or matching plans. For example, on the basis of the aforementioned "AI matching" function, the "AI brush" function can also be used to allow users to add more design elements based on the images generated by AI, including adding some stickers and patterns on clothing. Or, color the clothing, or replace the matching objects, etc. After that, you can also use the drawing capabilities of the AI large-scale parameter model to redraw the designed image to make the generated image more realistic and natural (otherwise, It may look like cartoon stickers affixed to the clothing photos, and it may not look like the actual patterns on the clothing, etc.), etc.

Furthermore, it can also provide the function of AI-generated clothing matching images based on specific scenes and user-specified character characteristics. That is, a variety of optional scenes can be provided in advance for users to choose. After selecting a scene, they can also use text and / Or upload images and other methods to describe the characteristics of the desired character. Afterwards, the AI large model can be used to generate clothing matching images that match the scene and the characteristics of the character, etc.

Regarding the above-mentioned "AI matching", "AI matching", scene matching and other functions, the images generated by AI can also be published through community network platforms, etc. In this platform, you can also view images posted by other users, or initiate Find the same style, etc.

In short, the embodiments of this application can apply the capabilities of the AI large-scale parameter model to scenarios related to clothing matching recommendations, by using the AI large-scale parameter model to "generate images based on images", "generate images based on text", and "generate images based on text + "Image-generated image" and other methods provide users with richer and more realistic matching recommendations.

Among them, AI large-scale parameter model can also be referred to as AI large model, which can refer to a type of foundation model. Specifically, it can refer to a large number of parameters trained using massive data and capable of adapting to a series of downstream tasks. Model. For large AI models, not only is there a huge number of parameters in terms of parameter scale (as the model continues to iterate, the number of parameters usually increases exponentially, from billions to trillions, to trillions, and even more (Multiple) characteristics, and from the perspective of modal support, AI large models have gradually developed from supporting a single task in a single modality such as pictures, images, text, voice, video, etc., to supporting a variety of tasks in multiple modalities. Task. That is to say, large models usually also have the ability to efficiently understand multi-modal information, cross-modal perception capabilities, and the ability to transfer and execute cross-differentiated tasks, etc., and may even have multi-modal information perception as embodied in the human brain. ability.

From another perspective, AI large model is the abbreviation of "artificial intelligence pre-training large model", which includes the two meanings of "pre-training" and "large model". The combination of the two creates a new artificial intelligence model. That is, after the model has completed pre-training on a large-scale data set, it does not require fine-tuning, or only requires fine-tuning with a small amount of data, and can support various downstream applications. In other words, large AI models benefit from its "large-scale pre-training + fine-tuning" paradigm and can adapt well to different downstream tasks, demonstrating its strong versatility. This large, versatile AI model can achieve superior performance by simply making corresponding fine-tuning in different downstream application scenarios when sharing parameters, breaking through the limitations of traditional AI models that are difficult to generalize to other tasks. .

From the perspective of processing results, the above-mentioned large AI model is also a generative model (GenerativeModel). This type of model can not only predict results based on features, but also "understand" how data is generated and "create" new data based on this.

With the support of the above-mentioned capabilities of the large AI model and the support of existing knowledge, some pre-training can be performed on the basic large AI model based on the scenario requirements in the embodiments of this application, so that the large AI model has multi-modality capabilities for commodities. understanding of dynamic information and content production capabilities. For example, in order to enable the AI large model to beautify and realistically process clothing matching images, some high-definition and beautiful clothing matching images can be collected in advance and input into the AI large model, which will produce the images according to the requirements. Specific content; for the content produced by the model, manual verification can also be carried out, and the accuracy information of the content is fed back to the AI large model, so that the AI large model can continue to iteratively learn, and ultimately make the produced content have higher quality. Accuracy, etc.

From the perspective of system architecture, referring to Figure 1, embodiments of the present application can provide corresponding functions in the product information service system. The product information service system can include a client and a server. The client mainly refers to consumer-oriented users. The client is mainly used to interact with users, including the user specifying the clothes to match, etc. It can also display images generated by AI, etc. The server side is mainly used to provide specific data support, including the storage of material libraries. In addition, the AI large model can also be run on the server side. After the AI large model generates specific images on the server side, it is returned to the client for display. etc.

The specific implementation solutions provided by the embodiments of this application are introduced in detail below.

Embodiment 1

First, Embodiment 1 provides a method of providing product matching information for the aforementioned "AI matching" function. See Figure 2. The method may include:

S201: Determine the first image to be matched with clothing; the first image includes: the first clothing and the first image content of the first partial human body part related to the human body wearing state.

