Image processing method and storage medium

The method of creating surface areas through image mapping and distorting control points solves the problems of low sharpness and loss of details in GAN model generation, and improves the image realism and detail retention effect in image processing.

CN116205787BActive Publication Date: 2024-12-03ALIBABA (CHINA) CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202310115561.5
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2023-02-10
Publication Date
2024-12-03
Estimated Expiration
2043-02-10

AI Technical Summary

Technical Problem

Existing image generation technology based on GAN model results in low image clarity, loss of details and pseudo-texture, which cannot effectively solve the problems of image details retention and realism in image processing.

Method used

By receiving image processing requests submitted by users, identify additional images and initial images, map and distort control points to create surface areas, use surface areas to update intermediate images, generate target images, avoid low resolution and loss of details caused by model calculation, and increase The three-dimensional effect and realism of the image.

Benefits of technology

The high-definition and detail retention of the target image in image processing are achieved, increasing the realistic image, and avoiding the problems of low resolution and detail loss in traditional methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN116205787B_ABST
    Figure CN116205787B_ABST
Patent Text Reader

Abstract

The embodiments of the present specification provide an image processing method and a storage medium, the image processing method comprising: receiving an image processing request submitted by a user, and determining an additional image, additional image information and an initial image according to the image processing request; mapping the additional image to the initial image according to the additional image information to obtain an intermediate image; determining a distortion control point corresponding to an image area in the intermediate image where the additional image is located, and creating a curved surface area corresponding to the image area based on the distortion control point; updating the intermediate image using the curved surface area, and obtaining a target image that integrates the additional image according to the update result, thereby reducing the rendering cost of the target image and increasing the realism of the target image.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The embodiments of the present specification relate to the field of computer technology, and in particular, to image processing methods. Background Art

[0002] With the development of the Internet, both enterprises and individuals inevitably have the need for image processing. Image processing technology is a technology that uses computers to process image information. In the image processing scenario where there are image materials, new images can be generated by processing these image materials. At present, the Generative Adversarial Nets (GAN) model obtained by training a large number of sample images is usually used for image generation. However, the images generated by the GAN model have low clarity, the details of the image materials are not fully retained, and pseudo textures are prone to appear. Therefore, it is urgent to provide a solution to the above problems. Summary of the invention

[0003] In view of this, the embodiments of this specification provide an image processing method. One or more embodiments of this specification also relate to an image processing apparatus, a computing device, a computer-readable storage medium and a computer program to solve the technical defects existing in the prior art.

[0004] According to a first aspect of an embodiment of this specification, there is provided an image processing method, including:

[0005] receiving an image processing request submitted by a user, and determining an additional image, additional image information, and an initial image according to the image processing request;

[0006] Mapping the additional image to the initial image according to the additional image information to obtain an intermediate image;

[0007] Determining a distortion control point corresponding to an image region in the intermediate image where the additional image is located, and creating a curved surface region corresponding to the image region based on the distortion control point;

[0008] The intermediate image is updated using the curved surface area, and a target image fused with the additional image is obtained according to the update result.

[0009] According to a second aspect of an embodiment of this specification, another image processing method is provided, including:

[0010] Receiving a clothing try-on request submitted by a user, and determining an additional image, additional image information, an initial clothing image, and a model image according to the clothing try-on request;

[0011] Mapping the additional image to the initial clothing image according to the additional image information to obtain an intermediate clothing image;

[0012] Determining a distortion control point corresponding to an image region in the intermediate clothing image where the additional image is located, and creating a curved surface region corresponding to the image region based on the distortion control point;

[0013] The intermediate clothing image is updated by using the curved surface area, and the updated intermediate clothing image is fused with the model image to obtain a clothing try-on image.

[0014] According to a third aspect of the embodiments of this specification, there is provided an image processing device, including:

[0015] A first receiving module is configured to receive an image processing request submitted by a user, and determine an additional image, additional image information, and an initial image according to the image processing request;

[0016] A first mapping module is configured to map the additional image to the initial image according to the additional image information to obtain an intermediate image;

[0017] A first creation module is configured to determine a distortion control point corresponding to an image region in the intermediate image where the additional image is located, and create a curved surface region corresponding to the image region based on the distortion control point;

[0018] The first updating module is configured to update the intermediate image using the curved surface area, and obtain a target image fused with the additional image according to the updating result.

[0019] According to a fourth aspect of the embodiments of this specification, another image processing device is provided, including:

[0020] A second receiving module is configured to receive a clothing try-on request submitted by a user, and determine an additional image, additional image information, an initial clothing image, and a model image according to the clothing try-on request;

[0021] A second mapping module is configured to map the additional image to the initial clothing image according to the additional image information to obtain an intermediate clothing image;

[0022] A second creation module is configured to determine a distortion control point corresponding to an image region in the intermediate clothing image where the additional image is located, and create a curved surface region corresponding to the image region based on the distortion control point;

[0023] The second updating module is configured to update the intermediate clothing image using the curved surface area, and fuse the updated intermediate clothing image with the model image to obtain a clothing fitting image.

[0024] According to a fifth aspect of an embodiment of this specification, a computing device is provided, including:

[0025] Memory and processor;

[0026] The memory is used to store computer executable instructions, and the processor is used to execute the computer executable instructions. When the computer executable instructions are executed by the processor, the steps of the above-mentioned image processing method are implemented.

[0027] According to a sixth aspect of the embodiments of this specification, a computer-readable storage medium is provided, which stores computer-executable instructions, and when the instructions are executed by a processor, the steps of the above-mentioned image processing method are implemented.

[0028] According to a seventh aspect of the embodiments of this specification, a computer program is provided, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned image processing method.

[0029] The image processing method provided in the embodiment of this specification, in order to ensure the generation effect of the target image, receives the image processing request submitted by the user, and determines the additional image, the additional image information and the initial image according to the image processing request, and maps the additional image to the initial image according to the additional image information, obtains the intermediate image, and realizes the preliminary fusion of the additional image and the initial image. Then, the distortion control point corresponding to the image area where the additional image is located in the intermediate image is determined, and the surface area corresponding to the image area is created based on the distortion control point, so as to realize the creation of the image area of ​​the additional image to be curved, and then the intermediate image is updated by using the surface area, and the target image fused with the additional image is obtained according to the update result, so as to avoid the low resolution and image detail loss caused by processing the image by the model calculation method, and make the generated target image have a three-dimensional effect, and increase the realism of the target image. BRIEF DESCRIPTION OF THE DRAWINGS

[0030] Figure 1 is a schematic diagram of an image processing method provided by an embodiment of this specification;

[0031] Figure 2 is a flow chart of an image processing method provided by an embodiment of this specification;

[0032] Figure 3 is a schematic diagram of control points and distortion control points of an image region in an image processing method provided by an embodiment of the present specification;

[0033] Figure 4 is a schematic diagram of a target image area and a template image in an image processing method provided by an embodiment of this specification;

[0034] Figure 5is a schematic diagram of image fusion in an image processing method provided by an embodiment of this specification;

[0035] Figure 6 It is a schematic diagram of the overall flow of an image processing method provided by an embodiment of this specification;

[0036] Figure 7 is a flow chart of another image processing method provided by an embodiment of this specification;

[0037] Figure 8 is a processing flow chart of an image processing method provided by an embodiment of this specification;

[0038] Fig. 9 is a structural schematic diagram of an image processing device provided by an embodiment of this specification;

[0039] Fig.10 is a schematic diagram of the structure of another image processing device provided by an embodiment of this specification;

[0040] Fig.11 It is a structural block diagram of a computing device provided by an embodiment of this specification. DETAILED DESCRIPTION

[0041] Many specific details are described in the following description to facilitate a full understanding of this specification. However, this specification can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of this specification, so this specification is not limited to the specific implementation disclosed below.

[0042] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms of "a", "said" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0043] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0044] First, the terms involved in one or more embodiments of this specification are explained.

[0045] Clipping mask: The principle is to use the layer below to limit the display and scope of the upper layer.

[0046] See also Figure 1 The schematic diagram shown, in order to ensure the generation effect of the target image, receives the image processing request submitted by the user through the client, and determines the additional image, the additional image information and the initial image according to the image processing request, and maps the additional image to the initial image according to the additional image information, obtains the intermediate image, and realizes the preliminary fusion of the additional image and the initial image. Then, the distortion control point corresponding to the image area where the additional image is located in the intermediate image is determined, and the surface area corresponding to the image area is created based on the distortion control point, so as to realize the creation of the image area of ​​the additional image to be curved, and then the intermediate image is updated by using the surface area, and the target image fused with the additional image is obtained according to the update result, avoiding the low resolution and image detail loss caused by processing the image by using the model calculation method, and making the generated target image have a three-dimensional effect, which increases the realism of the target image.

[0047] In this specification, an image processing method is provided. This specification also relates to an image processing apparatus, a computing device, a computer-readable storage medium, and a computer program, which are described in detail one by one in the following embodiments.

[0048] See also Figure 2 , Figure 2 A flowchart of an image processing method provided according to an embodiment of the present specification is shown, which specifically includes the following steps.

[0049] Step 202: receiving an image processing request submitted by a user, and determining an additional image, additional image information, and an initial image according to the image processing request.

