Unlock instant, AI-driven research and patent intelligence for your innovation.

Image generation method of improved GAN model

An image generation and model technology, applied in the field of image processing, can solve problems such as model difficulty in training and model collapse, and achieve cost-saving effects

Active Publication Date: 2020-04-17
HUBEI UNIV OF TECH
View PDF6 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] At present, the GAN model is widely used in image generation, but there are problems of model collapse and model difficulty in training

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Image generation method of improved GAN model
  • Image generation method of improved GAN model
  • Image generation method of improved GAN model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0022] In order to facilitate those of ordinary skill in the art to understand and implement the present invention, the present invention will be described in further detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the implementation examples described here are only used to illustrate and explain the present invention, and are not intended to limit this invention.

[0023] please see figure 1 , a kind of image generation method of improving GAN model provided by the invention, comprises the following steps:

[0024] Step 1: Read 5,000 original image samples of local images, including two categories. Take crayfish as an example: one category is bads (black shrimp and damaged shrimp, etc.), and the other category is goods (good Lobster), the corresponding labels are 0 and 1 respectively;

[0025] Step 2: Take 5000 samples as the input of the improved LeNet model, and train the model parameters by the method of convolutio...

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

No PUM Login to View More

Abstract

The invention discloses an image generation method of an improved GAN model. The method comprises the following steps: firstly, reading a local existing original image sample; pre-training an improvedLeNet model; taking the improved LeNet model as a discrimination model of the improved GAN model; designing a generation model of the improved GAN model according to the generation model of the original GAN model; taking random noise as input of a generation model, and obtaining a generated image sample after deconvolution operation; taking the generated image sample and the existing original image sample as the input of the discrimination model to obtain the probability that the picture is true, and returning the result to the discrimination model in the step 3 and the generation model in the step 4; judging whether the model updates parameters to better distinguish the authenticity of the picture, and generating a more real picture by the generation model; wherein the generation model and the discrimination model compete with each other and promote each other, training is finished after N steps, and finally the generation model generates a real sample. According to the method, the cost can be saved for manufacturers, and more samples can be provided for subsequent classification problems.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to an image generation method, in particular to an image generation method for improving a GAN model. Background technique [0002] When dealing with the classification problem on the production line, the number of samples provided by the manufacturer is limited. In order to increase the characteristics of the samples and improve the stability of the classification model, it is necessary to generate pictures based on the existing samples. [0003] At present, the GAN model is widely used in image generation, but there are problems of model collapse and difficult training of the model. Contents of the invention [0004] In order to solve the above technical problems, the present invention combines the improved LeNet model and the GAN model to provide an image generation method for the improved GAN model. [0005] The technical scheme adopted in the present invention is: a k...

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
no application Login to View More
IPC IPC(8): G06K9/62G06N3/04G06N3/08
CPCG06N3/08G06N3/045G06F18/241
Inventor 舒军李振亚杨露吴柯蒋明威邓明舟舒心怡潘健王淑青
Owner HUBEI UNIV OF TECH
Features
  • R&D
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic, Popular Technical Reports.

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com