An Unsupervised-Learning-Based Approach for Automated Defect Inspection on Textured Surfaces
The DOI matches its Crossref record; 50 of the 51 references it cites were found in a registry or an index.
Abstract
Automated defect inspection has long been a challenging task especially in industrial applications, where collecting and labeling large amounts of defective samples are usually harsh and impracticable. In this paper, we propose an approach to detect and localize defects with only defect-free samples for model training. This approach is carried out by reconstructing image patches with convolutional denoising autoencoder networks at different Gaussian pyramid levels, and synthesizing detection results from these different resolution channels. Reconstruction residuals of the training patches are used as the indicator for direct pixelwise defect prediction, and the reconstruction residual map generated in each channel is combined to generate the final inspection result. This novel method has two prominent characteristics, which benefit the implementation of automatic defect inspection in practice. First, it is absolutely unsupervised that no human intervention is needed throughout the inspection process. Second, multimodal strategy is utilized in this method to synthesize results from multiple pyramid levels. This strategy is capable of improving the robustness and accuracy of the method. To evaluate this approach, experiments on convergence, noise immunity, and defect inspection accuracy are conducted. Furthermore, comparative tests with some excellent algorithms on actual and simulated data sets are performed. Experimental results demonstrated the effectiveness and superiority of the proposed method on homogeneous and nonregular textured surfaces.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Moreover, AE-based algorithms also demonstrate strong competitiveness in steel surface defect detection, which are reported to be fairly noise-robust.”
From Automated Visual Defect Detection for Flat Steel Surface: A Survey (2020) · cites this paper as [114]
In the citing paper
“Mei et al. [114] utilized convolutional denoising AE network to reconstruct image patches, combined with the reconstruction residual maps, this scheme can reliably learning final detection results, where no manual intervention is needed throughout all the detection process.”
From Automated Visual Defect Detection for Flat Steel Surface: A Survey (2020) · cites this paper as [114]
Citations
Cites
Scikit-learn: Machine Learning in Python
Pedregosa, F., Varoquaux, G., Gramfort, A., Michel, V., Thirion, B., Grisel, O., Blondel, M., Müller, A., Nothman, J., Louppe, G., et al. (2012) · arXiv · doi:10.48550/arxiv.1201.0490 · Cited by 71,016 · cited 1×
“The packages involved in the implementation mainly include Keras [47], Theano [48], Scikit-Learn [49], Numpy [50], and Scipy [51].”
MethodThe NumPy Array: A Structure for Efficient Numerical Computation
van der Walt, S., Colbert, S. C., & Varoquaux, G. (2011) · Computing in Science & Engineering · doi:10.1109/mcse.2011.37 · Cited by 9,381 · cited 1×
“The packages involved in the implementation mainly include Keras [47], Theano [48], Scikit-Learn [49], Numpy [50], and Scipy [51].”
Method(2010) · Journal of Machine Learning Research · doi:10.5555/1756006.1953039 · Cited by 4,051 · cited 1×
“This is done by first corrupting the initial input x into x̃ by means of a stochastic mapping x̃ ∼ q(x̃|x) [29].”
MethodPractical Recommendations for Gradient-Based Training of Deep Architectures
Bengio, Y. (2012) · Lecture Notes in Computer Science · doi:10.1007/978-3-642-35289-8_26 · Cited by 2,259 · cited 1×
“As for the learning rate α, a heuristic principle is applied by choosing a value close to the largest learning rate that does not cause divergence [42].”
MethodStatic Analysis of Shape in TensorFlow Programs
The Theano Development Team, Al‐Rfou, R., Alain, G., Almahairi, A., Angermueller, C., Bahdanau, D., Ballas, N., Bastien, F., Bayer, J., Belikov, A., et al. (2016) · arXiv (Cornell University) · doi:10.4230/lipics.ecoop.2020.15 · Cited by 1,589 · cited 1×
“The packages involved in the implementation mainly include Keras [47], Theano [48], Scikit-Learn [49], Numpy [50], and Scipy [51].”
