An image generation method based on improved cycle GAN
An image generation and image technology, applied in the field of image processing, to achieve good processing effects, improve background distortion, and achieve realistic effects
Active Publication Date: 2022-08-02
TIANJIN UNIV
View PDF10 Cites 0 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] The purpose of the present invention is to provide an image generation method based on the improved Cycle GAN, which introduces the feature fusion mechanism into the Cycle GAN network. background distortion problem
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment Construction
[0034] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, but not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.
[0035] like Image 6 As shown, an embodiment provided by the present invention, an image generation method based on improved Cycle GAN, includes:
[0036] Get the input image from the X-domain image database;
[0037] inputting the input image to an encoder and outputting a feature image;
[0038] Inputting the feature image into the feature weight adaptive module, extracting background information and target feature information fro...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses an image generation method based on an improved Cycle GAN, comprising: acquiring an input image from an X-domain image database; inputting the input image into an encoder and outputting a feature image; inputting the feature image into a feature weight automatic The adaptation module extracts background information and target feature information from the feature image, and performs feature fusion with different weights on the extracted background information and target feature information; the processed feature images are sequentially input into the converter and the decoder to restore and generate output image. The image generation method of the present invention introduces the feature fusion mechanism into the Cycle GAN network, the image effect generated after improvement is more realistic, the feature detail processing effect is better, and the background distortion problem after the original network conversion is improved, so that the converted image is in the structure, The brightness and color are closer to the image in the real scene.
Description
technical field [0001] The invention relates to the technical field of image processing, in particular to an image generation method based on an improved Cycle GAN. Background technique [0002] Cycle GAN is a well-known algorithm that applies generative adversarial network to unsupervised image-to-image translation. Its biggest feature is unsupervised, and it only needs to provide images of different domains to successfully train images between different domains. The mapping solves the problem that paired data is not easy to obtain. like figure 1 shown, Cycle GAN learns and Two maps, the discriminator To determine whether the generated image is a real image in the Y domain, the discriminator Determine whether the generated image is a real image in the X domain. The samples in the domain are generated by the generator to the samples in the domain , Then generate samples through generator F , through network optimization to make as close as possible to the r...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & Authority Patents(China)
IPC IPC(8): G06T5/00G06T11/00G06K9/62G06N3/04G06N3/08G06V10/80G06V10/82
CPCG06T11/001G06N3/084G06T2207/10024G06T2207/20081G06T2207/20084G06N3/045G06F18/25G06T5/00
Inventor 侯永宏侯春羽李斌朱新山李施琦屈璐瑶曾筠婷李亚霖钱统玉
Owner TIANJIN UNIV