Fashionable garment image generation method based on garment category and texture pattern control

Abstract

The invention discloses a fashion clothing image generation method based on clothing category and texture pattern control. According to the invention, the method comprises the steps: constructing a brand-new data set, and carrying out innovation on a network structure and validity verification; enabling a reference designer to draw a fashion costume design sketch in the costume design process, and then selecting corresponding texture pattern cloth to be applied to the design sketch to complete costume design; dividing the whole task into two parts of fashion costume design sketch generation based on costume category control and fashion costume image generation based on texture pattern and design sketch control, wherein the two parts are trained separately, and after the training of the two parts is completed, the output of the first part is used as the input of the fashion design sketch of the second part, so that two networks can be spliced. According to the invention, the generation quality of fashion clothes is improved, the texture extension module is added to optimize the generation of textures, and the gradient reconstruction loss is increased to make the generated fashion clothes picture clearer.

Description

technical field [0001] The invention proposes a fashion clothing image generation method based on clothing category and texture pattern control, and mainly relates to the field of fashion clothing image generation. Background technique [0002] Image generation based on deep learning refers to a computer technology that generates virtual digital images through deep learning technology. Using this technology, a large number of virtual pictures similar to real pictures can be quickly generated. To improve the availability of image generation techniques, more research is now focusing on conditional image generation techniques. In the conditional image generation network, the user can input some relevant conditions to control the generation of the image, which allows the user to have certain control over the generation of the image. [0003] The fashion clothing image generation method is also based on conditional image generation technology. Users can generate fashion clothing...

AI Technical Summary

Problems solved by technology

However, these methods also have some disadvantages

First of all, the control conditions selected by some methods are difficult to obtain, such as fashion clothing sketches. Drawing fashion clothing sketches requires a very professional knowledge base and takes a lot of time

Secondly, some existing generation methods based on deep learning are not particularly effective in generating clothing textures

Images