MethodRefaeilzadeh, P., Tang, L., & Liu, H. (2009) · Encyclopedia of Database Systems · doi:10.1007/978-0-387-39940-9_565 · Cited by 1,549
Adaptive deconvolutional networks for mid and high level feature learning
Zeiler, M. D., Taylor, G. W., & Fergus, R. (2011) · 2011 International Conference on Computer Vision · doi:10.1109/iccv.2011.6126474 · Cited by 1,217
Salt-and-pepper noise removal by median-type noise detectors and detail-preserving regularization
Chan, R. H., Chung-Wa, & Nikolova, M. (2005) · IEEE Transactions on Image Processing · doi:10.1109/tip.2005.852196 · Cited by 1,152 · cited 1×
“c) Noise corruption: As stated, the salt and pepper noise [34] is utilized for data corruption in the MSCDAE”
MethodWLD: A Robust Local Image Descriptor
Chen, J., Shan, S., He, C., Zhao, G., Pietikäinen, M., Chen, X., & Gao, W. (2010) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2009.155 · Cited by 1,024
Weimer, D., Scholz-Reiter, B., & Shpitalni, M. (2016) · CIRP Annals · doi:10.1016/j.cirp.2016.04.072 · Cited by 577 · cited 1×
“Weimer et al. [23] and Park et al. [24] also proposed novel deep CNN architectures, which obviously improve the accuracy of automated defect detection.”
Background
Research works printed but not linked: 4
- mario fritz, 2017, The KTH-TIPS and KTH-TIPS2 Image Databases
- jones, 2014, SciPy Open source scientific tools for Python
- al-rfou, 2016, Theano A Python framework for fast computation of mathematical expressions
- chollet, 2017, Keras
Cited by
State of the Art in Defect Detection Based on Machine Vision
Ren, Z., Fang, F., Yan, N., & Wu, Y. (2022) · International Journal of Precision Engineering and Manufacturing-Green Technology · doi:10.1007/s40684-021-00343-6 · Cited by 799
Automated Visual Defect Detection for Flat Steel Surface: A Survey
Luo, Q., Fang, X., Liu, L., Yang, C., & Sun, Y. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2019.2963555 · Cited by 559
“[114] utilized convolutional denoising AE network to reconstruct image patches, combined with the reconstruction residual maps, this scheme can reliably learning final detection results, where no manual intervention is needed throughout all the detection process.”
BackgroundMethodSurface Defect Detection Methods for Industrial Products: A Review
Chen, Y., Ding, Y., Zhao, F., Zhang, E., Wu, Z., & Shao, L. (2021) · Applied Sciences · doi:10.3390/app11167657 · Cited by 457
“Literature [61] proposed a multi-scale convolutional denoising autoencoder (MSCDAE), which uses a multi-modal strategy to synthesize the results of multiple pyramid levels and tests on LCD panels, tiles, and textiles; the experiment proves that this method obtains high accuracy and robustness.”
MethodVisual-Based Defect Detection and Classification Approaches for Industrial Applications—A SURVEY
Czimmermann, T., Ciuti, G., Milazzo, M., Chiurazzi, M., Roccella, S., Oddo, C. M., & Dario, P. (2020) · Sensors · doi:10.3390/s20051459 · Cited by 441
Image-Based Surface Defect Detection Using Deep Learning: A Review
Bhatt, P. M., Malhan, R. K., Rajendran, P., Shah, B. C., Thakar, S., Yoon, Y. J., & Gupta, S. K. (2021) · Journal of Computing and Information Science in Engineering · doi:10.1115/1.4049535 · Cited by 362
Defect Detection Methods for Industrial Products Using Deep Learning Techniques: A Review
Saberironaghi, A., Ren, J., & El-Gindy, M. (2023) · Algorithms · doi:10.3390/a16020095 · Cited by 317
“[48] Multi-scale convolutional denoising autoencoder High precision and robustness by combining results from multiple pyramid levels; can effectively remove noise from the input data, which can improve the performance of defect detection in noisy images May not be able to generalize well to new unseen data, especially…”
BackgroundEbayyeh, A. A. R. M. A., & Mousavi, A. (2020) · IEEE Access · doi:10.1109/access.2020.3029127 · Cited by 310
“in [250] proposed an algorithm to detect texture surface defects in general such as LCD panels, ceramic tiles, and textiles.”
