A novel defect generation model based on two-stage GAN

The fabric defect models based on deep learning often demand numerous training samples to achieve high accuracy. However, obtaining a complete dataset containing all possible fabric textures and defects is a big challenge due to the sophisticated and various fabric textures and defect forms. This st...

Full description

Saved in:
Bibliographic Details
Published ine-Polymers Vol. 22; no. 1; pp. 793 - 802
Main Authors Zhang, Yuming, Gao, Zhongyuan, Zhi, Chao, Chen, Mengqi, Zhou, Youyong, Wang, Shuai, Fu, Sida, Yu, Lingjie
Format Journal Article
LanguageEnglish
Published Berlin De Gruyter 15.09.2022
Walter de Gruyter GmbH
Subjects
Accuracy
Datasets
date enhancement
Deep learning
Defects
Engineering
fabric defect detection
Game theory
generative adversarial network
Genetic algorithms
Laboratories
Methods
Neural networks
Performance evaluation
Radiation
Training
Online AccessGet full text

Cover

More Information
Summary:The fabric defect models based on deep learning often demand numerous training samples to achieve high accuracy. However, obtaining a complete dataset containing all possible fabric textures and defects is a big challenge due to the sophisticated and various fabric textures and defect forms. This study created a two-stage deep pix2pixGAN network called Dual Deep pix2pixGAN Network (DPGAN) to address the above problem. The defect generation model was trained based on the DPGAN network to automatically “transfer” defects from defected fabric images to clean, defect-free fabric images, thus strengthening the training data. To evaluate the effectiveness of the defect generation model, extensive comparative experiments were conducted to assess the performance of the fabric defect detection before and after data enhancement. The results indicate that the detection accuracy was improved regarding the belt_yarn, hole, and stain defect.
ISSN:1618-7229
2197-4586
1618-7229
DOI:10.1515/epoly-2022-0071