In the "AI matching" scenario, mainly based on the clothing-related image specified by the user (which can be called the first image), other clothing product images that can be matched with it can be determined, and the target can be generated through the AI large model Image, this target image is used to simulate the matching effect of clothing with a matching relationship when worn on the same human body.

Among them, the first image to be matched with clothing can be specified by the user. For example, the user can be provided with an "AI matching" entrance on the client homepage and other pages, through which the user can upload the first image; or, It can be that the details page of a specific clothing product provides an "AI matching" entrance, through which an image related to the product can be used as the first image.

Alternatively, the "AI matching" function can be combined with functions such as product search. For example, product search results can be provided after the user initiates a search request for apparel products, and then the product search results can be displayed on the product search results list page. The product provides an "AI matching" portal, through which users can initiate an "AI matching" request for a certain product in the product search results. At this time, a certain target image associated with the product can be determined as the first image. Among them, a specific product search request can be initiated by entering keywords, or it can also be initiated by uploading a picture, and so on.

For example, as shown in Figure 3(A), it shows a product search result page. This page is generated after the user uploads an image and initiates a search. See point 31, which can be a specific search. As a result, a "one-click matching" option is provided, and the user can initiate a matching request based on the product, and accordingly, a certain target image related to the product can be determined as the first image.

As mentioned above, in this embodiment of the present application, in order to provide users with a more realistic and natural matching effect diagram, the specific first image may include image content about the clothing product itself, and may also include the clothing product. The image content of the first partial body part related to the product when it is worn on the human body. For example, the image shown in Figure 3(A) shows a shirt-like clothing product, and is a photo taken with a model wearing the shirt, which also includes the model's head, The hands, etc., can reflect the posture of the model wearing the shirt, as well as the state of the specific clothing when worn by the model in the corresponding posture, for example, some naturally occurring wrinkles, etc.

Among them, the above-mentioned first image can also be determined in different ways according to different request initiation methods. For example, if the AI matching is directly based on the image uploaded by the user, it can be determined whether the image uploaded by the user meets the conditions. For example, is it a photo taken with a person wearing specific clothing? If so, you can cut out the image, for example, cut out the parts of the human body and clothing, remove the background, and also cut out the image. Parts related to specific clothing, for example, if the specific clothing is a shirt, the lower body part can be removed, etc.

Or, in another way, since a material library needs to be established in advance, this material library can include multiple specific images. Such images can also be cut out from pre-collected upper body pictures of clothing and Obtained by interception, etc. In other words, the material library has already saved a lot of material libraries generated based on certain products. In this case, the user first initiates a search for clothing products based on uploaded images or keywords, etc., and then searches for them from the search results. When specifying a product for AI matching, the product image displayed on the search results list page can also be directly displayed as the material image corresponding to the specific product in the above material library. For example, in the example shown in Figure 3(A), information about multiple products that meet the search conditions is displayed on the search results list page, where each product corresponds to its own product image. In the embodiment of the present application, The material pictures corresponding to specific products in the material library can be displayed on the product search results page. In this way, when users initiate AI matching based on a product in the product search results, they can directly use the material pictures corresponding to the product as First image. The situation is similar for initiating AI matching requests from the product details page. The material image corresponding to the specific product can be determined from the material library as the first image, and so on.

Of course, during specific implementation, the material library may also include material pictures generated based on upper body pictures of clothing outside the current information service system (that is, the clothing in the specific image may not be a commodity on sale in the current information service system, but It can also be used to generate specific material pictures), and can even include some virtual material pictures generated by AI large models instead of images taken from real clothing product pictures, etc.

S202: Determine at least one second image that can be matched from the preset material library; the second image includes: the second clothing and the second image content of the second partial human body part related to the human body wearing state. .

After the first image is determined, the clothing included in the first image becomes the first clothing product that needs to be matched. After that, at least one second image that can be matched can be determined from the material library. Among them, there can be many different ways to determine the second image. For example, in one way, since the material library in the specific material library may be associated with corresponding products, a product matching relationship library can be established in advance. In this way, the second clothing product that can match the first clothing product can be determined, and then the second image corresponding to the second clothing product can be determined from the material library. Or, in another way, you can also make full use of the computer vision capabilities of algorithm models such as AI large models, and then directly determine the second image that can match the first image by understanding the model of the material images in the material library. etc. The second image may also include the image content of the second clothing product and the second image content of the second partial human body part related to the human body wearing state. For example, assuming that the second clothing product is a pants product, the second partial human body part may include legs, feet, and so on.