BackgroundMethodIndustrial Internet of Things Intelligence Empowering Smart Manufacturing: A Literature Review
Hu, Y., Jia, Q., Yao, Y., Lee, Y., Lee, M., Wang, C., Zhou, X., Xie, R., & Yu, F. R. (2024) · IEEE Internet of Things Journal · doi:10.1109/jiot.2024.3367692 · Cited by 256
“A denoise autoencoder networks is used to rebuild the defect free image, thus the reconstruction residuals of the training patches are used as the indicator for defect prediction [164].”
MethodA Review on Recent Advances in Vision-based Defect Recognition towards Industrial Intelligence
Gao, Y., Li, X., Wang, X. V., Wang, L., & Gao, L. (2022) · Journal of Manufacturing Systems · doi:10.1016/j.jmsy.2021.05.008 · Cited by 253
Deep Learning for Unsupervised Anomaly Localization in Industrial Images: A Survey
Tao, X., Gong, X., Zhang, X., Yan, S., & Adak, C. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3196436 · Cited by 228
“The earliest defect images are generated by adding random noise; for instance, Nakazawa et al. [93] used synthesized noisy images for AD in wafer images, and the multi-scale AE method proposed by Mei et al. [37] also adopted the similar synthesized noisy images.”
Unsupervised anomaly segmentation via deep feature reconstruction
Shi, Y., Yang, J., & Qi, Z. (2021) · Neurocomputing · doi:10.1016/j.neucom.2020.11.018 · Cited by 224
“A general practice to tackle the problem is by multi-scale modeling [8], [10], [20], [23].”
InfluentialSteel Surface Defect Recognition: A Survey
Wen, X., Shan, J., He, Y., & Song, K. (2022) · Coatings · doi:10.3390/coatings13010017 · Cited by 210
Yang, H., Chen, Y., Song, K., & Yin, Z. (2019) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2018.2886031 · Cited by 176
“MSCDAE [5] achieves a reasonable Recall but relatively low Precision.”
BackgroundMethodSelf-citationUnsupervised fabric defect detection based on a deep convolutional generative adversarial network
Hu, G., Huang, J., Wang, Q., Li, J., Xu, Z., & Huang, X. (2020) · Textile Research Journal · doi:10.1177/0040517519862880 · Cited by 167
SDDNet: A Fast and Accurate Network for Surface Defect Detection
Cui, L., Jiang, X., Xu, M., Li, W., Lv, P., & Zhou, B. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3056744 · Cited by 165
“The supervised detectors [17] often involve pixel-level ground truth for training, while the results of the unsupervised algorithms [18] are particularly sensitive to the segmentation threshold, which affects the detection robustness.”
BackgroundMethodLee, Y., & Kang, P. (2022) · IEEE Access · doi:10.1109/access.2022.3171559 · Cited by 122
“In addition, multi-scale encoding and denoised CAE are used (Mei et al., 2018), or a latent vector is extracted by adding a memory module (Gong et al., 2019) to construct a representation that reflects features of the normal data well.”
BackgroundMethodA smart surface inspection system using faster R-CNN in cloud-edge computing environment
Wang, Y., Liu, M., Zheng, P., Yang, H., & Zou, J. T. (2020) · Advanced Engineering Informatics · doi:10.1016/j.aei.2020.101037 · Cited by 156
Self-citationZhao, Z., Qi, H., Qi, Y., Zhang, K., Zhai, Y., & Zhao, W. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.2969057 · Cited by 137
Unsupervised Anomaly Detection for Surface Defects With Dual-Siamese Network
Tao, X., Zhang, D., Ma, W., Hou, Z., Lu, Z., & Adak, C. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3142326 · Cited by 135
“The image reconstruction-based methods have recently shown great promise for anomaly segmentation, which mainly adopt the deep autoencoder (AE) [4], [5], [14], [15], [25].”
BackgroundMethodResultInfluentialExploring plain ViT features for multi-class unsupervised visual anomaly detection
Zhang, J., Chen, X., Wang, Y., Wang, C., Liu, Y., Li, X., Yang, M., & Tao, D. (2025) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2025.104308 · Cited by 134
Lee, Y., & Kang, P. (2022) · arXiv · doi:10.48550/arxiv.2203.10808 · Cited by 123
“In addition, multi-scale encoding and denoised CAE are used (Mei et al., 2018), or a latent vector is extracted by adding a memory module (Gong et al., 2019) to construct a representation that reflects features of the normal data well.”