[0050] Specifically, a user refers to an individual who has image processing needs, such as a merchant, a consumer, or a designer. Correspondingly, an image processing request refers to a request for applying for image processing. In practical applications, in order to facilitate users to request image processing, a corresponding request page can be provided to users. Users can input image processing requirements based on operation controls such as upload controls and selection controls in the request page, and after completing the input, an image processing request is generated and submitted.

[0051] Additional images refer to images that are added to the initial image. The additional images can be images of any content, such as landscape images, human images, animal images, logo images, cartoon images, etc., and are not limited here. Correspondingly, additional image information refers to information associated with the additional images. The additional image information can be information such as the image size, image fusion area, and image fusion effect of the additional images. The initial image refers to the image to which the additional image is to be added. The initial image can be a clothing image, a car image, a human image, a curtain image, a model image, etc. In actual applications, users may need to fuse additional images with the initial image for the purpose of creating a new image, reflecting the difference of the initial image, or increasing the richness of the initial image.

[0052] Taking merchants as an example, for the e-commerce industry, especially in the field of clothing, the appearance of clothing images will greatly affect customers' desire to buy. However, every time a new style of clothing is released, the shooting cost not only includes the cost of models, photographers, and venues, but also the time cost of post-production. For many small and medium-sized merchants, this part of the cost is difficult to bear, so it will lead to the appearance of some stolen or infringing pictures. In order to fundamentally solve the problem of merchants stealing pictures, the image processing method provided in the embodiments of this specification can be used to generate clothing images.

[0053] It should be noted that the image method provided in this specification can support the processing of any initial image. For the convenience of description, this embodiment uses the initial image as an initial clothing image for illustration. The processing process of other types of initial images can refer to the same or corresponding description content of this embodiment, and this embodiment will not be elaborated in detail here.

[0054] Considering that the image processing process is usually complicated and requires more computing resources, image processing is usually performed on the server to ensure the efficiency of image processing. In order to facilitate users to interact with the server, an interactive page (i.e., the request page) is usually provided to users on the client, so that users can submit image processing requests through the client, and the server receives the image processing requests submitted by users through the client.

[0055] In specific implementation, after receiving the image processing request submitted by the user, in order to clarify the user's image processing requirements, the image processing request needs to be parsed. In this embodiment, the additional image, additional image information and initial image are determined according to the image processing request, which may specifically include the following two methods:

[0056] Method 1:

[0057] The image processing request is parsed to obtain initial image identification information, additional image and additional image information; and according to the initial image identification information, an initial image corresponding to the initial image identification information is searched in an image library.

[0058] Specifically, the initial image identification information refers to information that can uniquely identify the initial image. The initial image identification information can be image code, image name, image type and other information, which are not limited here. In the case where the initial image is an initial clothing image, the initial image identification information can be clothing style information, such as: T-shirt, dress, shirt and other clothing styles. The image library refers to the storage space used to store the initial image, such as a database, folder, image server, etc.

[0059] The received image processing request is parsed, specifically, based on the correspondence between the pre-set request parameters and the information, the specific information corresponding to the request parameters carried in the image processing request is determined. For example, the image processing request carries three request parameters, which are: IMAGE_ID, ADD_IMAGE, and ADD_IMAGE_INFO. Among them, IMAGE_ID corresponds to the initial image identification information, ADD_IMAGE corresponds to the additional image, and ADD_IMAGE_INFO corresponds to the additional image information. The information corresponding to IMAGE_ID in the image processing request is used as the initial image identification information, the image corresponding to ADD_IMAGE is used as the additional image; and the information corresponding to ADD_IMAGE_INFO is used as the additional image information.

[0060] In this embodiment, the user can manually upload an additional image in the image processing request page, input the fusion information (additional image information) associated with the additional image, and input the initial image identification information corresponding to the initial image stored in the server, and submit the image processing request on this basis. Then the submitted image processing request directly carries the initial image identification information, the additional image, and the additional image information. The server then searches for the corresponding initial image in the image library according to the initial image identification information.

[0061] For example, a clothing processing request submitted by a merchant through a client is received, the clothing processing request is parsed, and the clothing type is a T-shirt, an additional image IMAGE1, and the positioning information and size information of the additional image are obtained. According to the clothing type T-shirt, an initial clothing image of the T-shirt type is searched in the clothing library.

[0062] In summary, by parsing the image processing request, obtaining the initial image identification information, additional images and additional image information, it is possible to obtain the additional images uploaded by the user, increasing the diversity and richness of the additional images. Then, by searching for the initial image corresponding to the initial image identification information, the initial image is obtained, avoiding the user providing the initial image himself, providing the user with image materials to be processed, and facilitating the user's processing flow.

[0063] Method 2:

[0064] The image processing request is parsed to obtain initial image identification information, additional image identification information and additional image information; based on the initial image identification information and the additional image identification information, the initial image corresponding to the initial image identification information and the additional image corresponding to the additional image identification information are searched in the image library.

[0065] Specifically, the additional image identification information refers to information that can uniquely identify the additional image. The additional image information may also be image code, image name, image type and other information, which is not limited here.

[0066] In this embodiment, the user can enter the initial image identification information corresponding to the initial image stored in the server, the additional image identification information corresponding to the additional image stored in the server, and the fusion information (additional image information) associated with the additional image in the image processing request page, and submit the image processing request on this basis. The submitted image processing request directly carries the initial image identification information, the additional image identification information, and the additional image information. The server then searches the image library for the corresponding initial image based on the initial image identification information, and searches the image library for the corresponding additional image based on the additional image identification information. It should be noted that the additional image and the initial image can also be stored in the same image library, or in different image libraries, which is not limited here.

[0067] It should be noted that the initial image and the additional image may be pre-stored by the server and provided to the user, or may be pre-uploaded to the server by the user before the user submits an image processing request, and there is no limitation here.

[0068] For example, a clothing processing request submitted by a merchant through a client is received, and the clothing processing request is parsed to obtain that the clothing type is a T-shirt, the image name of the additional image is IMAGE1, and the positioning information and size information of the additional image are obtained; according to the clothing type T-shirt, the initial clothing image of the T-shirt type is searched in the clothing library, and according to the image name IMAGE1, the additional image IMAGE1 corresponding to IMAGE1 is searched in the material library.

[0069] In summary, by parsing the image processing request, the initial image identification information, additional image identification information and additional image information are obtained, and then the initial image corresponding to the initial image identification information and the additional image corresponding to the additional image identification information are searched, thereby avoiding directly carrying image files in the image processing request and improving the submission efficiency of the image processing request. In addition, the user is provided with the search and processing of the initial image to be processed and the additional image, which facilitates the user's processing flow.

[0070] Considering that if the image processing program needs to process various types of images, the image processing program may be too complicated. Therefore, in order to simplify the image processing program, the image to be processed can be calibrated in advance so as to perform image processing on the standardized image. In this embodiment, the additional image is determined according to the image processing request, and the specific implementation is as follows:

[0071] Obtaining an initial additional image carried in the image processing request;

[0072] The initial additional image is subjected to image format conversion and / or pixel channel conversion to obtain an additional image.

[0073] Specifically, the initial additional image refers to an additional image carried in the image processing request that has not been calibrated, that is, a non-standardized additional image. That is to say, there is a difference in image format and / or pixel channel between the initial additional image and the additional image required during image processing. Therefore, it is necessary to convert the image format and / or pixel channel of the initial additional image.

[0074] Correspondingly, image format refers to the format in which image files are stored, such as JPEG, BMP, GIF, EPS, etc. Pixel channel refers to the color representation of pixels in an image. For example, an RGB pixel has three color representations, and the pixel has three pixel channels, which are R (red), G (green) and B (blue).

[0075] In a specific implementation, after obtaining the initial additional image carried in the image processing request, it is determined whether the image format of the initial additional image is a preset image format; if not, it indicates that the image format of the initial additional image does not meet the format standard of image processing, and the image format of the initial additional image is converted into the preset image format to obtain a standardized additional image; if yes, it indicates that the image format of the initial additional image meets the format standard of image processing, and the initial additional image is directly used as a standardized additional image.

[0076] Similarly, it is also possible to determine whether the pixel channels of the initial additional image are preset pixel channels; if not, it indicates that the image channels of the initial additional image do not meet the format standard for image processing, and the pixel channels of the initial additional image are converted into preset pixel channels to obtain a standardized additional image; if yes, it indicates that the pixel channels of the initial additional image meet the channel standard for image processing, and the initial additional image is directly used as a standardized additional image.

[0077] For example, an initial additional image IMAGE_I carried in an image processing request submitted by a merchant is obtained, and when the image format of the initial additional image IMAGE_I is not a vector image format, the image format of the initial additional image IMAGE_I is converted into an initial additional image IMAGE_I in a vector image format, and when the image channel of the initial additional image IMAGE_I is not an RGBA channel, the image channel of the initial additional image IMAGE_I is converted into an additional image IMAGE1 with an RGBA channel.

[0078] In summary, by performing image format conversion and / or image channel conversion on the initial additional image to obtain the additional image, it is ensured that the image format and / or image channel of the additional image meets the image processing standard and the smoothness and stability of the image processing process are ensured.

[0079] Step 204: Map the additional image to the initial image according to the additional image information to obtain an intermediate image.

