A novel class-level weighted partial domain adaptation network for defect detection
Authors: 
Yulong Zhang, Yilin Wang, Zhiqiang Jiang, Li Zheng, Jinshui Chen, Jiangang LuAuthors Info & Claims
Pages 23083 - 23096
Abstract
Recently, unsupervised domain adaptation methods have been increasingly applied to address the domain shift problems in defect detection. However, the effectiveness of most existing methods is based on the identical category space in the source and target domains. A more practical scenario is when the target domain only contains a subset of source categories, i.e., partial domain adaptation, which has not been well-resolved. To this end, a novel class-level weighted partial domain adaptation network (CWPDAN) is proposed for defect detection. Specifically, a hybrid weighting mechanism is derived from a defect classifier and an auxiliary domain classifier. In this case, the weighting mechanism is injected into both the defect classifier and the fine-grained domain adaptation strategy. As such, the shared category space across domains can be aligned well and the outlier categories can be identified and filtered out to alleviate negative transfer. Comprehensive partial domain adaptation experiments verify that the proposed CWPDAN can achieve 95.07% and 98.27% average accuracy on a tire defect dataset and a benchmark dataset, respectively, outperforming other state-of-the-art methods.
References
[1]
Tulbure A-A, Tulbure A-A, and Dulf E-H A review on modern defect detection models using dcnns-deep convolutional neural networks J Adv Res 2022 35 33-48
[2]
Zheng Z, Yang H, Zhou L, Yu B, and Zhang Y Hlu 2-net: A residual u-structure embedded u-net with hybrid loss for tire defect inspection IEEE Trans Instrum Meas 2021 70 1-11
[3]
Rudolph M, Wehrbein T, Rosenhahn B, Wandt B (2022) Fully convolutional cross-scale-flows for image-based defect detection. In Proc IEEE/CVF Winter Conf Appl Comput Vis, pp 1088–1097
[4]
Zheng Y, Cui L (2022) Defect detection on new samples with siamese defect-aware attention network. Appl Intell 1–16
[5]
Zhang W, Li X, Ma H, Luo Z, and Li X Universal domain adaptation in fault diagnostics with hybrid weighted deep adversarial learning IEEE Trans Ind Inform 2021 17 12 7957-7967
[6]
Zhuang F, Qi Z, Duan K, Xi D, Zhu Y, Zhu H, Xiong H, and He Q A comprehensive survey on transfer learning Proc IEEE 2020 109 1 43-76
[7]
Chen Z, Wang C, Wu J, Deng C, Wang Y (2022) Deep convolutional transfer learning-based structural damage detection with domain adaptation. Appl Intell 1–15
[8]
Fu S, Chen J, Chen D, He C (2023) Cnns/vits-cnns/vits: Mutual distillation for unsupervised domain adaptation. Inf Sci 622:83–97
[9]
Marino S, Beauseroy P, and Smolarz A Unsupervised adversarial deep domain adaptation method for potato defects classification Comput Electron Agric 2020 174
[10]
Zhang S, Zhang Q, Gu J, Su L, Li K, and Pecht M Visual inspection of steel surface defects based on domain adaptation and adaptive convolutional neural network Mechanical Syst Signal Process 2021 153
[11]
Gong Y, Luo J, Shao H, and Li Z A transfer learning object detection model for defects detection in x-ray images of spacecraft composite structures Compos Struct 2022 284
[12]
Cao Z, Ma L, Long M, Wang J (2018) Partial adversarial domain adaptation. In Proc Eur Conf Comput Vis (ECCV), pp 135–150
[13]
Zhang J, Ding Z, Li W, Ogunbona P (2018) Importance weighted adversarial nets for partial domain adaptation. In Proc IEEE/CVF Conf Comput Vis Pattern Recognit, pp 8156–8164
[14]
Ren C-X, Ge P, Yang P, and Yan S Learning target-domain-specific classifier for partial domain adaptation IEEE Trans Neural Networks Learn Sys 2020 32 5 1989-2001
[15]
