CN113051778A

CN113051778A - Clothing design method and device, electronic equipment and storage medium

Info

Publication number: CN113051778A
Application number: CN202110518507.6A
Authority: CN
Inventors: 魏辰峰; 陈达勤; 浣军; 李曈; 娄明; 宋博宁
Original assignee: Beijing Shenshang Technology Co ltd
Current assignee: Beijing Huimei Future Technology Co ltd
Priority date: 2021-05-12
Filing date: 2021-05-12
Publication date: 2021-06-29
Anticipated expiration: 2041-05-12
Also published as: CN113051778B

Abstract

The application provides a clothing design method, a clothing design device, electronic equipment and a storage medium, and relates to the field of clothing design. The method comprises the steps of taking a received design request as an input of a pre-trained first model to carry out operation to obtain a plurality of clothing design pictures; counting the current popular multiple clothing commodities in the clothing class corresponding to the design request; extracting common characteristics of the multiple clothes commodities to obtain target characteristics; calculating the feature similarity of the corresponding feature of each of the plurality of clothing design pictures and the target feature; and sending the at least one clothing design picture with the highest feature similarity between the features and the target features to a client so as to carry out clothing design according to the at least one clothing design picture. The garment design method, the garment design device, the electronic equipment and the storage medium can generate the garment with the appointed design points according to the requirements.

Description

Clothing design method and device, electronic equipment and storage medium

Technical Field

The present application relates to the field of garment design, and in particular, to a garment design method and apparatus, an electronic device, and a storage medium.

Background

As machine learning techniques and image generation techniques are becoming mature, machine learning is increasingly widely used in garment design. In particular, the generation of a very realistic and high-resolution clothing image has been enabled by the generation of a network for confrontation (GAN) technology in 2014, which mostly uses a multidimensional normal distribution sampling gaussian noise as an input of a GAN model, and the GAN model obtains a very realistic image through a series of operations.

However, such approaches have a common disadvantage of lacking control, i.e., difficulty in controlling the creation of a garment at a given design point.

Therefore, how to provide an effective solution to ensure that a garment with a specified design point can be generated becomes an urgent problem in the prior art.

Disclosure of Invention

The embodiment of the application provides a garment design method and device, electronic equipment and a storage medium, and aims to solve the problem that in the prior art, a garment with specified design points is difficult to generate.

The embodiment of the application adopts the following technical scheme:

in a first aspect, an embodiment of the present application provides a garment design method, including:

taking the received design request as the input of a pre-trained first model for operation to obtain a plurality of clothing design pictures;

counting the current popular multiple clothing commodities in the clothing class corresponding to the design request;

extracting common characteristics of the multiple clothes commodities to obtain target characteristics;

calculating the feature similarity of the feature corresponding to each of the plurality of clothing design pictures and the target feature;

sending at least one clothing design picture with the highest feature similarity between the features and the target features to a client so as to carry out clothing design according to the at least one clothing design picture;

the design request comprises a garment type and a garment style parameter, the first model is obtained by taking a known first garment design picture as output and taking the garment type and the garment style parameter corresponding to the first garment design picture as input for training.

In one possible design, the counting multiple clothing items currently popular in the clothing category corresponding to the design request includes:

according to the clothing commodity sales record of the third-party server, counting the commodity sales volume of the clothing commodity of the clothing class corresponding to the design request in the current time period;

and selecting the multiple clothing commodities with the highest commodity sales volume as the currently popular multiple clothing commodities.

In one possible design, the extracting common features of the multiple items of clothing to obtain target features includes:

respectively taking the pictures of the multiple clothing commodities as the input of a pre-trained second model for operation, and splicing the feature vectors output by the last two layers of the second model to obtain a plurality of initial features which are in one-to-one correspondence with the multiple clothing commodities;

extracting a feature vector common to the plurality of initial features to obtain the target feature;

the second model is obtained by training with a known second clothing design picture as input and clothing categories corresponding to the second clothing design picture as output.

In one possible design, the method further includes:

calling a third model corresponding to the clothing category;

taking the selected target clothing design picture in at least one clothing design picture as the input of the third model for operation to obtain a sheet parameter corresponding to the target clothing design picture;

3D rendering is carried out on the target clothing design picture according to the plate parameters;

the third model is obtained by training with a known third clothing design picture as input and plate parameters of the third clothing design picture as input.