S203: Splice the first image content and the second image content.

After the first image and at least one second image are determined, the first image and the second image can be spliced. For example, if the first image is an image related to tops and clothing, and the second image is related to pants and clothing. image, the first image and the second image can be spliced up and down to form a complete body and the matching effect of specific clothing on the body.

During specific implementation, since the first partial human body part included in the first image content and the second partial human body part included in the second image content may come from different human bodies, even the same human body may be photographed in different postures. Images, therefore, if directly spliced, the poses of the upper body and lower body in the spliced image may not match. Therefore, in a preferred manner, the clothing key points and human body key point information contained in the first image content and the second image content can also be identified through a specific algorithm. After that, the clothing key points and human body key point information can be identified based on the identified clothing key points and human body key point information. point information to perform splicing processing on the first image content and the second image content. For example, based on the identification of key points of the human body, the specific first image content or the second image content is subjected to affine change processing and then spliced, so that the postures of different local body parts after splicing are more coordinated.

S204: Use the artificial intelligence AI large-scale parameter model to redraw the spliced image to generate a target image. The target image is used to simulate the wearing state of the first clothing and the second clothing on the same human body. The matching effect below.

After the splicing of the first image content and the second image content is completed, in order to make the final generated image more realistic and natural, the AI large model can also be used to redraw the spliced image. This redrawing process can mainly Beautify the spliced image and remove some defects existing in the spliced image. For example, since the first image content may be the upper body of character A wearing clothing 1, the second image content may be the lower body of character B wearing clothing 2, and there may be a height gap between character A and character B. In this case, after splicing, , the length of the upper and lower body may be inconsistent, etc. In addition, the material pictures in the material library themselves may also have some flaws in the process of cutting out the pictures. For example, the hands may not cut out naturally enough, etc. These problems can be improved by using large AI models. For example, the length ratio of the upper and lower bodies can be adjusted, flaws existing in the cutout process can be repaired, and image joints can be processed to make the final target image more natural. , is more realistic and better reflects the matching effect of the first clothing product and the second clothing product when they are worn on the same human body. In addition, you can also use AI large models to add backgrounds to specific matching plans, and so on.

For example, as shown in Figure 4, assuming that Figures 4(A) and (B) are the first image and the second image respectively, after splicing according to key points of clothing, key points of the human body, etc., the resulting image can be as shown in Figure 4( C), and after redrawing using the AI large model, the resulting image can be shown in Figure 4(D). By comparing Figure 4(C) and Figure 4(D), we can see that after redrawing the AI large model, the beauty and realism of the image have been improved. For example, the human body shown at 41 in Figure 4(C) The edges of the arm lines are not smooth enough. After redrawing the AI large model, as shown at 42 in Figure 4(D), the lines at the corresponding positions have become smoother, and the wrists and finger joints have become more natural; in addition, , the top connection shown at 43 in Figure 4(C) also has traces of cutout lines left during cutout. After processing with the AI large model, the corresponding position can be found in Figure 4(D) As shown at 44, the cutout lines have been removed, and the condition of the clothes appears more natural, etc.

After the above processing, the target image generated by the AI large model can be returned to the client for display. For example, after initiating AI matching for the first image in Figure 3(A), the displayed target image can be as shown in Figure 3(B) ), it can be seen that four matching schemes are provided for the AI large model, each showing the corresponding matching effects when the upper body is worn. Of course, more matching plans can also be generated, which can be viewed by clicking "Change Batch" and so on.

During specific implementation, in addition to the "one-click matching" option to generate specific matching solutions from the AI large model, the "advanced matching" option shown at 32 in Figure 3(B) can also be provided. Through this option, users You can specify the specific clothing type, style, color tone, etc. that need to be matched to shorten the matching range. For example, as shown in Figure 5(A), options such as "Select Type", "Select Style", and "Select Color" can be provided, and the user can make selections under these options respectively, for example, as shown in Figure 5(A) , select the type as "wide-leg pants"; as shown in Figure 5(B), select the style as "urban style"; as shown in Figure 5(C), select the color as "taro purple", and then, By clicking the "AI generated outfit" option, you can get the target image generated by the AI large model. In this way, if the user does not have clear requirements for the matching objects, he can use the "One-click matching" option to generate a specific matching plan completely by the AI large model. If the user has some clear requirements for the matching objects, he can use the "Advanced Matching" option. Filter, etc.