Zhang C and Zhao Q Attention guided for partial domain adaptation Inf Sci 2021 547 860-869
[16]
Li S, Gong K, Xie B, Liu CH, Cao W, Tian S (2022) Critical classes and samples discovering for partial domain adaptation. IEEE Trans Cybern
[17]
Cao Z, You K, Zhang Z, Wang J, Long M (2022) From big to small: Adaptive learning to partial-set domains. IEEE Trans Pattern Anal Mach Intell
[18]
Liu Z-H, Lu B-L, Wei H-L, Chen L, Li X, Wang C-T (2020) A stacked auto-encoder based partial adversarial domain adaptation model for intelligent fault diagnosis of rotating machines. IEEE Trans Ind Inform
[19]
Zhang X, Wang J, Jia S, Han B, Zhang Z (2022) Partial domain adaptation method based on class-weighted alignment for fault diagnosis of rotating machinery. IEEE Trans Instrum Meas
[20]
Li W, Chen Z, and He G A novel weighted adversarial transfer network for partial domain fault diagnosis of machinery IEEE Trans Ind Inform 2020 17 3 1753-1762
[21]
Li S, Liu CH, Lin Q, Wen Q, Su L, Huang G, and Ding Z Deep residual correction network for partial domain adaptation IEEE Trans Pattern Anal Mach Intell 2021 43 7 2329-2344
[22]
Zhu Y, Zhuang F, Wang J, Ke G, Chen J, Bian J, Xiong H, and He Q Deep subdomain adaptation network for image classification IEEE Trans Neural Netw Learn Syst 2020 32 4 1713-1722
[23]
Wang M and Deng W Deep visual domain adaptation: A survey Neurocomputing 2018 312 135-153
[24]
Gretton A, Borgwardt KM, Rasch MJ, Schölkopf B, Smola A (2012) A kernel two-sample test. J Mach Learning Res 13(1):723–773
[25]
Long M, Cao Y, Wang J, Jordan M (2015) Learning transferable features with deep adaptation networks. In Proc 32nd Int Conf Mach Learn, pp 97–105
[26]
Zhang Y, Wang Y, Jiang Z, Zheng L, Chen J, Lu J (2022) Subdomain adaptation network with category isolation strategy for tire defect detection. Measurement 204:112046
[27]
Ganin Y, Ustinova E, Ajakan H, Germain P, Larochelle H, Laviolette F, Marchand M, and Lempitsky V Domain-adversarial training of neural networks J Mach Learn Res 2016 17 1 2096-2030
[28]
Chen G, Guo Y, Huang M, Gong D, Yu Z (2022) A domain adaptation learning strategy for dynamic multiobjective optimization. Inf Sci
[29]
Li J, Lü S, and Li Z Unsupervised domain adaptation via softmax-based prototype construction and adaptation Inf Sci 2022 609 257-275
[30]
Sharma A, Kalluri T, Chandraker M (2021) Instance level affinity-based transfer for unsupervised domain adaptation. In Proc IEEE/CVF Conf Comput Vis Pattern Recognit. pp 5361–5371
[31]
Chai Z, Zhao C, Huang B (2021) Multisource-refined transfer network for industrial fault diagnosis under domain and category inconsistencies. IEEE Trans Cybern
[32]
He C, Tan T, Fan X, Zheng L, Ye Z (2022) Noise-residual mixup for unsupervised adversarial domain adaptation. Appl Intell 1–14
[33]
Gholenji E and Tahmoresnezhad J Joint discriminative subspace and distribution adaptation for unsupervised domain adaptation Appl Intell 2020 50 7 2050-2066
[34]
Chen P, Zhao R, He T, Wei K, Yang Q (2022) Unsupervised domain adaptation of bearing fault diagnosis based on join sliced wasserstein distance. ISA Trans
[35]
Goodfellow I, Pouget-Abadie J, Mirza M, Xu B, Warde-Farley D, Ozair S, Courville A, and Bengio Y Generative adversarial nets In Proc Adv Neural Inf Process Syst 2014 27 2672-2680
[36]
Jiao J, Zhao M, Lin J, and Ding C Classifier inconsistency-based domain adaptation network for partial transfer intelligent diagnosis IEEE Trans Ind Inform 2019 16 9 5965-5974
[37]
Wang B, Wang S, Zhang Z, Zhao X, Fu Z (2022) Decomposed-distance weighted optimal transport for unsupervised domain adaptation. Appl Intell 1–15
[38]
Li X and Zhang W Deep learning-based partial domain adaptation method on intelligent machinery fault diagnostics IEEE Trans Ind Electron 2020 68 5 4351-4361