In one possible design, the 3D rendering of the target apparel design picture according to the panel parameters includes:

3D rendering is carried out on the target clothing design picture according to the slab parameters and the generated fabric; or

And 3D rendering is carried out on the target clothing design picture according to the slab parameters and the generated printing.

In one possible design, the method further includes:

adjusting the slab parameters in response to the client side adjusting operation aiming at the slab parameters;

the 3D rendering of the target dress design picture according to the slab parameters comprises the following steps:

and 3D rendering is carried out on the target clothing design picture according to the adjusted plate parameters.

In one possible design, the first model is a styleGAN model or a BigGAN model.

In a second aspect, an embodiment of the present application provides a garment design device, including:

the first operation unit is used for operating the received design request as the input of a pre-trained first model to obtain a plurality of garment design pictures;

the statistical unit is used for counting the current popular multiple clothing commodities in the clothing class corresponding to the design request;

the extraction unit is used for extracting the common characteristics of the multiple clothing commodities to obtain target characteristics;

the second operation unit is used for calculating the feature similarity of the feature corresponding to each of the plurality of clothes design pictures and the target feature;

the sending unit is used for sending the at least one clothing design picture with the highest feature similarity between the features and the target features to a client so as to carry out clothing design according to the at least one clothing design picture;

In a third aspect, an embodiment of the present application provides an electronic device, including a processor, a communication interface, a memory, and a communication bus, where the processor, the communication interface, and the memory complete mutual communication through the bus;

a memory for storing a computer program;

the processor is used for executing the program stored in the memory and realizing the following processes:

In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, in which a computer program is stored, and when executed by a processor, the computer program implements the following procedures:

The above-mentioned at least one technical scheme that this application one or more embodiments adopted can reach following beneficial effect:

the design request is used as input operation of the first model to obtain a plurality of clothing design pictures, a plurality of clothing items which are popular at present in clothing types corresponding to the design request are counted, common features of the clothing items are extracted to obtain target features, then at least one clothing design picture with the highest feature similarity of the features and the target features is sent to the client side so as to carry out clothing design according to the at least one clothing design picture, the design request comprises clothing types and clothing style parameters, and the clothing design picture sent to the client side meets the clothing types and the clothing style parameters, meanwhile, the features of the clothing design picture are matched with the same features of the clothing items popular at present to the greatest degree, so that clothing with the designated design key points can be generated according to requirements.

Drawings

The accompanying drawings, which are included to provide a further understanding of the disclosure and are incorporated in and constitute a part of this specification, illustrate embodiments of the disclosure and together with the description serve to explain the disclosure without limiting the disclosure in any way. In the drawings:

fig. 1 is a schematic flow chart of a garment design method according to an embodiment of the present application.

Fig. 2 is a schematic structural diagram of an electronic device according to an embodiment of the present application.

Fig. 3 is a schematic structural diagram of a garment design device according to an embodiment of the present application.

Detailed Description

In order to solve the problem that a garment with designated design points is difficult to generate in the prior art, embodiments of the present application provide a garment design method, an apparatus, an electronic device and a storage medium, and the design method, the apparatus, the electronic device and the storage medium can generate a garment with designated design points according to requirements.

The clothing design method provided by the embodiment of the application can be applied to a server, and the server can be a network server or a database server.

For convenience of description, the embodiments of the present application are described with a server as an execution subject unless otherwise specified.

It is to be understood that the described execution body does not constitute a limitation of the embodiments of the present application.

Optionally, the garment design method as shown in fig. 1 may include the following steps:

and step S101, taking the received design request as an input of a pre-trained first model to carry out operation, and obtaining a plurality of clothes design pictures.

Wherein the design request includes a garment type and a garment style parameter. For example, the garment type may be T-shirt, skirt, shorts, etc., and the garment style parameter may be hole, wash, herben style, etc.

In the embodiment of the application, a first model for generating a clothing design picture is trained in advance, and the first model is obtained by training with a known first clothing design picture as an output and a clothing type and a clothing style parameter corresponding to the first clothing design picture as inputs. The known first design picture can capture clothing data aiming at the e-commerce data, the store and the show data through a crawler tool, and uses a classification network to perform tiling/sub-tiling classification on the captured images, and uses a semantic segmentation algorithm to remove background noise aiming at the tiling picture, and then erases patterns and prints on the clothing to obtain a clean standard effect picture, namely the first design picture.