In addition, when the target image is displayed through the client, an operation option for searching for the same product for the second clothing product in the target image can be provided, so that after receiving the search request for the same product, the same product search request can be provided. product search results. For example, as shown in Figure 3(B), the "Search for the same style" option can be provided in the lower right corner of each target image. If the user is interested in a certain set of matching, he can click the corresponding "Search for the same style" option. , the search results for the same product as the corresponding matching object can be displayed, and the user can perform subsequent link operations such as purchasing the same product. In this way, users can quickly find other products that can match the products they are interested in and meet their needs. In this way, the target image generated by the AI large model can help users express the matching objects they need. Although the clothing products in the target image may not be directly available for purchase, they may not even be on sale in the current information service system. products, however, you can find products of the same style in the current information service system through "Search for the same style".

In an optional implementation, when the target image is displayed through the client, operation options for designing the clothing image in the target image can also be provided, that is, the "AI brush" function mentioned above . In other words, after the AI large model generates the target image, the user can also use the "AI brush" function to make a customized design based on the target image. After receiving a request for custom design of the clothing image in the target image, optional design elements may be provided to add the target design element to the clothing image. For example, you can add a design element at a certain position on a specific clothing image, or you can change the matching object for the clothing image, or modify the color of the clothing image through "graffiti", etc. Among them, because specific design elements usually exist in the form of stickers, etc., when added directly to the target image, it may not be three-dimensional enough and does not look like it is printed on specific clothing. Therefore, it can also provide users with Option for AI image generation of the design results. After receiving the AI generated image request, the AI large model can be used to redraw the designed target image to match the clothing image in the target image with the target image. Design elements are merged to generate images with realistic clothing effects.

For example, as shown in Figure 6, assume that the target image is as shown in Figure 6(A). During the design process, as shown in Figure 6(B), the user added some "flower" pattern texture pictures to the top. , after clicking the "Go to Generate" option, the AI large model can redraw the design results and generate the image as shown in Figure 6(C). Among them, during specific implementation, the AI large model can provide a variety of images with different redrawing degrees. For example, as shown in Figure 6(C), four images generated at different redrawing degrees are provided. Among them, the redrawing degree The lower it is, the closer it is to the original image. The higher the degree of redrawing, the more three-dimensional it is, and the closer it is to the actual state of the corresponding pattern on the specific clothing.

As another example, as shown in Figure 7, assume that the target image is as shown in Figure 7(A). During the user's design process, as shown in Figure 7(B), the user modified the color by "graffiti", but , it can be clearly seen that it is the color applied on, rather than the state when the clothing actually has the corresponding color. After clicking "Go to Generate", the image shown in Figure 7(C) can be generated. Similarly, the image can also be generated with a variety of different redrawing degrees. Among them, when the redrawing degree is relatively high, as shown in Figure 7(C) As shown in the "V4" picture in 7(C), it can be seen that the specific clothing is made up of multiple pieces, each piece has a different color, so it looks more real.

In addition, you can also provide an option to search for the same type of products on the above designed clothing images. After receiving a request to search for the same type of products on the designed clothing images, you can also provide search results for the same type of products. That is to say, for example, if the original target image does not have a "flower" pattern, if you directly search for the same style, the search results will mainly be styles without patterns. However, if the user needs a style with a "flower" pattern, Style, you can use this "AI brush" function to add the "flower" pattern, and use the AI large model to generate the clothing state after adding the design elements. If you are satisfied, you can initiate a search request for the same style, and you can get the "flower" pattern with "flower" "Similar products with pattern elements, etc.

In short, through the embodiments of the present application, a second image that can be matched with other first clothing can be determined for the first image specified by the user, and both the first image and the second image can include specific clothing. The image content and the image content of the relevant local human body parts in the human body wearing state. In this way, when the matching results are returned, the above image content can be spliced and redrawn through the AI large-scale parameter model to generate a model for simulating the matching of the first garment and the second garment when worn on the same human body. The target image for the effect. In this way, it can not only recommend matching objects for the user, but also provide the matching effect of the first clothing and the second clothing in the same matching scheme when worn on the same body. Moreover, this matching effect will be more real and natural, and can be more Effectively serve as a reference.

Embodiment 2

In the first embodiment mentioned above, the "AI Brush" function is mentioned. In practical applications, the "AI Brush" function can also exist independently from the "AI Matching" function. For example, the user can first enter the "AI Brush" function. "Interface, then upload any image, and then use the design elements provided in this function to customize the clothing image in the image. After that, it can also be redrawn through the AI large model to generate an image with more realistic clothing effects. . Specifically, Embodiment 2 provides an image processing method. See Figure 8. The method may include:

S801: After receiving a request to design the clothing image in the target image, provide candidate design elements for designing the clothing image, so as to add the target design element to the clothing image;

S802: In response to the request for image generation of the design result, use the AI large-scale parameter model to redraw the designed target image so as to integrate the clothing image in the target image with the target design element to generate Images with realistic clothing effects.

Specifically, an operation option for initiating a search for the same product on the designed image can also be provided, so that the user can search for the same product in the current system, and can also perform subsequent purchase and other processes.

Embodiment 3

In the first embodiment mentioned above, after the user specifies the first image of a certain clothing, the algorithm determines the second image corresponding to the clothing that can be matched, and then performs splicing and beautification of the AI large model to generate the target image and return it. to users. In the third embodiment, you can also initiate a specific outfit request by specifying the style/scene and character characteristics, and then use the AI large model to generate an outfit that matches the scene and is tried on the person with the character characteristics. Outfit renderings. Specifically, the third embodiment provides a method for providing clothing matching information. Referring to Figure 9, the method may include:

S901: Receive a clothing matching request, which includes the required target dressing style and/or scene information, as well as text and/or image information used to describe the characteristics of the target person.

In the third embodiment, a specific matching request can be initiated directly based on a certain style and/or scene. That is to say, the user does not need to specify a certain piece of clothing in advance, but the algorithm can provide a complete matching plan. However, Users can specify the human body characteristics of the character, and then use the AI large model to generate a specific matching plan when trying on the human body with the characteristics. Among them, users can describe human body characteristics through text or uploaded images, or they can also use text + image to describe. For example, after entering the interface related to scene matching, the user can select the required scene, etc. Suppose the user selects the "wedding style" scene, the interface can also provide input controls for inputting character characteristics, for example, It can be a text input box, a control for uploading images, etc. For example, as shown in Figure 10(A), the user can enter the text "Red-haired girl wearing a short top and glasses", or, as shown in Figure 10(B), the user can also upload images such as photos to describe themselves. Desired human characteristics are described. Or, as shown in Figure 10(C), it can also be described in the form of text + image, and so on.

S902: Determine a clothing matching scheme that meets the requirements of the target dressing style and/or scene, and use an AI large-scale parameter model to generate a target image. The target image is used to express the clothing matching scheme when worn by the target person. The matching effect under the state.

After receiving the matching request, you can determine the clothing matching scheme that meets the needs of the target dressing style and/or scene, and use the AI large model to generate a target image. The target image can express the clothing matching scheme based on the target. The matching effect of the character's clothing. During specific implementation, multiple styles and/or scenes can be set in advance, and high-definition clothing matching images (with specific character upper body effects) in various styles and/or scenes can be used to train the AI large model, so that the AI large model Has the ability to generate corresponding images. For example, after selecting the "wedding style" scene and uploading a character image as shown in Figure 10(B), the target image generated by the AI large model can be as shown in Figure 10(D). It can be seen that the characters in the target image The body shape, posture, etc. are consistent with the images uploaded by the user, thereby expressing the upper body effect of the clothing matching scheme on the target person in the specific scene.

Of course, for this kind of target image generated based on the scene and the human body characteristics of the target person, the operation option for initiating a search for the same model can also be provided. The user can obtain information about the same product by initiating a search for the same model, and then execute Subsequent purchase and other processes.

It should be noted that the above-mentioned "AI matching", "AI brush", AI matching based on style and/or scene, etc., can exist independently. For example, they can be multiple independent sub-modules in the same functional module. Users You can enter different modules to interact separately, or you can also use them in combination with each other. For example, the aforementioned "AI matching" can be combined with the "AI brush" function, and AI matching based on style and/or scene can also be combined with the "AI matching" function. Brush” functionality combined, and more.

In practical applications, it can also provide users with community platform services. After obtaining the target image generated by the AI large model through the above-mentioned multiple interaction methods, it can also be published through this community platform. Correspondingly, you can also view other users' through The AI-generated works published by the community platform, for example, as shown in Figure 11(A), can display the AI-generated works published by multiple users on the community platform. If the current user is interested in a certain work, he can also search for the same product. Alternatively, for AI works generated under the "AI collocation" method, as shown in Figure 11(B), the "Generate my collocation" option can also be provided, so that the current user can re-enter his or her needs and regenerate a new AI works; or, if it is an AI work generated based on scene and human body feature description information, as shown in Figure 11(C), the "Use current scene" option can also be provided, allowing users to generate the AI matching solution they need based on the scene. ,etc.

It should be noted that the embodiments of this application may involve the use of user data. In actual applications, this can be done in compliance with the applicable laws and regulations of the country where the user is located (for example, the user explicitly agrees, the user is effectively notified, etc.), use user-specific personal data in the scenarios described herein to the extent permitted by applicable laws and regulations.

Corresponding to Embodiment 1, this embodiment of the present application also provides a device for providing product matching information. The device may include:

The first image determination unit is used to determine the first image to be matched with clothing; the first image includes: the first image content of the first clothing and its related first partial human body part in the human body wearing state;

The second image determination unit is used to determine at least one second image that can be matched from the preset material library; the second image includes: the second clothing and its related second partial human body when the human body is wearing it. The second image content of the part;

A splicing processing unit configured to splice the first image content and the second image content;

The AI redrawing unit is used to use the artificial intelligence AI large-scale parameter model to redraw the spliced image to generate a target image. The target image is used to simulate the first clothing and the second clothing in the same The matching effect is carried out on the human body when worn.

Wherein, the images stored in the material library correspond to the commodities provided in the associated information service system;

The first image determination unit may be specifically used for:

Provide product search results based on product search requests submitted by users;

In response to a clothing matching request initiated for a designated product in the product search results, a target image associated with the designated product in the material library is determined as the first image.

Wherein, the product search request may be: a product search request submitted based on clothing-related images uploaded by the user, and the product search results include products of the same style as the clothing body included in the image uploaded by the user.

Specifically, the splicing processing unit can be used for:

Identify clothing key points and human body key point information contained in the first image content and the second image content;

According to the identified clothing key points and human body key point information, the first image content and the second image content are spliced.

Additionally, the device may include:

The first same-style search unit is configured to also provide an operation option for searching for same-style products on the second clothing in the target image when displaying the target image through the client, so that when receiving the same-style product search After request, search results for the same product will be provided.

In addition, when the target image is displayed through the client, operation options for designing the clothing image in the target image are also provided;

At this time, the device may also include:

A design element providing unit, configured to provide optional design elements after receiving a request for customized design of the clothing image in the target image, so as to add the target design element to the clothing image;

The AI redrawing unit may also be used to: in response to a request for image generation of the design result, use the AI large-scale parameter model to redraw the designed target image, so as to compare the clothing image in the target image with the The target design elements are combined to generate an image with a realistic clothing effect.

Additionally, the device may include:

The second same-style search unit is configured to provide search results for same-style products in response to a request for searching for same-style products on the designed clothing image.

Corresponding to Embodiment 2, this embodiment of the present application also provides an image processing device, which may include:

A design element providing unit, configured to provide candidate design elements for designing the clothing image after receiving a request to design the clothing image in the target image, so as to add the target design element to the clothing image;

The AI redrawing unit is configured to respond to a request for image generation of the design result and use the AI large-scale parameter model to redraw the designed target image, so as to match the clothing image in the target image with the target design. Elements are combined to generate images with realistic clothing effects.

Specifically, the device may also include:

A same-style search unit is configured to provide search results for same-style products in response to a request for searching for same-style products on the designed clothing image.

Corresponding to Embodiment 3, this embodiment of the present application also provides a device for providing clothing matching information. The device may include:

A request receiving unit, configured to receive a clothing matching request, which includes the required target dressing style and/or scene information, as well as text and/or image information used to describe the human body characteristics of the target person;

An AI generation unit is used to determine a clothing matching scheme that meets the needs of the target dressing style and/or scene, and uses an AI large-scale parameter model to generate a target image, where the target image is used to express the clothing matching scheme based on the requirements of the target clothing style and/or scene. Describe the matching effect when the target person is wearing it.

Additionally, the device may include:

A same-style search unit is configured to provide search results for same-style products in response to a request to search for the same-style products for the clothing in the clothing matching scheme.

In addition, embodiments of the present application also provide a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps of the method described in any one of the foregoing method embodiments are implemented.

and an electronic device including:

one or more processors; and

A memory associated with the one or more processors. The memory is used to store program instructions. When the program instructions are read and executed by the one or more processors, the program instructions execute any one of the foregoing method embodiments. steps of the method.

Among them, Figure 12 exemplarily shows the architecture of an electronic device. For example, the device 1200 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. , aircraft, etc.

Referring to Figure 12, device 1200 may include one or more of the following components: processing component 1202, memory 1204, power supply component 1206, multimedia component 1208, audio component 1210, input/output (I/O) interface 1212, sensor component 1214, and communications component 1216.