[39]
Saenko K, Kulis B, Fritz M, Darrell T (2010) Adapting visual category models to new domains. In Lect Notes Comput Sci pp 213–226 . Springer
[40]
He K, Zhang X, Ren S, Sun J (2016) Deep residual learning for image recognition. In Proc IEEE Conf Comput Vis Pattern Recognit, pp 770–778
[41]
Liu Z, Lin Y, Cao Y, Hu H, Wei Y, Zhang Z, Lin S, Guo B (2021) Swin transformer: Hierarchical vision transformer using shifted windows. In Proc IEEE Int Conf Comput Vision, pp 10012–10022
[42]
Xu R, Li G, Yang J, Lin L (2019) Larger norm more transferable: An adaptive feature norm approach for unsupervised domain adaptation. In Proc IEEE Int Conf Comput Vision, pp 1426–1435
[43]
Guo P, Zhu J, Zhang Y (2022) Selective partial domain adaptation. In Br Mach Vis Conf
[44]
Guo L, Lei Y, Xing S, Yan T, and Li N Deep convolutional transfer learning network: A new method for intelligent fault diagnosis of machines with unlabeled data IEEE Trans Ind Electron 2018 66 9 7316-7325
[45]
Chen X, Hsieh C-J, Gong B (2022) When vision transformers outperform resnets without pre-training or strong data augmentations. In Proc Int Conf Learn Represent
[46]
Van der Maaten L, Hinton G (2008) Visualizing data using t-sne. J Mach Learn Res 9(11):2579–2625
[47]
Hinton GE, Roweis S (2002) Stochastic neighbor embedding. Adv Neural Inf Process Syst Proc Conf 15
Recommendations
Similarity-based domain adaptation network
AbstractDomain adaptation utilizes labeled source domains to solve classification problems in the unlabeled target domain. Previous domain adaptation methods consider global domain adaptation while neglecting class-wise information, thus ...
Domain adaptation via Transferable Swin Transformer for tire defect detection
AbstractIn data-driven tire defect detection, scarce defect samples and the poor transferability of models make the defect detection a time-consuming and expensive process. In recent years, domain adaptation methods have been applied to solve the cross-...
Highlights- A novel Transferable Swin Transformer is proposed for transfer tire defect detection.
- A region-level weighting mechanism is designed for the subdomain adaptation.
- The proposed method performs well on the tire defect dataset and a ...
Deep Joint Semantic Adaptation Network for Multi-source Unsupervised Domain Adaptation
AbstractMulti-source Unsupervised Domain Adaptation (MUDA) transfers knowledge learned from multiple labeled source domains to an unlabeled target domain by minimizing the domain shift between multiple source domains and the target domain. Recent studies ...
Highlights- We propose a simple and effective multi-source unsupervised domain adaptation method for classification.
- We propose a metric to align the cross-domain joint distribution of category-corresponded subdomains on multiple task-specific ...
Comments
Please enable JavaScript to view thecomments powered by Disqus.Information & Contributors
Information
Published In
© The Author(s), under exclusive licence to Springer Science+Business Media, LLC, part of Springer Nature 2023. Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law.
Publisher
Kluwer Academic Publishers
United States
Publication History
Published: 05 July 2023
Accepted: 25 May 2023
Author Tags
Qualifiers
- Research-article
Funding Sources
- National Key Research and Development Plan of China
- National Natural Science Foundation of China
- Major Scientific Project of Zhejiang Laboratory
- Zhejiang University Robotics Institute (Yuyao) Project
Contributors
Other Metrics
Bibliometrics & Citations
Bibliometrics
Article Metrics
- 0Total Citations
- 0Total Downloads
- Downloads (Last 12 months)0
- Downloads (Last 6 weeks)0
Reflects downloads up to 09 Jan 2025