During the garment design, the design request (garment type and garment style parameters) can be quantized, and after the quantization, the design request can be used as the input of the first model for calculation to obtain a plurality of garment design pictures.

In the embodiment of the present application, the first model may be, but is not limited to, a styleGAN model, a BigGAN model, or a Conditional Generative Adaptive Networks (CGAN) model.

And S102, counting the currently popular multiple clothing commodities in the clothing class corresponding to the design request.

Specifically, counting the currently popular multiple types of clothing items in the clothing category corresponding to the design request may include the following steps:

and step S1021, counting the commodity sales volume of the clothing commodities of the clothing class corresponding to the design request in the current time period according to the clothing commodity sales record of the third-party server.

The third-party server can be a server of a clothing commodity sales platform, such as a server of a Taobao platform, a server of a Jingdong platform, and the like. The current time period may be set according to actual conditions, such as 1 month or 1 week before the current time. The clothing category may refer to a clothing type, for example, in the field of clothing design, a skirt may be considered as one category, and a shirt may be considered as another category different from the one corresponding to the skirt, and thus the clothing category corresponding to the design request may refer to the clothing type carried in the design request.

For example, the clothing type carried in the design request is a skirt, the current time period is one week before the current time, and the commodity sales amount of the clothing commodity of the clothing type corresponding to the design request in the current time period is counted.

And step S1022, selecting the multiple clothing commodities with the highest commodity sales amount as the currently popular multiple clothing commodities.

After the commodity sales volume of the clothing commodities of the clothing category corresponding to the design request in the current time period is counted, the clothing commodities of different styles can be sequenced according to the sales volume according to the clothing styles, and then the clothing commodity with the highest commodity sales volume is selected as the current popular multiple clothing commodity.

For example, the top 100 clothing items with the highest sales volume may be selected as the currently popular multiple clothing items, and the clothing items with the top 1% sales volume may also be selected as the currently popular multiple clothing items.

And step S103, extracting common characteristics of the multiple clothes commodities to obtain target characteristics.

In the embodiment of the application, a second model for commodity feature extraction is also pre-established, and the second model can be obtained by training with a known second clothing design picture as input and clothing categories corresponding to the second clothing design picture as output.

When the common features of the multiple types of clothing commodities are extracted, the pictures of the multiple types of clothing commodities can be respectively used as the input of a pre-trained second model for operation, and the feature vectors output by the last two layers of the second model are spliced to obtain a plurality of initial features which are in one-to-one correspondence with the multiple types of clothing commodities. And then extracting the common characteristic vector of the plurality of initial characteristics to obtain the target characteristic, wherein the target characteristic is the common characteristic of the plurality of clothes commodities.

For example, there are 3 items of clothing, which are respectively labeled clothing a, clothing B, and clothing C, where when the picture of clothing a is operated as the input of the second model, the feature vectors output from the last two layers of the second model are a1 and a2, respectively, when the picture of clothing B is operated as the input of the second model, the feature vectors output from the last two layers of the second model are B1 and B2, respectively, and when the picture of clothing C is operated as the input of the second model, the feature vectors output from the last two layers of the second model are C1 and C2, respectively. Then, the eigenvector a1 and a2 are spliced to obtain eigenvector a3, the eigenvector b1 and b2 are spliced to obtain eigenvector b3, the eigenvector c1 and c2 are spliced to obtain eigenvector c3, and then the common eigenvector in the eigenvectors a3, b3 and c3 is extracted to obtain the target characteristic.

In this embodiment of the application, the second garment design picture may be the first garment design picture, or may be a garment design picture obtained according to the obtaining manner of the first garment design picture, which is not specifically limited in this embodiment of the application.

The second model may be, but is not limited to, a styleGAN model, a BigGAN model, or a Conditional Generative Adaptive Networks (CGAN) model, among others.

And step S104, calculating the feature similarity of the corresponding feature of each of the plurality of clothing design pictures and the target feature.