BackgroundMethodTiny Defect Detection in High-Resolution Aero-Engine Blade Images via a Coarse-to-Fine Framework
Li, D., Li, Y., Xie, Q., Wu, Y., Yu, Z., & Wang, J. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3062175 · Cited by 121
“In recent years, the techniques [5]–[10] based on computer vision (CV) and deep learning (DL) have been developed rapidly, which are applied in many fields for defect detection [11]–[14], [15].”
BackgroundCross Position Aggregation Network for Few-Shot Strip Steel Surface Defect Segmentation
Feng, H., Song, K., Cui, W., Zhang, Y., & Yan, Y. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3246519 · Cited by 115
“works [24], [25] introduce unsupervised-learning methods to”
Multi-Scale Patch-Based Representation Learning for Image Anomaly Detection and Segmentation
Tsai, C.-C., Wu, T.-H., & Lai, S.-H. (2022) · 2022 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV) · doi:10.1109/wacv51458.2022.00312 · Cited by 72
“Previous deep learning methods for anomaly detection and segmentation are usually based on reconstructionbased neural network architectures, such as autoencoders (AE) [3], [5], [11], [15], [17], variational autoencoders (VAE) [14], [23], and generative adversarial networks (GAN) [1], [20].”
Wen, G., Gao, Z., Cai, Q., Wang, Y., & Mei, S. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.3007292 · Cited by 99
“The traditional vision methods for surface defect inspection can be further categorized into statistical methods, structural methods, filter-based methods, and model-based methods [4].”
BackgroundMethodInfluentialSelf-citationGaussian Anomaly Detection by Modeling the Distribution of Normal Data in Pretrained Deep Features
Rippel, O., Mertens, P., Konig, E., & Merhof, D. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3098381 · Cited by 82
ZJU-Leaper: A Benchmark Dataset for Fabric Defect Detection and a Comparative Study
Zhang, C., Feng, S., Wang, X., & Wang, Y. (2020) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2021.3057027 · Cited by 90
“Basic auto-encoders [6]–[8] are used to regenerate an input after its passing through certain bottleneck layers.”
BackgroundMethodInfluentialSchlosser, T., Friedrich, M., Beuth, F., & Kowerko, D. (2022) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-021-01906-9 · Cited by 88
A Comprehensive Survey on Machine Learning Driven Material Defect Detection
Bai, J., Wu, D., Shelley, T., Schubel, P. J., Twine, D., Russell, J., Zeng, X., & Zhang, J. (2025) · ACM Computing Surveys · doi:10.1145/3730576 · Cited by 82
“By reducing the reliance on human inspection, these technologies diminish the impact of human error, contributing to more reliable and uniform inspection results.”
A Real-Time Steel Surface Defect Detection Approach With High Accuracy
Wang, W., Mi, C., Wu, Z., Lu, K., Long, H., Pan, B., Li, D., Zhang, J., Chen, P., & Wang, B. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3127648 · Cited by 77
Visual Anomaly Detection for Images: A Systematic Survey
Yang, J., Xu, R., Qi, Z., & Shi, Y. (2022) · Procedia Computer Science · doi:10.1016/j.procs.2022.01.057 · Cited by 77
Leveraging computer vision towards high-efficiency autonomous industrial facilities
Yousif, I., Burns, L., Kalach, F. E., & Harik, R. (2024) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-024-02396-1 · Cited by 77
“An unsupervised learning algorithm was implemented to detect surface defects on textured surfaces (Mei et al., 2018).”
A Multilevel Information Fusion-Based Deep Learning Method for Vision-Based Defect Recognition
Gao, Y., Gao, L., Li, X., & Wang, X. V. (2019) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2019.2947800 · Cited by 73
“[12] proposed an unsupervised learning method that used a convolutional autoencoder to identify the defect and defect-free.”
BackgroundMethodInfluentialAMI-Net: Adaptive Mask Inpainting Network for Industrial Anomaly Detection and Localization
Luo, W., Yao, H., Yu, W., & Li, Z. (2024) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2024.3368142 · Cited by 71
“To address this issue, numerous methods [15], [16], [17], [18], [19] have been proposed.”