Processing component 1202 generally controls the overall operations of device 1200, such as operations associated with display, phone calls, data communications, camera operations, and recording operations. The processing component 1202 may include one or more processors 1220 to execute instructions to complete all or part of the steps of the method provided by the technical solution of the present disclosure. Additionally, processing component 1202 may include one or more modules that facilitate interaction between processing component 1202 and other components. For example, processing component 1202 may include a multimedia module to facilitate interaction between multimedia component 1208 and processing component 1202.

Memory 1204 is configured to store various types of data to support operations at device 1200 . Examples of such data include instructions for any application or method operating on device 1200, contact data, phonebook data, messages, pictures, videos, etc. Memory 1204 may be implemented by any type of volatile or non-volatile storage device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EEPROM), Programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.

Power supply component 1206 provides power to various components of device 1200. Power supply components 1206 may include a power management system, one or more power supplies, and other components associated with generating, managing, and distributing power to device 1200 .

Multimedia component 1208 includes a screen that provides an output interface between device 1200 and the user. In some embodiments, the screen may include a liquid crystal display (LCD) and a touch panel (TP). If the screen includes a touch panel, the screen may be implemented as a touch screen to receive input signals from the user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. A touch sensor can not only sense the boundaries of a touch or swipe action, but also detect the duration and pressure associated with the touch or swipe action. In some embodiments, multimedia component 1208 includes a front-facing camera and/or a rear-facing camera. When the device 1200 is in an operating mode, such as a shooting mode or a video mode, the front camera and/or the rear camera may receive external multimedia data. Each front-facing camera and rear-facing camera can be a fixed optical lens system or have a focal length and optical zoom capabilities.

Audio component 1210 is configured to output and/or input audio signals. For example, audio component 1210 includes a microphone (MIC) configured to receive external audio signals when device 1200 is in operating modes, such as call mode, recording mode, and speech recognition mode. The received audio signals may be further stored in memory 1204 or sent via communications component 1216 . In some embodiments, audio component 1210 also includes a speaker for outputting audio signals.

The I/O interface 1212 provides an interface between the processing component 1202 and a peripheral interface module. The peripheral interface module may be a keyboard, a click wheel, a button, etc. These buttons may include, but are not limited to: Home button, Volume buttons, Start button, and Lock button.

Sensor component 1214 includes one or more sensors that provide various aspects of status assessment for device 1200 . For example, the sensor component 1214 can detect the open/closed state of the device 1200, the relative positioning of components, such as the display and keypad of the device 1200, and the sensor component 1214 can also detect the position change of the device 1200 or a component of the device 1200. , the presence or absence of user contact with device 1200 , device 1200 orientation or acceleration/deceleration and temperature changes of device 1200 . Sensor assembly 1214 may include a proximity sensor configured to detect the presence of nearby objects without any physical contact. Sensor assembly 1214 may also include a light sensor, such as a CMOS or CCD image sensor, for use in imaging applications. In some embodiments, the sensor component 1214 may also include an acceleration sensor, a gyroscope sensor, a magnetic sensor, a pressure sensor, or a temperature sensor.

Communication component 1216 is configured to facilitate wired or wireless communications between device 1200 and other devices. Device 1200 can access wireless networks based on communication standards, such as WiFi, or mobile communication networks such as 2G, 3G, 4G/LTE, and 5G. In one exemplary embodiment, the communication component 1216 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel. In one exemplary embodiment, the communications component 1216 also includes a near field communications (NFC) module to facilitate short-range communications. For example, the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.

In an exemplary embodiment, device 1200 may be configured 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 array (FPGA), controller, microcontroller, microprocessor or other electronic components are implemented for executing the above method.

In an exemplary embodiment, a non-transitory computer-readable storage medium including instructions, such as a memory 1204 including instructions, is also provided. The above instructions can be executed by the processor 1220 of the device 1200 to complete the method provided by the technical solution of the present disclosure. . For example, the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.

From the above description of the embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software plus the necessary general hardware platform. Based on this understanding, the technical solution of the present application can be embodied in the form of a software product in essence or that contributes to the existing technology. The computer software product can be stored in a storage medium, such as ROM/RAM, disk , optical disk, etc., including a number of instructions to cause a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various embodiments or certain parts of the embodiments of this application.

Each embodiment in this specification is described in a progressive manner. The same and similar parts between the various embodiments can be referred to each other. Each embodiment focuses on its differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple. For relevant details, please refer to the partial description of the method embodiment. The system and system embodiments described above are only illustrative, in which the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, It can be located in one place, or it can be distributed over multiple network elements. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Persons of ordinary skill in the art can understand and implement the method without any creative effort.

The method and electronic equipment for providing product matching information provided by this application have been introduced in detail above. Specific examples are used in this article to illustrate the principles and implementation methods of this application. The description of the above embodiments is only used to help understand this application. The method of the application and its core idea; at the same time, for those of ordinary skill in the field, there will be changes in the specific implementation and application scope based on the ideas of the application. In summary, the contents of this specification should not be construed as limiting this application.

Claims (13)

1. A method for providing merchandise collocation information, comprising:

determining a first image to be subjected to clothing collocation; the first image includes: a first image content of a first localized human body portion associated with the first garment in a worn state of the human body;

determining at least one collocation second image from a preset material library; the second image includes: a second image content of a second partial human body part associated with the second garment in the worn state of the human body;

splicing the first image content and the second image content;

and redrawing the spliced image by using the artificial intelligence AI large-scale parameter model to generate a target image, wherein the target image is used for simulating the matching effect of the first garment and the second garment in the wearing state of the same human body.

2. The method of claim 1, wherein the step of determining the position of the substrate comprises,

the images stored in the material library correspond to commodities provided in the associated information service system;

the determining a first image to be subjected to clothing collocation includes:

providing commodity searching results according to commodity searching requests submitted by users;

and responding to a clothing collocation request initiated for a specified commodity in the commodity search results, and determining a target image associated with the specified commodity in the material library as the first image.

3. The method of claim 2, wherein the step of determining the position of the substrate comprises,

the commodity search request is: and submitting a commodity search request based on the image related to the clothes uploaded by the user, wherein the commodity search result comprises commodities belonging to the same type as the clothes main body included in the image uploaded by the user.

4. The method of claim 1, wherein the step of determining the position of the substrate comprises,

the splicing processing of the first image content and the second image content comprises the following steps:

identifying clothing key points and human body key point information contained in the first image content and the second image content;

and according to the identified clothing key points and the human body key point information, performing splicing processing on the first image content and the second image content.

5. The method of claim 1, wherein the step of determining the position of the substrate comprises,

and when the target image is displayed through the client, an operation option for searching the same type of commodity for the second clothes in the target image is also provided, so that the same type of commodity searching result is provided after the same type of commodity searching request is received.

6. The method of claim 1, wherein the step of determining the position of the substrate comprises,

when the target image is displayed through the client, operation options for designing the clothing image in the target image are also provided;

the method further comprises the steps of:

providing optional design elements after receiving a request for custom-designing a clothing image in the target image so as to add the target design elements to the clothing image;

and responding to a request for generating an image of a design result, and redrawing the designed target image by utilizing an AI large-scale parameter model so as to fuse the clothing image in the target image with the target design element and generate an image with a real clothing effect.

7. The method as recited in claim 6, further comprising:

and responding to the request of searching the same commodity for the designed clothing image, and providing the same commodity searching result.

8. An image processing method, comprising:

after receiving a request for designing a clothing image in a target image, providing candidate design elements for designing the clothing image so as to add the target design elements to the clothing image;

and responding to a request for generating an image of a design result, and redrawing the designed target image by utilizing an AI large-scale parameter model so as to fuse the clothing image in the target image with the target design element and generate an image with a real clothing effect.

9. The method as recited in claim 8, further comprising:

and responding to the request of searching the same commodity for the designed clothing image, and providing the same commodity searching result.

10. A method of providing apparel collocation information, comprising:

receiving a clothing collocation request, wherein the request comprises required target dressing style and/or scene information and text and/or image information for describing human body characteristics of a target person;

and determining a clothing matching scheme meeting the requirements of the target dressing style and/or scene, and generating a target image by utilizing an AI large-scale parameter model, wherein the target image is used for expressing the matching effect of the clothing matching scheme in the wearing state of the target person.

11. The method as recited in claim 10, further comprising:

and responding to a request for searching the same commodity for the clothes in the clothes collocation scheme, and providing the same commodity searching result.

12. A computer readable storage medium, on which a computer program is stored, characterized in that the program, when being executed by a processor, implements the steps of the method of any of claims 1 to 11.

13. An electronic device, comprising:

one or more processors; and

a memory associated with the one or more processors for storing program instructions that, when read for execution by the one or more processors, perform the steps of the method of any of claims 1 to 11.