In the embodiment of the present application, when the feature similarity between the feature corresponding to each picture and the target feature is calculated, the similarity between the features may be obtained by calculating a KL (Kullback-Leibler) distance, a JS (Jensen-Shannon) distance, or a cosine distance between the feature corresponding to each picture and the target feature, normalizing the KL (Kullback-Leibler) distance, the JS (JS-Shannon) distance, or the cosine distance, and performing weighting calculation according to weights.

And step S105, sending the at least one clothing design picture with the highest feature similarity between the features and the target features to the client.

After the client receives the sent at least one clothing design picture, clothing design can be carried out according to the at least one clothing design picture.

The clothing design method provided by the embodiment of the application obtains a plurality of clothing design pictures through input operation taking a design request as a first model, counts a plurality of clothing commodities which are popular at present in clothing categories corresponding to the design request, extracts common characteristics of the clothing commodities to obtain target characteristics, and then sends at least one clothing design picture with the highest characteristic similarity between the characteristics and the target characteristics to a client side so as to carry out clothing design according to the at least one clothing design picture, wherein the design request comprises clothing types and clothing style parameters, so that the characteristics of the clothing design picture sent to the client side are matched with the same characteristics of the clothing commodities with the popular at present most when the clothing design picture meets the clothing types and the clothing style parameters, and therefore clothing with specified design key points can be generated through supervision according to requirements, and can be designed for substantially all clothing items on the market. Meanwhile, the generation process of the costume design picture is convenient and fast, and the costume design efficiency is improved.

On the basis of the clothing design method in the first aspect, the embodiment of the application further provides a first possible design for 3D rendering of clothing design pictures. It may include, but is not limited to, the following steps:

step S201, a third model corresponding to the clothing category is called.

In the embodiment of the application, a plurality of third models are pre-established for the clothing design pictures of different categories, the third models are respectively used for determining the plate parameters of the clothing design pictures of different categories, the third models are obtained by training with the known third clothing design pictures as input and the plate parameters of the third clothing design pictures as input. The third model may be, but is not limited to, a styleGAN model, a BigGAN model, or a Conditional Generative Adaptive Networks (CGAN) model, etc.

The sheet parameters may include, but are not limited to, sheet seam point position, sheet vertex position, radian between each point of the sheet, sheet side length, sheet printing position, etc.

And after the server sends the at least one clothing design picture to the client, the server can call a third model corresponding to the clothing category in the design request.

In this embodiment of the application, the third clothing design picture may be the first clothing design picture or the second clothing design picture, or may be a clothing design picture obtained according to the obtaining manner of the first clothing design picture, which is not specifically limited in this embodiment of the application.

Step S202, the selected target clothing design picture in the at least one clothing design picture is used as the input of the third model for operation, and the plate parameter corresponding to the target clothing design picture is obtained.

In the embodiment of the application, after the server sends the at least one clothing design picture to the client, a user at the client can select one or more clothing design pictures as target clothing design pictures according to needs.

And step S203, performing 3D rendering on the target costume design picture according to the plate parameters.

Furthermore, when the target clothing design picture is subjected to 3D rendering, different fabrics can be generated or corresponding fabrics can be generated according to the selection of a user, and then the target clothing design picture is subjected to 3D rendering according to the slab parameters and the generated fabrics.

Or different types of prints can be generated or corresponding prints can be generated according to the selection of the user, and then the target garment design picture is rendered in a 3D mode according to the plate parameters and the generated prints.

Further, before 3D rendering of the target costume design picture, the user can also initiate an adjustment operation for the slab parameters to the server through the client, the server adjusts the slab parameters in response to the adjustment operation for the slab parameters by the client, and the 3D rendering can be performed on the target costume design picture according to the adjusted slab parameters during 3D rendering, so that the standard of commercial generation is achieved.

Therefore, through the first possible design, 3D garment effects of different fabrics or different prints can be generated during 3D rendering so as to guide customers to produce. Meanwhile, the parameters of the plate sheets of the costume design picture can be adjusted, so that the standard of commercial generation can be achieved.

Fig. 2 is a schematic structural diagram of an electronic device according to an embodiment of the present application. Referring to fig. 2, at a hardware level, the electronic device includes a processor, and optionally further includes an internal bus, a network interface, and a memory. The Memory may include a Memory, such as a Random-Access Memory (RAM), and may further include a non-volatile Memory, such as at least 1 disk Memory. Of course, the electronic device may also include hardware required for other services.