A new graph-based semi-supervised method for surface defect classification
Wang, Y., Gao, L., Gao, Y., & Li, X. (2020) · Robotics and Computer-Integrated Manufacturing · doi:10.1016/j.rcim.2020.102083 · Cited by 71
Yao, H., Yu, W., & Wang, X. (2022) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2022.3204368 · Cited by 69
“[19] first proposed a multiscale convolutional denoising autoencoder (MSCDAE) to localize texture defects while using only defect-free samples for model training.”
BackgroundMethodAttention-Based Multiscale Feature Fusion for Efficient Surface Defect Detection
Zhao, Y., Liu, Q., Su, H., Zhang, J., Ma, H., Zou, W., & Liu, S. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3372229 · Cited by 69
“Considering the consistency and regularity of the background in fabric images, Mei et al. [17] proposed a convolutional denoising autoencoder network (CDAE) to locate irregular fabric textures, which uses only defect-free samples for training.”
BackgroundA Survey of Vision-Based Methods for Surface Defects’ Detection and Classification in Steel Products
Ibrahim, A. A. M. S., & Tapamo, J. (2024) · Informatics · doi:10.3390/informatics11020025 · Cited by 70
“Mei et al. [139] proposed an unsupervised learning method involving the construction of a convolutional denoising autoencoder architecture based on a Gaussian pyramid to discern defective and defect-free regions.”
MethodA Contactless PCBA Defect Detection Method: Convolutional Neural Networks With Thermographic Images
Jeon, M., Yoo, S., & Kim, S.-W. (2022) · IEEE Transactions on Components Packaging and Manufacturing Technology · doi:10.1109/tcpmt.2022.3147319 · Cited by 68
Unsupervised fabric defects detection based on spatial domain saliency and features clustering
Zhao, S., Zhong, R. Y., Wang, J., Xu, C., & Zhang, J. (2023) · Computers & Industrial Engineering · doi:10.1016/j.cie.2023.109681 · Cited by 67
Learning Global-Local Correspondence With Semantic Bottleneck for Logical Anomaly Detection
Yao, H., Yu, W., Luo, W., Qiang, Z., Luo, D., & Zhang, X. (2023) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2023.3314801 · Cited by 66
“In addition, MSCDAE was proposed in [11] for texture defect detection, MS-FACE utilizing latent feature clustering was proposed in [12], and a multi-level reconstruction method was proposed in [13].”
MethodGaussian-IoU loss: Better learning for bounding box regression on PCB component detection
Liu, X., Hu, J., Wang, H., Zhang, Z., Lu, X., Sheng, C., Song, S., & Nie, J. (2021) · Expert Systems with Applications · doi:10.1016/j.eswa.2021.116178 · Cited by 65
Liu, T., Li, B., Zhao, Z., Du, X., Jiang, B., & Geng, L. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2210.14485 · Cited by 63
Learning deep feature correspondence for unsupervised anomaly detection and segmentation
Yang, J., Shi, Y., & Qi, Z. (2022) · Pattern Recognition · doi:10.1016/j.patcog.2022.108874 · Cited by 62
A Survey of Defect Detection Applications Based on Generative Adversarial Networks
He, X., Chang, Z., Zhang, L., Xu, H., Chen, H., & Luo, Z. (2022) · IEEE Access · doi:10.1109/access.2022.3217227 · Cited by 62
“The second is the defect detection method based on unsupervised learning [34], which usually only needs to input normal non-defect samples into the network for training, also known as one-class learning.”
MethodA Systematic Review of Machine-Vision-Based Leather Surface Defect Inspection
Chen, Z., Deng, J., Zhu, Q., Wang, H., & Chen, Y. (2022) · Electronics · doi:10.3390/electronics11152383 · Cited by 61
“[10] proposed a Gaussian pyramid-based multiscale convolutional denoising auto-encoder architecture (MSCDAE) to detect and localize defects with only defect-free samples, which is an unsupervised learning-based defect inspection approach.”
BackgroundAn Unsupervised Fault-Detection Method for Railway Turnouts
Guo, Z., Wan, Y., & Ye, H. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.2998863 · Cited by 61
“, [37]–[40]), to improve the condition monitoring of turnouts, this article develops a novel unsupervised fault-detection method using current signals.”