[0080] Specifically, based on the above determination of the additional image, the additional image information and the initial image, since different additional images may have differences in mapping positions and / or image sizes, in order to ensure the accuracy of the fusion of the additional image and the initial image, it is necessary to first map the additional image to the initial image to obtain an intermediate image, and then perform further fusion processing based on the intermediate image. The intermediate image refers to an image generated after mapping the additional image to the initial image.

[0081] In a specific implementation, the additional image information may include information related to mapping the additional image to the initial image, such as the additional image information including information related to translation, rotation and / or scaling of the additional image, a mapping matrix is ​​constructed based on this information, and then the additional image is mapped according to the mapping matrix to achieve mapping the additional image to the initial image. In addition, the additional image information may also directly include the mapping matrix, which is not limited here.

[0082] Furthermore, considering that the user usually has mapping requirements such as position and size for the additional image to be mapped, in order to facilitate the user to submit information corresponding to the image processing request and ensure the accuracy of the additional image mapping, in this embodiment, the additional image is mapped to the initial image according to the additional image information to obtain the intermediate image, and the specific implementation includes the following steps:

[0083] Step 2042: Determine the mapping center point corresponding to the additional image in the initial image according to the positioning information in the additional image information.

[0084] Specifically, the positioning information refers to the mapping position information corresponding to the additional image in the initial image. The positioning information can be the vertex coordinates corresponding to the four vertices of the additional image in the initial image, or can be the reference position corresponding to the additional image in the initial image, such as in the middle of the initial image, or in the upper right corner of the initial image. Correspondingly, the mapping center point refers to the mapping point corresponding to the center point of the additional image in the initial image.

[0085] Based on this, according to the positioning information, the mapping center point corresponding to the additional image in the initial image can be calculated. For example, according to the vertex coordinates corresponding to the four vertices of the additional image in the initial image, the intersection point of the lines connecting the intersecting vertices is calculated and the intersection point is used as the mapping center point.

[0086] Furthermore, due to the wide range of user groups, most users usually do not have the ability to know specific coordinates. Therefore, in order to facilitate the user's description of the positioning information, the user can provide a description of the area, and then locate the location area corresponding to the additional image in the initial image. In this embodiment, the mapping center point corresponding to the additional image in the initial image is determined according to the positioning information in the additional image information, and the specific implementation is as follows:

[0087] Determining, according to the positioning information in the additional image information, a location area corresponding to the additional image in the initial image;

[0088] Determine a center point of the location area, and use the center point of the area as a mapping center point corresponding to the additional image in the initial image.

[0089] Specifically, the position area refers to the area in the initial image that corresponds to the additional image and can represent its position. In practical applications, the initial image can be pre-divided into multiple position areas, and the positioning information can be description information of the position area to which the image is to be mapped. Based on the positioning information, the position area corresponding to the additional image in the initial image can be determined.

[0090] Correspondingly, the area center point refers to the center point of the location area. For example, if the location area is a circle, the area center point is the center of the circular area.

[0091] Continuing with the above example, based on determining the additional image IMAGE1, the positioning information of the additional image IMAGE1 in the additional image information is the 5th position area, then the 5th position area is determined according to the position areas pre-divided for the initial clothing image, and the area center point P1 of the 5th position area is used as the mapping center point P1 corresponding to the additional image IMAGE1 in the initial clothing image.

[0092] In summary, based on the positioning information, the position area corresponding to the additional image in the initial image is determined, and the area center point of the position area is used as the mapping center point, which simplifies the user's description of the mapping position corresponding to the additional image and ensures the mapping accuracy.

[0093] Step 2044: Map the additional image to the initial image according to the correspondence between the mapping center point and the image center point of the additional image, and scale the additional image mapped to the initial image according to the size information in the additional image information to obtain an intermediate image.

[0094] Specifically, the image center point refers to the center point of the additional image. The size information refers to information used to describe the size of the additional image. The size information can be a description of the length and width in pixels or in centimeters. For example, the size information is: length = 1024 pixels and width = 512 pixels.

[0095] Since the original size information of the additional image determined based on the image processing request may not match the size information required after mapping, it is necessary to scale the size of the additional image so that the additional image meets the size mapping requirement.

[0096] It should be noted that the scaling process of the additional image can also be performed before the additional image is mapped. In addition, considering that the additional image may be required to have an angle offset with the additional image after mapping. Therefore, the additional image can be rotated. In addition to determining the mapping center point as described above, it is also necessary to determine the rotation angle, which can also be part of the positioning information. The additional image is then rotated based on the rotation angle.

[0097] Using the above example, based on determining the mapping center point P1 corresponding to the additional image IMAGE1 in the initial clothing image, the corresponding relationship between the image center point P2 of the additional image IMAGE1 and the mapping center point P1 is established, and the additional image is mapped to the initial image so that the mapped image center point P2 coincides with the mapping center point P1. The size information in the additional image information is: length = 512 pixels, width = 256 pixels, and the image size of the additional image IMAGE1 is: length = 1024 pixels, width = 512 pixels. Then, according to the size information, the area corresponding to the additional image IMAGE1 is enlarged to twice the size of the additional image IMAGE1, and the intermediate clothing image IMAGE2 of the mapped additional image is obtained.

[0098] In summary, after determining the mapping center point corresponding to the additional image, the additional image is mapped, and the additional image is scaled according to the size information, which ensures that the position and size of the additional image after mapping meet the image processing requirements of the user.

[0099] Step 206: Determine the distortion control points corresponding to the image region where the additional image is located in the intermediate image, and create a curved surface region corresponding to the image region based on the distortion control points.

[0100] Specifically, based on the intermediate image obtained above, since the image area where the additional image is located and other areas in the intermediate image are spliced ​​together, there is a lack of real visual effect. In order to improve the realism of the image, the image area where the additional image is located in the intermediate image can be distorted to obtain a curved surface area corresponding to the image area.

[0101] The image area refers to the area where the additional image is located in the intermediate image. Correspondingly, the distortion control point refers to the control point formed by distorting the control point set in the image area. The number of the distortion control points is one or more. In practical applications, the number of distortion control points can be determined according to the specific distortion method, and is not limited here. The curved surface area refers to the area formed by the distortion processing based on the image area.

[0102] Since the image area is distorted to form a curved surface, the intermediate image can present a more realistic or better visual effect. In order to ensure the controllability of the distortion effect, the distortion control points corresponding to the image area can be determined first, and then the curved surface area can be created based on the distortion control points. In specific implementation, the distortion control points corresponding to the image area where the additional image is located can be determined by selecting a number of distortion control points in the intermediate image according to a preset selection rule, such as randomly selecting 16 distortion control points within a preset range from the image area. Or, according to a preset calculation formula, the distortion control points corresponding to the four vertices of the area are calculated, etc., which are not limited here.

[0103] On the basis of determining the distortion control points, the distortion control points are used as the determined points in the surface, and then other surface parts are created based on these distortion control points to obtain the surface area corresponding to the image area. It should be noted that in the process of creating the surface area, the image area needs to be affine transformed into the surface so that the created surface area contains the image features of the image area.

[0104] Furthermore, if only the image area corresponding to the additional image is distorted, while other areas connected to the image area remain unchanged, the image may be discontinuous and appear abrupt. Therefore, in this embodiment, the distortion control point corresponding to the image area where the additional image is located in the intermediate image is determined, and the specific implementation is as follows:

[0105] Determining a mapping relationship between the intermediate image and at least one preset area wireframe;

[0106] According to the mapping relationship, determining a target area wireframe corresponding to the image area where the additional image is located in at least one preset area wireframe;

[0107] Selecting a region control point in the wireframe region included in the target region wireframe, and using the region control point as a control point corresponding to the image region;

[0108] Based on the correspondence between the preset control point and the preset distortion control point, the distortion control point corresponding to the control point is determined.

[0109] The preset area wireframe refers to a preset wireframe, and there is a mapping relationship between the wireframe and the area in the image. Among them, the mapping relationship refers to the one-to-one correspondence between the area in the image and the wireframe. Correspondingly, the target area wireframe refers to the preset area wireframe corresponding to the image area. The wireframe area refers to the area in the image corresponding to the target area wireframe. In practical applications, the wireframe area can be the same area as the image area of ​​the attached image. In addition, the wireframe area can also be an area that includes the image area, or an area that overlaps with the image area. There is no restriction here. Correspondingly, the area control point refers to the control point selected in the wireframe area to create the surface area. The preset control point refers to the preset control point. Correspondingly, the preset distortion control point refers to the preset distortion control point. In specific implementation, preset control points and distortion control points corresponding to the preset control points can be set in advance for the wireframe area.

[0110] In practical applications, since the distortion is to make the image look more natural and real, the target object that needs to be wrinkled or bent in the intermediate image is usually distorted, such as clothing, curtains, plants, sea surface, etc. Based on this, at least one preset area wireframe corresponding to the target object setter contained in the intermediate image can be pre-set, and then the target area wireframe can be determined in the preset area wireframe. Specifically, the method for determining the target area wireframe can be to determine the area center point of the image area where the additional image is located in the target object, and use the preset area wireframe corresponding to the area center point as the target area wireframe. It can also be determined whether there is an overlapping part between the image area where the additional image is located in the target object and the wireframe area contained in the preset area wireframe, and the preset area wireframe with the overlapping part is used as the target area wireframe. In addition, other methods can also be used to determine the target area wireframe, which is not limited here.

[0111] On the basis of determining the wireframe of the target area, the wireframe area contained in the wireframe of the target area is taken as the area to be distorted, and at least one control point is selected in the wireframe area according to the preset control point selection rule. The selected control point is used as the control point of the image area corresponding to the additional image. Specifically, the control points can be selected in the wireframe area in an equidistant manner, for example, the wireframe area is divided into 4 rows and 4 columns, and 4 control points are selected in each row and column in an equidistant manner, so a total of 16 control points are selected, and these 16 control points are used as the control points corresponding to the image area. In addition, a number of control points can be randomly selected, and then the selected control points are used as the control points corresponding to the image area, which is not limited here.

[0112] After the control points are selected, in order to construct the curved surface area corresponding to the image area, it is necessary to determine the distortion control points corresponding to the selected control points. In specific implementation, the correspondence between the preset control points and the preset distortion control points can be pre-set, and then the distortion control points corresponding to the control points can be determined based on the corresponding relationship after the control points are determined. The corresponding relationship can be calculated based on a preset distortion matrix or a preset distortion algorithm, which is not limited here.

[0113] Using the above example, based on the intermediate clothing image IMAGE2, the intermediate clothing image IMAGE2 contains a T-shirt, and 6 preset area wireframes are pre-set for the T-shirt. The target area wireframe corresponding to the image area where the additional image is located in these 6 preset area wireframes is the 5th area wireframe. Then 16 control points are selected in the wireframe area contained in the 5th area wireframe, as shown in the following figure: Figure 3 According to the correspondence between the preset control points and the preset distortion control points, the 16 distortion control points corresponding to the 16 controls are determined, as shown in (a). Figure 3 As shown in (b).

[0114] In summary, by selecting control points in the wireframe area contained in the wireframe of the target area corresponding to the image area, a reasonable selection of the area to be distorted is achieved, and then the distortion control points corresponding to these control points are determined to ensure the rationality of the distortion degree of the distortion control points.

[0115] Furthermore, there are various ways to create a curved surface area based on the distortion control points. In this embodiment, the curved surface area corresponding to the image area is created based on the distortion control points, and the specific implementation is as follows:

[0116] Obtaining distortion position information corresponding to the distortion control point;

[0117] Performing surface calculation based on the distorted position information to obtain an initial surface area corresponding to the wireframe area;

[0118] Based on the mesh correspondence between the wireframe area and the initial curved surface area, the wireframe area is affine transformed to the initial curved surface area to obtain a curved surface area.

[0119] Specifically, the distortion position information refers to the position coordinate information of the distortion control point. The initial surface area refers to the surface before pixel filling based on the wireframe area.

[0120] The surface calculation is performed based on the distortion position information, and the construction of the initial surface area can be realized according to a preset surface calculation formula. Specifically, the surface calculation formula can be a Bezier surface calculation formula shown in Formula 1.

[0121]

[0122] Among them, B m,i (u) and B n,j (v) is the basis function, m represents the number of control points in a single row in the horizontal axis direction minus 1, and n represents the number of control points in a single column in the vertical axis direction minus 1. In addition, the surface calculation formula may also be a spherical surface calculation formula, an ellipsoid surface calculation formula, a single-leaf hyperboloid surface calculation formula, a double-leaf hyperboloid surface calculation formula, or a hyperbolic paraboloid surface calculation formula. In specific implementation, a suitable surface calculation formula may be selected according to the needs of the actual application, and no limitation is made here.

[0123] On the basis of obtaining the initial surface area, since the initial surface area does not fill the pixels of the wireframe area, it is necessary to affine transform the wireframe area to the initial surface area according to the area correspondence relationship to obtain the surface area, thereby ensuring the integrity of the regional surface. Among them, the grid correspondence relationship can be understood as the correspondence between the grids in the two areas. Affine transformation is a linear transformation from two-dimensional coordinates to two-dimensional coordinates, and keeps the relative positions between two-dimensional graphics unchanged, that is, parallel lines before transformation are still parallel lines after transformation, straight lines before transformation are still straight lines after transformation, and the position order and length ratio of points on the same straight line remain unchanged before and after the change. Therefore, based on the grid correspondence relationship, the area contained in the grid in the wireframe area is affine transformed to the area contained in the corresponding grid in the initial surface area to achieve pixel filling of the initial surface area.

[0124] In summary, the surface calculation is performed based on the distortion position information of the distortion control point to obtain the initial surface area. Then the wireframe area is affine transformed to the initial surface area to obtain the surface area, realizing the surface deformation based on the wireframe area, and ensuring the rationality and aesthetics of the created surface area.

[0125] Furthermore, in order to ensure that the image features of the wireframe area can be better presented in the obtained regional surface, in this embodiment, based on the mesh correspondence between the wireframe area and the initial surface area, the wireframe area is affine transformed to the initial surface area to obtain the surface area, which is specifically implemented as follows:

[0126] Performing meshing processing on the wireframe area and the initial curved surface area respectively to obtain a mesh area corresponding to the wireframe area and a curved surface mesh area corresponding to the initial curved surface area;

[0127] Constructing a region correspondence relationship between the grid region and the curved surface grid region;

[0128] Based on the region correspondence, the mesh region is affine transformed into the curved surface mesh region to obtain a curved surface region.

[0129] Specifically, meshing refers to dividing an area into at least one subdivision area through a mesh. For example, for a wireframe area, the wireframe area can be divided into 10*10 mesh areas at equal intervals. For the initial surface area, since the initial surface area is a curved surface, the initial surface area can also be divided into 10*10 surface mesh areas according to the surface subdivision function. Among them, the mesh area refers to the area corresponding to the mesh generated after the wireframe area is meshed. Assuming that the wireframe area is divided into 100 meshes, these 100 meshes have corresponding mesh areas. Similarly, the surface mesh area refers to the area corresponding to the mesh generated after the initial surface area is meshed. In specific implementation, the mesh area obtained by meshing can be a square, a rectangle, or a triangle.

[0130] Correspondingly, the region correspondence refers to the correspondence between the mesh area and the surface mesh area obtained by division. Specifically, the region correspondence can be constructed according to the correspondence between the arrangement order of the mesh area in the wireframe area and the arrangement order of the surface mesh area in the surface mesh area. For example, if mesh area 1 is arranged in the first position from left to right and from top to bottom in the wireframe area, then the mesh area 1 has a region correspondence with the surface mesh area arranged in the first position from left to right and from top to bottom in the surface mesh area. Based on the region correspondence, an affine transformation is performed between the mesh area and the surface mesh area to obtain the surface area R of the fused additional image.

[0131] Continuing with the above example, after determining the 16 distortion control points corresponding to these 16 controls, obtain the position coordinates corresponding to these 16 distortion control points. Input these 16 position coordinates into the Bezier surface calculation formula for surface calculation to obtain the initial surface area corresponding to the 5th wireframe area. Then divide the 5th wireframe area into 10*10 grids according to equidistant spacing, connect the grid diagonals of these 100 grids to generate 200 triangles, and obtain the grid areas corresponding to these 200 triangles. And divide the initial surface area into 10*10 grids according to the surface subdivision function, then connect the grid diagonals of these 100 grids to generate 200 triangles, and obtain the surface grid areas corresponding to these 200 triangles. The 200 triangular mesh areas are numbered in order of arrangement, and the 200 surface mesh areas are numbered in the same order of arrangement, and then a region correspondence between the mesh areas with the same number and the surface mesh areas is constructed. Based on the region correspondence, the 200 mesh areas are affine transformed to the corresponding surface mesh areas to obtain the surface areas.

[0132] In summary, on the basis of mesh division, the mesh area is affine transformed to the surface mesh area based on the area correspondence to obtain the surface area, thereby ensuring the accuracy of the affine transformation and the aesthetics of the surface area.

[0133] Step 208: Update the intermediate image using the curved surface area, and obtain a target image fused with the additional image according to the update result.

[0134] Specifically, based on the above creation of the curved surface region, the intermediate image is updated using the curved surface region to obtain a target image fused with the additional image.

[0135] The target image refers to the image obtained by updating the intermediate image based on the curved surface area. Specifically, updating the intermediate image using the curved surface area can be understood as changing the color, transparency, cropping, etc. of the intermediate image based on the curved surface area; in addition, if there is no other change requirement, the intermediate image can also be directly used as the target image, which is not limited here.

[0136] Using the above example, on the basis of obtaining the surface area R, the intermediate clothing image IMAGE2 of the fused surface area R is directly used as the target clothing image IMAGE_T.

[0137] On the basis of obtaining the target image fused with the additional image, considering that the target image also needs to be fused with other aspects of background, color or style. In this embodiment, before the intermediate image is updated using the curved surface area and the target image fused with the additional image is obtained according to the update result, the method further includes:

[0138] Determining a template image according to the template information carried in the image processing request;

[0139] Correspondingly, after the intermediate image is updated by using the curved surface area and a target image fused with the additional image is obtained according to the update result, the method further includes:

[0140] The target image is fused with the template image to obtain a fused image.

[0141] Specifically, template information can be understood as relevant information used to describe the template image, and the template image can be determined based on the template information. Among them, the template image refers to the image that needs to be fused with the target image, and the specific content of the template image needs to correspond to the image content of the target object. For example, if the target image is a clothing image, then the template image can be a model image; if the target image is an animal image, then the template image is a natural scenery image such as a park. In addition, the template image can also be an image used to change the style or color of the target image, which is not limited here. Correspondingly, the fused image is an image generated by fusing the target image with the template image.