The processor, the network interface, and the memory may be connected to each other via an internal bus, which may be an ISA (Industry Standard Architecture) bus, a PCI (Peripheral Component Interconnect) bus, an EISA (Extended Industry Standard Architecture) bus, or the like. The bus may be divided into an address bus, a data bus, a control bus, etc. For ease of illustration, only one double-headed arrow is shown in FIG. 2, but this does not indicate only one bus or one type of bus.

And the memory is used for storing programs. In particular, the program may include program code comprising computer operating instructions. The memory may include both memory and non-volatile storage and provides instructions and data to the processor.

The processor reads the corresponding computer program from the nonvolatile memory into the memory and then runs the computer program to form the clothing design device on the logic level. The processor is used for executing the program stored in the memory and is specifically used for executing the following operations:

The method performed by the clothing design device disclosed in the embodiment of fig. 2 of the present application can be applied to or implemented by a processor. The processor may be an integrated circuit chip having signal processing capabilities. In implementation, the steps of the above method may be performed by integrated logic circuits of hardware in a processor or instructions in the form of software. The Processor may be a general-purpose Processor, including a Central Processing Unit (CPU), a Network Processor (NP), and the like; but also Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components. The various methods, steps and logic blocks disclosed in one or more embodiments of the present application may be implemented or performed. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like. The steps of a method disclosed in connection with one or more embodiments of the present application may be embodied directly in the hardware decoding processor, or in a combination of the hardware and software modules included in the decoding processor. The software module may be located in ram, flash memory, rom, prom, or eprom, registers, etc. storage media as is well known in the art. The storage medium is located in a memory, and a processor reads information in the memory and completes the steps of the method in combination with hardware of the processor.

The electronic device may also execute the method of fig. 1 and implement the functions of the clothing design apparatus in the embodiment shown in fig. 2, which are not described herein again in this application embodiment.

Of course, besides the software implementation, the electronic device of the present application does not exclude other implementations, such as a logic device or a combination of software and hardware, and the like, that is, the execution subject of the following processing flow is not limited to each logic unit, and may also be hardware or a logic device.

Embodiments of the present application also provide a computer-readable storage medium storing one or more programs, where the one or more programs include instructions, which, when executed by an electronic device including a plurality of application programs, enable the electronic device to perform the method of the embodiment shown in fig. 1, and are specifically configured to perform the following operations:

Fig. 3 is a schematic structural diagram of a garment design device according to an embodiment of the present application. Referring to fig. 3, in one software implementation, a garment design apparatus includes:

In short, the above description is only a preferred embodiment of this document, and is not intended to limit the scope of protection of this document. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of this document shall be included in the protection scope of this document.

Computer-readable media, including both non-transitory and non-transitory, removable and non-removable media, may implement information storage by any method or technology. The information may be computer readable instructions, data structures, modules of a program, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), other types of Random Access Memory (RAM), Read Only Memory (ROM), Electrically Erasable Programmable Read Only Memory (EEPROM), flash memory or other memory technology, compact disc read only memory (CD-ROM), Digital Versatile Discs (DVD) or other optical storage, magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information that can be accessed by a computing device. As defined herein, a computer readable medium does not include a transitory computer readable medium such as a modulated data signal and a carrier wave.

All the embodiments in this document are described in a progressive manner, and the same and similar parts among the embodiments can be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.

Claims

1. A garment design method, comprising:

2. The method of claim 1, wherein the counting the plurality of items of clothing currently popular in the category of clothing corresponding to the design request comprises:

3. The method of claim 1, wherein extracting common features of the plurality of items of clothing to obtain target features comprises:

4. The method of claim 1, further comprising:

calling a third model corresponding to the clothing category;

5. The method of claim 4, wherein the 3D rendering of the target garment design picture in accordance with the panel parameters comprises:

6. The method of claim 4, further comprising:

7. The method of claim 1, wherein the first model is a styleGAN model or a BigGAN model.

8. A garment design device, comprising:

9. An electronic device is characterized by comprising a processor, a communication interface, a memory and a communication bus, wherein the processor and the communication interface are used for realizing the communication between the processor and the memory through the bus;

a memory for storing a computer program;

10. A computer-readable storage medium, in which a computer program is stored, which computer program, when being executed by a processor, carries out the following procedure: