Eternal-MAML: a meta-learning framework for cross-domain defect recognition

Defect recognition tasks for industrial product suffer from a serious lack of samples, greatly limiting the generalizability of deep learning models. Addressing the imbalance of defective samples often involves leveraging pre-trained models for transfer learning. However, when these models, pre-trai...

Full description

Saved in:
Bibliographic Details
Main Authors: Jipeng Feng, Haigang Zhang, Zhifeng Wang
Format: Article
Language:English
Published: PeerJ Inc. 2025-05-01
Series:PeerJ Computer Science
Subjects:
Online Access:https://peerj.com/articles/cs-2757.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850143779874406400
author Jipeng Feng
Haigang Zhang
Zhifeng Wang
author_facet Jipeng Feng
Haigang Zhang
Zhifeng Wang
author_sort Jipeng Feng
collection DOAJ
description Defect recognition tasks for industrial product suffer from a serious lack of samples, greatly limiting the generalizability of deep learning models. Addressing the imbalance of defective samples often involves leveraging pre-trained models for transfer learning. However, when these models, pre-trained on natural image datasets, are transferred to pixel-level defect recognition tasks, they frequently suffer from overfitting due to data scarcity. Furthermore, significant variations in the morphology, texture, and underlying causes of defects across different industrial products often lead to a degradation in performance, or even complete failure, when directly transferring a defect classification model trained on one type of product to another. The Model-Agnostic Meta-Learning (MAML) framework can learn a general representation of defects from multiple industrial defect recognition tasks and build a foundational model. Despite lacking sufficient training data, the MAML framework can still achieve effective knowledge transfer among cross-domain tasks. We noticed there exists serious label arrangement issues in MAML because of the random selection of recognition tasks, which seriously affects the performance of MAML model during both training and testing phase. This article proposes a novel MAML framework, termed as Eternal-MAML, which guides the update of the classifier module by learning a meta-vector that shares commonality across batch tasks in the inner loop, and addresses the overfitting phenomenon caused by label arrangement issues in testing phase for vanilla MAML. Additionally, the feature extractor in this framework combines the advantages of the Squeeze-and-Excitation module and Residual block to enhance training stability and improve the generalization accuracy of model transfer with the learned initialization parameters. In the simulation experiments, several datasets are applied to verified the cross-domain meta-learning performance of the proposed Eternal-MAML framework. The experimental results show that the proposed framework outperforms the state-of-the-art baselines in terms of average normalized accuracy. Finally, the ablation studies are conducted to examine how the primary components of the framework affect its overall performance. Code is available at https://github.com/zhg-SZPT/Eternal-MAML.
format Article
id doaj-art-88de8d6a27104bce9cb41e90bc8f83a5
institution OA Journals
issn 2376-5992
language English
publishDate 2025-05-01
publisher PeerJ Inc.
record_format Article
series PeerJ Computer Science
spelling doaj-art-88de8d6a27104bce9cb41e90bc8f83a52025-08-20T02:28:36ZengPeerJ Inc.PeerJ Computer Science2376-59922025-05-0111e275710.7717/peerj-cs.2757Eternal-MAML: a meta-learning framework for cross-domain defect recognitionJipeng Feng0Haigang Zhang1Zhifeng Wang2School of Computer Science and Software Engineering, University of Science and Technology Liaoning, Anshan, ChinaInstitute of Applied Artificial Intelligence of the Guangdong-Hong Kong-Macao Greater Bay Area, Shenzhen Polytechnic University, Shenzhen, ChinaSchool of Computer Science and Software Engineering, University of Science and Technology Liaoning, Anshan, ChinaDefect recognition tasks for industrial product suffer from a serious lack of samples, greatly limiting the generalizability of deep learning models. Addressing the imbalance of defective samples often involves leveraging pre-trained models for transfer learning. However, when these models, pre-trained on natural image datasets, are transferred to pixel-level defect recognition tasks, they frequently suffer from overfitting due to data scarcity. Furthermore, significant variations in the morphology, texture, and underlying causes of defects across different industrial products often lead to a degradation in performance, or even complete failure, when directly transferring a defect classification model trained on one type of product to another. The Model-Agnostic Meta-Learning (MAML) framework can learn a general representation of defects from multiple industrial defect recognition tasks and build a foundational model. Despite lacking sufficient training data, the MAML framework can still achieve effective knowledge transfer among cross-domain tasks. We noticed there exists serious label arrangement issues in MAML because of the random selection of recognition tasks, which seriously affects the performance of MAML model during both training and testing phase. This article proposes a novel MAML framework, termed as Eternal-MAML, which guides the update of the classifier module by learning a meta-vector that shares commonality across batch tasks in the inner loop, and addresses the overfitting phenomenon caused by label arrangement issues in testing phase for vanilla MAML. Additionally, the feature extractor in this framework combines the advantages of the Squeeze-and-Excitation module and Residual block to enhance training stability and improve the generalization accuracy of model transfer with the learned initialization parameters. In the simulation experiments, several datasets are applied to verified the cross-domain meta-learning performance of the proposed Eternal-MAML framework. The experimental results show that the proposed framework outperforms the state-of-the-art baselines in terms of average normalized accuracy. Finally, the ablation studies are conducted to examine how the primary components of the framework affect its overall performance. Code is available at https://github.com/zhg-SZPT/Eternal-MAML.https://peerj.com/articles/cs-2757.pdfComputer visionModel-agnostic meta-learningIndustrial visual detection
spellingShingle Jipeng Feng
Haigang Zhang
Zhifeng Wang
Eternal-MAML: a meta-learning framework for cross-domain defect recognition
PeerJ Computer Science
Computer vision
Model-agnostic meta-learning
Industrial visual detection
title Eternal-MAML: a meta-learning framework for cross-domain defect recognition
title_full Eternal-MAML: a meta-learning framework for cross-domain defect recognition
title_fullStr Eternal-MAML: a meta-learning framework for cross-domain defect recognition
title_full_unstemmed Eternal-MAML: a meta-learning framework for cross-domain defect recognition
title_short Eternal-MAML: a meta-learning framework for cross-domain defect recognition
title_sort eternal maml a meta learning framework for cross domain defect recognition
topic Computer vision
Model-agnostic meta-learning
Industrial visual detection
url https://peerj.com/articles/cs-2757.pdf
work_keys_str_mv AT jipengfeng eternalmamlametalearningframeworkforcrossdomaindefectrecognition
AT haigangzhang eternalmamlametalearningframeworkforcrossdomaindefectrecognition
AT zhifengwang eternalmamlametalearningframeworkforcrossdomaindefectrecognition