[0142] Based on this, the template image is determined according to the template information, and the template image can be determined according to the correspondence between the template information and the template image established in advance. In addition, the template image can also be determined by searching the template image according to the template information. On the basis of determining the template image, the target image and the template image are fused, which can be pixel fusion or feature fusion, etc., which is not limited here.

[0143] In summary, on the basis of obtaining the target image, the template image to be fused is determined, and the target image is fused with the template image, thereby improving the richness of the target image.

[0144] Furthermore, in order to achieve a better fusion effect, a series of image processing can be performed in the process of determining the template image, so as to obtain the template image to be fused. In this embodiment, the template image is determined according to the template information carried by the image processing request, and the specific implementation is as follows:

[0145] Obtaining an initial template image corresponding to the template information carried in the image processing request;

[0146] Reading a channel value of at least one pixel channel in the initial template image, and calculating a channel mean based on the channel value of at least one pixel channel;

[0147] Selecting a target pixel channel from at least one pixel channel according to the channel mean, and determining a target channel value of the target pixel channel;

[0148] Screening target pixels in the initial template image based on the target channel value, and calculating the brightness adjustment amplitude corresponding to the target pixels;

[0149] The brightness of the initial template image is adjusted based on the brightness adjustment amplitude to obtain a template image.

[0150] Specifically, the initial template image refers to the unprocessed image directly corresponding to the template information. Correspondingly, the channel value refers to the value of the pixel channel. The channel mean refers to the channel mean calculated for at least one pixel channel. The target pixel channel refers to the pixel channel selected from at least one pixel channel according to the pixel mean. Correspondingly, the target channel value is the pixel value of the target pixel channel. The target pixel point refers to the pixel point selected from the pixel points contained in the initial template image according to the target channel value, and the number of the target pixel points can be one or more.

[0151] The brightness adjustment amplitude refers to the amplitude used to describe the brightness adjustment of the initial template image. The brightness adjustment amplitude can be a percentage value, such as 40% means that a 40% brightness adjustment is performed on the basis of the initial template image. In addition, the brightness adjustment amplitude can also be expressed as a decimal, such as 0.4, -0.4, etc., which is not limited here.

[0152] Based on this, on the basis of obtaining the initial template image corresponding to the template information, when the initial template image is an RGB image, the channel values ​​of the three pixel channels in the initial template image can be read. And the pixel channel with a larger channel mean value is selected as the target pixel channel. In addition, the pixel channel with a smaller channel mean value can be selected as the target pixel channel from at least one pixel channel. In specific implementation, the selection method can be determined according to actual application requirements, and no limitation is made here.

[0153] Further, on the basis of determining the target pixel channel, the channel value of the target pixel channel (i.e., the target channel value) is obtained. Then, the target pixel points are screened in the initial template image. The specific method of screening the target pixel points includes: selecting at least one pixel point whose target channel value of the target pixel channel is greater than a preset channel threshold (a threshold value pre-set for the channel value), and based on the difference between the target channel value of the target pixel channel of at least one pixel point and the preset channel threshold, selecting a target pixel point whose target channel value is closer to the preset channel threshold from at least one pixel point. On the basis of screening the target pixel points, the brightness adjustment amplitude corresponding to the target pixel point is calculated. Specifically, the pixel value difference between the preset pixel value (such as the pixel value corresponding to white) and the target pixel value of the target pixel point (the pixel value of the target pixel point) can be calculated, and then the brightness adjustment amplitude corresponding to the above pixel value difference can be determined according to the corresponding relationship between the preset pixel value difference and the preset brightness adjustment amplitude.

[0154] Using the above example, if the model posture information carried by the clothing processing request is front-to-pocket, then the initial model image corresponding to the front-to-pocket is obtained. The initial model image contains 100 pixels, and the pixels contain 3 pixel channels, namely R channel, G channel and B channel. Then the channel mean Rm of the R channel value of the 100 pixels in the initial template image, the channel mean Gm of the G channel and the channel mean Bm of the B channel are calculated respectively. Among them, Rm>Gm>Bm, then the R channel is selected as the target pixel channel among these three channel means. Determine the channel value of the R channel of the 100 pixels. And select the target pixel points greater than the preset channel threshold C from the R channel value of the 100 pixels. Specifically, the target pixel points include pixel points 70 to pixel point 85. Then calculate according to the target pixel values ​​of the target pixel points from pixel points 70 to pixel point 85, and obtain the brightness adjustment range corresponding to adjusting the pixel values ​​of these pixels to 255 as Range. Then adjust the brightness of the initial model image according to the brightness adjustment range Range to obtain the model image.

[0155] In summary, the brightness adjustment amplitude corresponding to the target pixel point is calculated based on the target pixel point of the initial template image, and the brightness of the initial template image is adjusted to obtain the template image, thereby improving the image brightness of the template image and reducing interference factors during subsequent image fusion.

[0156] Furthermore, in order to avoid interference factors during image fusion and improve the image fusion effect, the target image can be cut, shadows can be added, and the processed image can be fused with the template image. In this embodiment, the target image is fused with the template image to obtain a fused image, and the specific implementation is as follows:

[0157] Determining a template region corresponding to the template image and an object region of a target object in the target image;

[0158] Based on the template area and the object area, determining an overlapping area, and removing the overlapping area in the object area to obtain a target object area;

[0159] Performing a shearing process on the target image based on the target object area, and adding a shadow layer corresponding to the target object to the sheared target image to obtain an object shadow image;

[0160] The object shadow image is mixed with the template image to obtain a fused image.

[0161] Specifically, the template area refers to the image area remaining after removing the object area of ​​the target object in the template image; wherein the target object is the image subject in the target image, such as clothing in a clothing image, an animal in an animal image, etc. Correspondingly, the object area refers to the image area corresponding to the target object in the target image. The overlapping area refers to the overlapping part of the template area and the object area after aligning the target image and the target object in the template image. Correspondingly, the target object area refers to the image area remaining after removing the overlapping area in the object area. The shadow layer refers to the layer pre-set for the target object for adding a shadow. Correspondingly, the object shadow layer refers to the image obtained after adding a shadow layer to the cut target image.

[0162] Based on this, on the basis of obtaining the target object area, the target image is sheared based on the target object area, which can be understood as performing a shearing mask on the target image. Specifically, it means shearing the image area except the target object area in the target image. Then the target image area remains in the sheared target image. Then a shadow layer is added to the target image area. Since different target objects have different shadows. Therefore, it is necessary to add a shadow layer pre-made for the target object to the sheared target image to obtain an object shadow image with a shadow effect. Then the object shadow image is mixed with the template image to obtain a fused image. Specifically, the image mixing method includes the positive film overlay method, the weighted average method, the Bayesian estimation method, and the like.

[0163] Specifically, the calculation formula for the multiplication is: A*B / 255, which means that the grayscale values ​​of each pixel corresponding to the red channel in the A image and the red channel in the B image are multiplied, and then divided by 255. A new red channel is obtained. Then the grayscale values ​​of each pixel corresponding to the green channel in the A image and the green channel in the B image are multiplied, and then divided by 255. A new green channel is obtained. Finally, the grayscale values ​​of each pixel corresponding to the blue channel in the A image and the blue channel in the B image are multiplied, and then divided by 255, to obtain a new blue channel. This process can also be simply called the process of matrix pixel multiplication.

[0164] Using the above example, determine the model area excluding the clothing area in the model image, and the clothing area corresponding to the T-shirt in the target clothing image IMAGE_T. Determine the overlapping area between the model area and the clothing area. Remove the overlapping area from the clothing area to obtain the target clothing area. Cut the clothing area based on the target clothing area to obtain the cut target clothing area, as shown in the following figure. Figure 4 As shown in (a), the shadow layer corresponding to the T-shirt is added to the target clothing area to obtain the clothing shadow image. Figure 4(b) in the figure) is used to perform positive overlay blending to obtain a clothing fitting image.

[0165] In summary, a fused image is obtained by removing the overlapping area in the object area of ​​the target image and then superimposing a shadow layer, and mixing the object shadow image generated after superimposing the shadow layer with the template image. This avoids the image area other than the object area in the target image affecting the fusion effect, and adds a shadow layer to achieve the addition of shadow effects, thereby increasing the authenticity of the fused image.

[0166] In practical applications, on the basis of obtaining a fused image, some image regions in the fused image can be modified or replaced. In this process, the edge regions can be fused by image difference. After modification or replacement, there may be jagged shapes at the edges of the modified or replaced regions. In this case, it is also necessary to smooth the edges of the regions, such as Gaussian blur processing and / or corrosion processing, so that the edges of the modified or replaced regions appear smooth.

[0167] Specifically, Figure 5 As shown in the figure, a clipping mask is applied to the distorted clothing image (target image) to remove the occlusion of the template image, and then a shadow layer is superimposed on the clipping mask to generate an object shadow image. For the model template image, the brightness is extracted, and the highlight range in the model template image is adjusted to white based on the brightness adjustment to obtain the template image. The object shadow image and the template image are matrix-pixel multiplied (multiply and blend) to obtain a fused image.

[0168] Based on the above-mentioned receiving of the image processing request submitted by the user through the client, the image processing result needs to be returned to the client. In order to improve the image transmission efficiency, the fused image can be compressed before being transmitted to the client. In this embodiment, after the target image is fused with the template image to obtain the fused image, the method further includes:

[0169] The fused image is compressed to obtain a compressed image, and the compressed image is sent to the client.

[0170] Specifically, the client can be an image processing APP (application) or other APP with image processing requirements, such as e-commerce APP, etc., which is not limited here. The compressed image refers to the image obtained by compressing the fused image. In specific implementation, the fused image can be compressed by compression algorithms such as run length coding, arithmetic coding and Huffman coding.

[0171] Continuing with the above example, based on the clothing try-on image obtained above, the clothing try-on image is compressed to obtain a compressed try-on image, and the compressed try-on image is returned to the client.

[0172] In summary, after compressing the fused image information, the compressed image is sent to the client, which can improve the image transmission efficiency and enhance the user's image processing experience.

[0173] Specifically, Figure 6 As shown, after the user inputs the clothing pattern and submits the demand file (the clothing pattern and the submitted demand file are carried in the clothing fitting request), the input clothing pattern is image calibrated (image format calibration and pixel channel calibration), and the demand file is parsed to obtain clothing pattern information (clothing style, pattern position size) and model parameters (face, matching and posture). The clothing pattern is mapped to the clothing image, and the Bezier surface deformation algorithm is used to perform texture distortion on the image area where the clothing pattern is located in the clothing image to obtain the surface area. The clothing image of the fused surface area and the model image are image fused by the positive film overlay algorithm to achieve wrinkle shadow synthesis. On the basis of generating a basic fused image, the parsed model parameters match the model's face, matching and posture. The fused image is further fused based on the image materials corresponding to these model parameters, and edge smoothing and background fusion are performed during the fusion process to output a virtual model composite image. The generated virtual model composite image is then returned to the user.

[0174] The image processing method provided in the embodiment of this specification, in order to ensure the generation effect of the target image, receives the image processing request submitted by the user, and determines the additional image, the additional image information and the initial image according to the image processing request, and maps the additional image to the initial image according to the additional image information, obtains the intermediate image, and realizes the preliminary fusion of the additional image and the initial image. Then, the distortion control point corresponding to the image area where the additional image is located in the intermediate image is determined, and the surface area corresponding to the image area is created based on the distortion control point, so as to realize the creation of the image area of ​​the additional image to be curved, and then the intermediate image is updated by using the surface area, and the target image fused with the additional image is obtained according to the update result, so as to avoid the low resolution and loss of image details caused by processing the image by using the model calculation method, and make the generated target image have a three-dimensional effect, and increase the realism of the target image.

[0175] See also Figure 7 , Figure 7 A flowchart of another image processing method provided according to an embodiment of the present specification is shown, which specifically includes the following steps.

[0176] Step 702: receiving a clothing try-on request submitted by a user, and determining an additional image, additional image information, an initial clothing image, and a model image according to the clothing try-on request;

[0177] Step 704: mapping the additional image to the initial clothing image according to the additional image information to obtain an intermediate clothing image;

[0178] Step 706: determining the distortion control points corresponding to the image region where the additional image is located in the intermediate clothing image, and creating a curved surface region corresponding to the image region based on the distortion control points;

[0179] Step 708: using the curved surface area to update the intermediate clothing image, and fusing the updated intermediate clothing image with the model image to obtain a clothing fitting image.

[0180] Specifically, a clothing try-on request submitted by a user is received, and according to the clothing try-on request, a clothing pattern image, clothing pattern image information, an initial clothing image, and a model image are determined, and according to the positioning information and size information in the clothing pattern image information, the clothing pattern image is mapped to the initial clothing image to obtain an intermediate clothing image. Then, 16 control points are selected in the image area where the clothing pattern image is located in the intermediate clothing image, and the distortion control points corresponding to the 16 control points are determined. Surface calculation is performed based on the distortion control points to generate an initial surface area corresponding to the image area. The image area is affine transformed to the initial surface area to generate a surface area. The intermediate clothing area is updated using the surface area to obtain an updated intermediate clothing image. The updated intermediate clothing image is fused with the model template image to obtain a clothing try-on image.

[0181] The image processing method provided in the embodiment of the present specification, in order to ensure the generation effect of clothing try-on images, receives a clothing try-on request submitted by a user, and determines an additional image, additional image information, an initial clothing image and a model image according to the clothing try-on request, and maps the additional image to the initial clothing image according to the additional image information, obtains an intermediate clothing image, and realizes the preliminary fusion of the additional image and the initial clothing image. Then, the distortion control points corresponding to the image area where the additional image is located in the intermediate clothing image are determined, and a curved surface area corresponding to the image area is created based on the distortion control points, so as to realize the creation of a curved surface of the image area of ​​the additional image, and then the intermediate clothing image is updated using the curved surface area, and the target clothing image fused with the additional image is obtained according to the update result, so as to avoid the low resolution and image detail loss caused by processing the clothing image in a model calculation manner, and make the generated target clothing image have a three-dimensional effect, thereby increasing the realism of the target clothing image.

[0182] The following combination Figure 8 , taking the application of the image processing method provided in this specification in automobile images as an example, the image processing method is further described. Figure 8A processing flow chart of an image processing method provided by an embodiment of the present specification is shown, which specifically includes the following steps.

[0183] Step 802: receiving a car image processing request submitted by a user through a client, and determining a flag image, flag image information, an initial car image, and a city background image according to the car image processing request.

[0184] Step 804: Map the flag image to the initial car image according to the flag image information to obtain an intermediate car image.

[0185] Step 806: Determine a mapping relationship between the intermediate vehicle image and at least one preset area wireframe, and determine a target area wireframe corresponding to the flag image region where the flag image is located in the at least one preset area wireframe according to the mapping relationship.

[0186] Step 808: Selecting a region control point in the wireframe area included in the target region wireframe, and using the region control point as a control point corresponding to the image region.

[0187] Step 810: Determine the distortion control point corresponding to the control point based on the correspondence between the preset control point and the preset distortion control point.

[0188] Step 812: Obtain the distortion position information corresponding to the distortion control point, and perform Bezier surface calculation based on the distortion position information to obtain an initial surface area corresponding to the wireframe area.

[0189] Step 814: Perform triangular meshing processing on the wireframe area and the initial curved surface area respectively to obtain a mesh area corresponding to the wireframe area and a curved surface mesh area corresponding to the initial curved surface area.

[0190] Step 816: constructing a region correspondence relationship between the mesh region and the surface mesh region, and based on the region correspondence relationship, affine transforming the mesh region to the surface mesh region to obtain the surface flag region.

[0191] Step 818: Use the curved flag region to update the intermediate car image, and obtain the target car image fused with the flag image according to the update result.

[0192] Step 820: Determine the city background area corresponding to the city background image and the car area of ​​the car in the target car image.

[0193] Step 822: Based on the city background area and the car area, determine the overlapping area, and remove the overlapping area in the car area to obtain the target car area.

[0194] Step 824: The target car image is sheared based on the target car area, and a shadow layer corresponding to the car is added to the sheared target car image to obtain a car shadow image.

[0195] Step 826: Blending the car shadow image with the city background image to obtain a city car image.

[0196] Step 828: compress the city car image to obtain a compressed city car image, and send the compressed city car image to the client.

[0197] The image processing method provided in the embodiment of the specification, in order to ensure the generation effect of the city car image, receives the car image processing request submitted by the user, and determines the flag image, flag image information, initial car image and city background image according to the car image processing request, and maps the flag image to the initial car image according to the additional image information, obtains the intermediate car image, and realizes the preliminary fusion of the flag image and the initial car image. Then, the distortion control point corresponding to the image area where the flag image is located in the intermediate car image is determined, and the surface area corresponding to the image area is created based on the distortion control point, so as to realize the creation of the image area of ​​the flag image to be curved, and then the intermediate car image is updated by using the surface area, and the target car image fused with the additional image is obtained according to the update result, so as to avoid the low resolution and image detail loss caused by processing the car image by using the model calculation method, and make the generated target car image have a three-dimensional effect, and increase the realism of the target car image.

[0198] Corresponding to the above method embodiment, this specification also provides an image processing device embodiment, Fig. 9 FIG. 2 shows a schematic diagram of the structure of an image processing device provided by an embodiment of the present specification. Fig. 9 As shown, the device comprises:

[0199] A first receiving module 902 is configured to receive an image processing request submitted by a user, and determine an additional image, additional image information, and an initial image according to the image processing request;

[0200] A first mapping module 904 is configured to map the additional image to the initial image according to the additional image information to obtain an intermediate image;

[0201] A first creation module 906 is configured to determine a distortion control point corresponding to an image region in the intermediate image where the additional image is located, and create a curved surface region corresponding to the image region based on the distortion control point;

[0202] The first updating module 908 is configured to update the intermediate image using the curved surface area, and obtain a target image fused with the additional image according to the updating result.

[0203] Optionally, the first creation module 902 includes:

[0204] A mapping relationship determination submodule is configured to determine a mapping relationship between the intermediate image and at least one preset area wireframe;

[0205] A wireframe determination submodule is configured to determine, according to the mapping relationship, a target area wireframe corresponding to the image area where the additional image is located in at least one preset area wireframe;

[0206] A selection submodule is configured to select a region control point in a wireframe region included in the target region wireframe, and use the region control point as a control point corresponding to the image region;

[0207] The control point determination submodule is configured to determine the distortion control point corresponding to the control point based on the correspondence between the preset control point and the preset distortion control point.

[0208] Optionally, the first creation module 902 includes:

[0209] An acquisition submodule, configured to acquire distortion position information corresponding to the distortion control point;

[0210] A calculation submodule is configured to perform surface calculation based on the distorted position information to obtain an initial surface area corresponding to the wireframe area;

[0211] The transformation submodule is configured to perform affine transformation of the wireframe area to the initial curved surface area based on the mesh correspondence between the wireframe area and the initial curved surface area to obtain a curved surface area.

[0212] Optionally, the transformation submodule is further configured to:

[0213] Performing meshing processing on the wireframe area and the initial curved surface area respectively to obtain a mesh area corresponding to the wireframe area and a curved surface mesh area corresponding to the initial curved surface area;

[0214] Constructing a region correspondence relationship between the grid region and the curved surface grid region;

[0215] Based on the region correspondence, the mesh region is affine transformed into the curved surface mesh region to obtain a curved surface region.

[0216] Optionally, the image processing method further includes:

[0217] An image determination module is configured to determine a template image according to the template information carried in the image processing request;

[0218] Accordingly, the image processing method further includes:

[0219] The fusion module is configured to fuse the target image with the template image to obtain a fused image.

[0220] Optionally, the image determination module is further configured to:

[0221] Obtaining an initial template image corresponding to the template information carried in the image processing request;

[0222] Reading a channel value of at least one pixel channel in the initial template image, and calculating a channel mean based on the channel value of at least one pixel channel;

[0223] Selecting a target pixel channel from at least one pixel channel according to the channel mean, and determining a target channel value of the target pixel channel;

[0224] Screening target pixels in the initial template image based on the target channel value, and calculating the brightness adjustment amplitude corresponding to the target pixels;

[0225] The brightness of the initial template image is adjusted based on the brightness adjustment amplitude to obtain a template image.

[0226] Optionally, the fusion module is further configured to:

[0227] Determining a template region corresponding to the template image and an object region of a target object in the target image;

[0228] Based on the template area and the object area, determining an overlapping area, and removing the overlapping area in the object area to obtain a target object area;

[0229] Performing a shearing process on the target image based on the target object area, and adding a shadow layer corresponding to the target object to the sheared target image to obtain an object shadow image;

[0230] The object shadow image is mixed with the template image to obtain a fused image.

[0231] Optionally, the first receiving module 902 is further configured to:

[0232] Obtaining an initial additional image carried in the image processing request;

[0233] The initial additional image is subjected to image format conversion and / or pixel channel conversion to obtain an additional image.

[0234] Optionally, the first mapping module 904 includes:

[0235] A center point determination submodule is configured to determine a mapping center point corresponding to the additional image in the initial image according to the positioning information in the additional image information;

[0236] The mapping submodule is configured to map the additional image to the initial image according to the correspondence between the mapping center point and the image center point of the additional image, and to scale the additional image mapped to the initial image according to the size information in the additional image information to obtain an intermediate image.

[0237] Optionally, the mapping submodule is further configured to:

[0238] Determining, according to the positioning information in the additional image information, a location area corresponding to the additional image in the initial image;

[0239] Determine a center point of the location area, and use the center point of the area as a mapping center point corresponding to the additional image in the initial image.

[0240] Optionally, the first receiving module 902 is further configured to:

[0241] Parsing the image processing request to obtain initial image identification information, additional image and additional image information; searching for the initial image corresponding to the initial image identification information in an image library according to the initial image identification information; or,

[0242] The image processing request is parsed to obtain initial image identification information, additional image identification information and additional image information; based on the initial image identification information and the additional image identification information, the initial image corresponding to the initial image identification information and the additional image corresponding to the additional image identification information are searched in the image library.

[0243] Optionally, the first receiving module 902 is further configured to:

[0244] Receive image processing requests submitted by users through the client;

[0245] Accordingly, the image processing method further includes:

[0246] The fused image is compressed to obtain a compressed image, and the compressed image is sent to the client.

[0247] An image processing device provided in an embodiment of the present specification, in order to ensure the generation effect of clothing try-on images, receives a clothing try-on request submitted by a user, and determines an additional image, additional image information, an initial clothing image and a model image according to the clothing try-on request, and maps the additional image to the initial clothing image according to the additional image information, obtains an intermediate clothing image, and realizes the preliminary fusion of the additional image and the initial clothing image. Then, the distortion control point corresponding to the image area where the additional image is located in the intermediate clothing image is determined, and a curved surface area corresponding to the image area is created based on the distortion control point, so as to realize the creation of a curved surface of the image area of ​​the additional image, and then the curved surface area is used to update the intermediate clothing image, and the target clothing image fused with the additional image is obtained according to the update result, so as to avoid the low resolution and loss of image details caused by processing the image using a model calculation method, and make the generated target image have a three-dimensional effect, thereby increasing the realism of the target image.

[0248] The above is a schematic scheme of an image processing device of this embodiment. It should be noted that the technical scheme of the image processing device and the technical scheme of the above-mentioned image processing method belong to the same concept, and the details not described in detail in the technical scheme of the image processing device can be referred to the description of the technical scheme of the above-mentioned image processing method.

[0249] Corresponding to the above method embodiment, this specification also provides another image processing device embodiment, Fig.10 FIG. 2 shows a schematic diagram of the structure of another image processing device provided by an embodiment of the present specification. Fig.10 As shown, the device comprises:

[0250] The second receiving module 1002 is configured to receive a clothing try-on request submitted by a user, and determine an additional image, additional image information, an initial clothing image, and a model image according to the clothing try-on request;

[0251] A second mapping module 1004 is configured to map the additional image to the initial clothing image according to the additional image information to obtain an intermediate clothing image;

[0252] The second creation module 1006 is configured to determine the distortion control points corresponding to the image region where the additional image is located in the intermediate clothing image, and create a curved surface region corresponding to the image region based on the distortion control points;

[0253] The second updating module 1008 is configured to update the intermediate clothing image using the curved surface area, and merge the updated intermediate clothing image with the model image to obtain a clothing fitting image.

[0254] The embodiment of the present specification provides another image processing device, which receives a clothing try-on request submitted by a user to ensure the generation effect of a clothing try-on image, and determines an additional image, additional image information, an initial clothing image, and a model image according to the clothing try-on request, and maps the additional image to the initial clothing image according to the additional image information, thereby obtaining a preliminary fusion of the additional image and the initial clothing image. Then, the distortion control points corresponding to the image area where the additional image is located in the intermediate clothing image are determined, and a curved surface area corresponding to the image area is created based on the distortion control points, thereby realizing the creation of a curved surface of the image area of ​​the additional image, and then the curved surface area is used to update the intermediate clothing image, and a target clothing image fused with the additional image is obtained according to the update result, so that the generated target clothing image has a three-dimensional effect, thereby increasing the realism of the target clothing image.

[0255] The above is a schematic scheme of another image processing device of this embodiment. It should be noted that the technical scheme of the image processing device and the technical scheme of the above another image processing method belong to the same concept, and the details not described in detail in the technical scheme of the image processing device can be referred to the description of the technical scheme of the above another image processing method.

[0256] Fig.11 The block diagram of a computing device 1100 according to one embodiment of the present specification is shown. The components of the computing device 1100 include but are not limited to a memory 1110 and a processor 1120. The processor 1120 is connected to the memory 1110 via a bus 1130, and a database 1150 is used to store data.

[0257] The computing device 1100 also includes an access device 1140 that enables the computing device 1100 to communicate via one or more networks 1160. Examples of these networks include a public switched telephone network (PSTN), a local area network (LAN), a wide area network (WAN), a personal area network (PAN), or a combination of communication networks such as the Internet. The access device 1140 may include one or more of any type of network interface (e.g., a network interface card (NIC)) whether wired or wireless, such as an IEEE 802.11 wireless local area network (WLAN) wireless interface, a World Wide Interoperability for Microwave Access (Wi-MAX) interface, an Ethernet interface, a universal serial bus (USB) interface, a cellular network interface, a Bluetooth interface, a near field communication (NFC) interface, and the like.

[0258] In one embodiment of the present specification, the above components of the computing device 1100 and Fig.11 Other components not shown in the figure may also be connected to each other, for example, via a bus. It should be understood that Fig.11The computing device structure block diagram shown is only for the purpose of illustration, and is not intended to limit the scope of this specification. Those skilled in the art can add or replace other components as needed.

[0259] The computing device 1100 may be any type of stationary or mobile computing device, including a mobile computer or mobile computing device (e.g., a tablet computer, a personal digital assistant, a laptop computer, a notebook computer, a netbook, etc.), a mobile phone (e.g., a smart phone), a wearable computing device (e.g., a smart watch, smart glasses, etc.), or other types of mobile devices, or a stationary computing device such as a desktop computer or PC. The computing device 1100 may also be a mobile or stationary server.

[0260] The processor 1120 is used to execute the following computer executable instructions, which implement the steps of the above-mentioned image processing method when executed by the processor.

[0261] The above is a schematic scheme of a computing device of this embodiment. It should be noted that the technical scheme of the computing device and the technical scheme of the above-mentioned image processing method belong to the same concept, and the details not described in detail in the technical scheme of the computing device can be referred to the description of the technical scheme of the above-mentioned image processing method.

[0262] An embodiment of the present specification further provides a computer-readable storage medium storing computer-executable instructions, which can implement the steps of the above-mentioned image processing method when executed by a processor.

[0263] The above is a schematic scheme of a computer-readable storage medium of this embodiment. It should be noted that the technical scheme of the storage medium and the technical scheme of the above-mentioned image processing method belong to the same concept, and the details not described in detail in the technical scheme of the storage medium can be referred to the description of the technical scheme of the above-mentioned image processing method.

[0264] An embodiment of the present specification further provides a computer program, wherein when the computer program is executed in a computer, the computer is caused to execute the steps of the above-mentioned image processing method.

[0265] The above is an illustrative solution of a computer program of this embodiment. It should be noted that the technical solution of the computer program and the technical solution of the above-mentioned image processing method belong to the same concept, and the details not described in detail in the technical solution of the computer program can be referred to the description of the technical solution of the above-mentioned image processing method.

[0266] The above is a description of a specific embodiment of the specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0267] The computer instructions include computer program codes, which may be in source code form, object code form, executable files or some intermediate forms, etc. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, USB flash drive, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electric carrier signal, telecommunication signal and software distribution medium, etc. It should be noted that the content contained in the computer-readable medium may be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media do not include electric carrier signals and telecommunication signals.

[0268] It should be noted that, for the above-mentioned method embodiments, for the sake of simplicity of description, they are all expressed as a series of action combinations, but those skilled in the art should be aware that the embodiments of this specification are not limited by the order of the actions described, because according to the embodiments of this specification, some steps can be performed in other orders or simultaneously. Secondly, those skilled in the art should also be aware that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily required by the embodiments of this specification.

[0269] In the above embodiments, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0270] The preferred embodiments of this specification disclosed above are only used to help explain this specification. The optional embodiments do not describe all the details in detail, nor do they limit the invention to only the specific implementation methods described. Obviously, many modifications and changes can be made according to the content of the embodiments of this specification. This specification selects and specifically describes these embodiments in order to better explain the principles and practical applications of the embodiments of this specification, so that technicians in the relevant technical field can well understand and use this specification. This specification is only limited by the claims and their full scope and equivalents.

Claims

1. An image processing method, include: Receiving an image processing request submitted by a user, the image processing request including: initial image identification information, additional images, and additional image information; Determining an additional image, additional image information, and an initial image according to the image processing request, wherein the additional image information includes image size information, image fusion area information, and image fusion effect information of the additional image; Mapping the additional image to the initial image according to the additional image information to obtain an intermediate image; Determine the distortion control points corresponding to the image area where the additional image is located in the intermediate image, and create a curved surface area corresponding to the image area based on the distortion control points, wherein the distortion control points are obtained by distortion of the control points set in the image area, and the determining of the distortion control points corresponding to the image area where the additional image is located in the intermediate image comprises: determining a mapping relationship between the intermediate image and at least one preset area wireframe; according to the mapping relationship, determine the area center point of the image area where the additional image is located in the target object, and use the preset area wireframe corresponding to the area center point as the target area wireframe, wherein the target object is an object presenting wrinkles or bends in the intermediate image; select area control points in the wireframe area included in the target area wireframe, and use the area control points as the control points corresponding to the image area; determine the distortion control points corresponding to the control points based on the correspondence between the preset control points and the preset distortion control points; The intermediate image is updated using the curved surface area, and a target image fused with the additional image is obtained according to the update result.

2. The image processing method according to claim 1, wherein the curved surface area corresponding to the image area is created based on the distortion control points. include: Obtaining distortion position information corresponding to the distortion control point; Performing surface calculation based on the distorted position information to obtain an initial surface area corresponding to the wireframe area; Based on the mesh correspondence between the wireframe area and the initial curved surface area, the wireframe area is affine transformed to the initial curved surface area to obtain a curved surface area.

3. The image processing method according to claim 2, wherein based on the mesh correspondence between the wireframe area and the initial curved surface area, the wireframe area is affine transformed into the initial curved surface area to obtain the curved surface area, include: Performing meshing processing on the wireframe area and the initial curved surface area respectively to obtain a mesh area corresponding to the wireframe area and a curved surface mesh area corresponding to the initial curved surface area; Constructing a region correspondence relationship between the grid region and the curved surface grid region; Based on the region correspondence, the mesh region is affine transformed into the curved surface mesh region to obtain a curved surface region.

4. The image processing method according to claim 1, wherein before the intermediate image is updated by using the curved surface area and a target image fused with the additional image is obtained according to the updated result, include: Determining a template image according to the template information carried in the image processing request; Correspondingly, after the intermediate image is updated by using the curved surface area and a target image fused with the additional image is obtained according to the update result, the method further includes: The target image is fused with the template image to obtain a fused image.

5. The image processing method according to claim 4, wherein the template image is determined according to the template information carried in the image processing request, include: Obtaining an initial template image corresponding to the template information carried in the image processing request; Reading a channel value of at least one pixel channel in the initial template image, and calculating a channel mean based on the channel value of at least one pixel channel; Selecting a target pixel channel from at least one pixel channel according to the channel mean, and determining a target channel value of the target pixel channel; Screening target pixels in the initial template image based on the target channel value, and calculating the brightness adjustment amplitude corresponding to the target pixels; The brightness of the initial template image is adjusted based on the brightness adjustment amplitude to obtain a template image.

6. The image processing method according to claim 5, wherein the target image is fused with the template image to obtain a fused image. include: Determining a template region corresponding to the template image and an object region of a target object in the target image; Based on the template area and the object area, determining an overlapping area, and removing the overlapping area in the object area to obtain a target object area; Performing a shearing process on the target image based on the target object area, and adding a shadow layer corresponding to the target object to the sheared target image to obtain an object shadow image; The object shadow image is mixed with the template image to obtain a fused image.

7. The image processing method according to claim 1, wherein the additional image is determined according to the image processing request. include: Obtaining an initial additional image carried in the image processing request; The initial additional image is subjected to image format conversion and / or pixel channel conversion to obtain an additional image.

8. The image processing method according to claim 1, wherein the additional image is mapped to the initial image according to the additional image information to obtain an intermediate image, include: Determining, according to the positioning information in the additional image information, a mapping center point corresponding to the additional image in the initial image; According to the correspondence between the mapping center point and the image center point of the additional image, the additional image is mapped to the initial image, and the additional image mapped to the initial image is scaled according to the size information in the additional image information to obtain an intermediate image.

9. The image processing method according to claim 8, wherein the mapping center point corresponding to the additional image in the initial image is determined according to the positioning information in the additional image information, include: Determining, according to the positioning information in the additional image information, a location area corresponding to the additional image in the initial image; Determine a center point of the location area, and use the center point of the area as a mapping center point corresponding to the additional image in the initial image.

10. The image processing method according to claim 1, wherein the additional image, the additional image information and the initial image are determined according to the image processing request. include: Parsing the image processing request to obtain initial image identification information, additional images, and additional image information; According to the initial image identification information, searching the image library for the initial image corresponding to the initial image identification information; or, The image processing request is parsed to obtain initial image identification information, additional image identification information and additional image information; based on the initial image identification information and the additional image identification information, the initial image corresponding to the initial image identification information and the additional image corresponding to the additional image identification information are searched in the image library.

11. The image processing method according to claim 4, wherein receiving an image processing request submitted by a user, include: Receive image processing requests submitted by users through the client; Correspondingly, after fusing the target image with the template image to obtain a fused image, the method further includes: The fused image is compressed to obtain a compressed image, and the compressed image is sent to the client.

12. An image processing method, include: Receive a clothing try-on request submitted by a user, and determine an additional image, additional image information, an initial clothing image, and a model image according to the clothing try-on request, wherein the additional image information includes image size information, image fusion area information, and image fusion effect information of the additional image; Mapping the additional image to the initial clothing image according to the additional image information to obtain an intermediate clothing image; Determine the distortion control points corresponding to the image area where the additional image is located in the intermediate clothing image, and create a curved surface area corresponding to the image area based on the distortion control points, wherein the distortion control points are obtained by distortion of the control points set in the image area, and the determination of the distortion control points corresponding to the image area where the additional image is located in the intermediate image includes: determining a mapping relationship between the intermediate image and at least one preset area wireframe; according to the mapping relationship, determine the area center point of the image area where the additional image is located in the target object, and use the preset area wireframe corresponding to the area center point as the target area wireframe, wherein the target object is an object showing wrinkles or bends in the intermediate image; select area control points in the wireframe area included in the target area wireframe, and use the area control points as control points corresponding to the image area; determine the distortion control points corresponding to the control points based on the corresponding relationship between the preset control points and the preset distortion control points; The intermediate clothing image is updated by using the curved surface area, and the updated intermediate clothing image is fused with the model image to obtain a clothing try-on image.

13. A computer-readable storage medium storing computer-executable instructions, wherein the computer-executable instructions, when executed by a processor, implement the steps of the image processing method according to any one of claims 1 to 12.