MethodYan, Y., Wang, D., Zhou, G., & Chen, Q. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3107586 · Cited by 60
An Anomaly Feature-Editing-Based Adversarial Network for Texture Defect Visual Inspection
Yang, H., Zhou, Q., Song, K., & Yin, Z. (2021) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2020.3015765 · Cited by 60
“In contrast, unsupervised learning [10], [11] shows great potential for automated online defect inspection in industrial processes, because it requires only readily available defect-free samples.”
BackgroundSelf-citationJiang, Y., Cao, Y., & Shen, W. (2023) · Knowledge-Based Systems · doi:10.1016/j.knosys.2023.110982 · Cited by 60
Few-shot semantic segmentation for industrial defect recognition
Shi, X., Zhang, S., Cheng, M., He, L. L., Tang, X., & Cui, Z. (2023) · Computers in Industry · doi:10.1016/j.compind.2023.103901 · Cited by 59
Zhang, R., Wang, H., Feng, M., Liu, Y., & Yang, G. (2024) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2024.3384583 · Cited by 59
“Most of the methods currently used for surface anomaly detection are image reconstruction-based methods [10], [11], [12], which mainly use autoencoders (AEs) as the architecture.”
MethodAn Efficient Network for Surface Defect Detection
Lin, Z., Ye, H., Zhan, B., & Huang, X. (2020) · Applied Sciences · doi:10.3390/app10176085 · Cited by 52
Multisensor Image Fusion for Automated Detection of Defects in Printed Circuit Boards
Li, M., Yao, N., Liu, S., Li, S., Zhao, Y., & Kong, S. G. (2021) · IEEE Sensors Journal · doi:10.1109/jsen.2021.3106057 · Cited by 57
A Survey of Surface Defect Detection of Industrial Products Based on A Small Number of Labeled DataPreprint
Jin, Q., & Chen, L. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2203.05733 · Cited by 55
“MEI [65] combined the idea of image pyramid hierarchy and Convolutional Denoising Autoencoder (CDAE) network to detect texture image defects.”
BackgroundFabric Defect Detection Method Combing Image Pyramid and Direction Template
Xie, H., Zhang, Y., & Wu, Z. (2019) · IEEE Access · doi:10.1109/access.2019.2959880 · Cited by 54
“Meanwhile, [19], [20] introduced the unsupervised learning model DAE to solve the data labeling problem and defect localization (in contrast, supervised learning requires pre-labeled data), and themethod only needs a few non-defective images to train models.”
BackgroundDeep Learning and Image Recognition
Li, C., Li, X., Chen, M., & Sun, X. (2023) · IEEE International Conference on Electronic Information and Communication Technology (ICEICT) · doi:10.1109/iceict57916.2023.10245041 · Cited by 51
“Industrial defect detection, which refers to the process of identifying and analyzing defects in a product or process in industrial settings [12-13].”
BackgroundDiffDD: A surface defect detection framework with diffusion probabilistic model
Zhou, X., Zhang, Y., Ren, Z., Mi, T., Feng, K., Zhou, S., & Jiang, Z. (2024) · Advanced Engineering Informatics · doi:10.1016/j.aei.2024.102637 · Cited by 51
Detecting textile micro-defects: A novel and efficient method based on visual gain mechanism
Wei, B., Hao, K., Gao, L., & Tang, X. (2020) · Information Sciences · doi:10.1016/j.ins.2020.06.035 · Cited by 51
Survey on industrial defect detection with deep learning
罗, 东., 蔡, 雨., 杨, 子., 章, 哲., 周, 瑜., & 白, 翔. (2022) · Scientia Sinica Informationis · doi:10.1360/ssi-2021-0336 · Cited by 50
Automatic quality inspection in additive manufacturing using semi-supervised deep learning
Manivannan, S. (2022) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-022-02000-4 · Cited by 48
“, 2020), fabric/texture inspection (Mei et al., 2018; Zheng et al., 2020), tile inspection (Rudolph et al.”
BackgroundFlexible electronics manufacturing technology and equipment
Yin, Z., Huang, Y., Yang, H. S., Chen, J., Duan, Y., & Chen, W. (2022) · Science China Technological Sciences · doi:10.1007/s11431-022-2098-1 · Cited by 48
Self-citationDG-GAN: A High Quality Defect Image Generation Method for Defect Detection
He, X., Luo, Z., Li, Q., Chen, H., & Li, F. (2023) · Sensors · doi:10.3390/s23135922 · Cited by 45
“One is based on supervised learning [21]; the other is based on unsupervised learning [22].”
BackgroundAutomatic defect detection of texture surface with an efficient texture removal network
Liang, Y., Xu, K., Zhou, P., & Zhou, D. (2022) · Advanced Engineering Informatics · doi:10.1016/j.aei.2022.101672 · Cited by 45
STD2: Swin Transformer-Based Defect Detector for Surface Anomaly Detection
Mia, M. S., & Li, C. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3492728 · Cited by 43
Visual Anomaly Detection for Images: A SurveyPreprint
Yang, J., Xu, R., Qi, Z., & Shi, Y. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2109.13157 · Cited by 44
“A typical method of image reconstruction is to compress and reconstruct the input image with the deep convolution autoencoder [45], [46], [47].”
MethodAutomated detection of textured-surface defects using UNet-based semantic segmentation network
Enshaei, N., Ahmad, S., & Naderkhani, F. (2020) · IEEE Conference on Prognostics and Health Management · doi:10.1109/icphm49022.2020.9187023 · Cited by 41
“Although the researchers for anomaly detection on textured surfaces in PHM context are vast, the results are not satisfactory due to existence of divers patterns in texture surfaces [23]-[24].”
BackgroundA Novel Pixel-Wise Defect Inspection Method Based on Stable Background Reconstruction
Lv, C., Shen, F., Zhang, Z., Xu, D., & He, Y. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.3038413 · Cited by 40
“The eight methods include AE [27], MSCDAE [18], AnoGAN [28], f-AnoGAN [19], CNN Feature Dictionary [29], Texture Inspection [30], Iterative Projection [21], and FCDD [32].”
BackgroundMethodInfluentialYee, T. S., Shrifan, N. H. M. M., Al-Gburi, A. J. A., Isa, N. A. M., & Akbar, M. F. (2022) · IEEE Access · doi:10.1109/access.2022.3197291 · Cited by 40
“[143] for the detection of CUI in a coated steel pipeline using an X-band OERW approach.”
MethodAutomatic modulation classification of radar signals utilizing X-net
Chen, K., Zhang, J., Chen, S., Zhang, S., & Zhao, H. (2022) · Digital Signal Processing · doi:10.1016/j.dsp.2022.103396 · Cited by 39
Crack identification for marine engineering equipment based on improved SSD and YOLOv5
Jia, Z., Su, X. B., Ma, G., Dai, T., & Sun, J. (2022) · Ocean Engineering · doi:10.1016/j.oceaneng.2022.113534 · Cited by 39
Normal Reference Attention and Defective Feature Perception Network for Surface Defect Detection
Luo, W., Yao, H., & Yu, W. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3268658 · Cited by 39
“[13] proposed a robust model named MSCDAE for texture defect inspection in the presence of noise, while Yang et al.”
MethodUnsupervised Defect Segmentation of Magnetic Tile Based on Attention Enhanced Flexible U-Net
Cao, X., Chen, B., & He, W. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3170989 · Cited by 38
PCB defect detection with self-supervised learning of local image patches
Yao, N., Zhao, Y., Kong, S. G., & Guo, Y. (2023) · Measurement · doi:10.1016/j.measurement.2023.113611 · Cited by 37
A new Feature-Fusion method based on training dataset prototype for surface defect recognition
Wang, Y., Li, X., Gao, Y., Wang, L., & Gao, L. (2021) · Advanced Engineering Informatics · doi:10.1016/j.aei.2021.101392 · Cited by 36
Anomaly Detection Using Siamese Network with Attention Mechanism for Few-Shot Learning
Takimoto, H., Seki, J., Situju, S. F., & Kanagawa, A. (2022) · Applied Artificial Intelligence · doi:10.1080/08839514.2022.2094885 · Cited by 36
“Mei et al. proposed a method to localize and detect using only defect-free data for model training (Mei, Yang, and Yin 2018).”
BackgroundMethodLiu, H., Hu, R., Dong, H., & Liu, Z. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3548193 · Cited by 35
“Mei et al. [12], for instance, combined convolutional layers with encoders to capture image features, utilizing decoders for reconstruction, ensuring that the encoded representations retain vital information [13].”
Automatic Wear Measurement of Pantograph Slider Based on Multiview Analysis
Lu, S., Liu, Z., Li, D., & Shen, Y. (2020) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2020.2997724 · Cited by 35
“The most common domains using the computer vision technology include the defect detection of fabric [15], [16], steel [17], [18], castings [19],”
Lu, H., Fang, M., Qiu, Y. N., & Xu, W. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3229728 · Cited by 35
“OBJECT detection is an important downstream task in the field of computer vision, which is widely used in object tracking [1], [2], defect detection [3], [4], and other”
BackgroundAn automated shearography system for cylindrical surface inspection
Ye, Y., Ma, K., Zhou, H., Arola, D. D., & Zhang, D. (2018) · Measurement · doi:10.1016/j.measurement.2018.11.085 · Cited by 35
DestripeCycleGAN: Stripe Simulation CycleGAN for Unsupervised Infrared Image Destriping
Yang, S., Qin, H., Yuan, S., Yan, X., & Rahmani, H. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3476560 · Cited by 34
“Benefiting from the refinement representation of probability distribution, unsupervised generative networks have demonstrated more promising technology in outdoor scenarios [35], [36], [37].”
BackgroundBrain-Inspired Interpretable Network Pruning for Smart Vision-Based Defect Detection Equipment
Wang, J., Zhao, S., Xu, C., Zhang, J., & Zhong, R. Y. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3188349 · Cited by 34
“Mei et al. [7] proposed a multiscale convolutional denoising autoencoder to realize unsupervised segmentation of several types of fabric defects.”
BackgroundFabric defect detection and classification using modified VGG network
Sabeenian, R. S., Paul, E., & Prakash, C. (2022) · Journal of the Textile Institute · doi:10.1080/00405000.2022.2105112 · Cited by 34
“The articles (Eldho Paul et al., 2021; Hanbay et al. 2016; Hu et al., 2020; Jing et al., 2019; Li et al., 2019; Mei et al., 2018; Sabeenian et al., 2019; Shi et al., 2021; Wang et al., 2021; Zhang et al., 2019; Zhang & Tang, 2019) also describe recent developments and new methodologies in the field…”
MethodUnseen-Material Few-Shot Defect Segmentation With Optimal Bilateral Feature Transport Network
Shan, D., Zhang, Y., Coleman, S., Kerr, D., Liu, S., & Hu, Z. (2022) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2022.3216900 · Cited by 34
“Although some methods [7], [8], [9] can segment defects from images without any labels, a large-scale defect dataset is still required [10].”
Zhang, H., Xiong, W., Lu, S., Chen, M., & Yao, L. (2023) · Textile Research Journal · doi:10.1177/00405175231158134 · Cited by 33
“Mei et al. 17 presented a multi-scale convolutional denoising auto-encoder method (MSCDAE), the method based on multi-scale convolution, reconstruction images on the different levels of Gaussian pyramid, and synthesizes reconstruction results of different resolution channels to realize defect detection.”
MethodBalanced multi-scale target score network for ceramic tile surface defect detection
Cao, T., Song, K., Xu, L., Feng, H., Yan, Y., & Guo, J. (2023) · Measurement · doi:10.1016/j.measurement.2023.113914 · Cited by 33
Asymmetric Teacher–Student Feature Pyramid Matching for Industrial Anomaly Detection
Zhu, J., Yan, P., Jiang, J., Cui, Y., & Xu, X. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3338681 · Cited by 32
“With the rapid development of convolutional neural networks (CNNs), researchers have proposed a wide range of CNN-based anomaly detection methods [4], [5], [6], [7], [8].”
BackgroundZhu, Y., Yang, R., He, Y., Ma, J., Guo, H., Yang, Y., & Zhang, L. (2021) · IEEE Access · doi:10.1109/access.2021.3064180 · Cited by 31
“A similar approach is adopted in [26]–[28].”
Low-Shot Unsupervised Visual Anomaly Detection via Sparse Feature Representation
Zhang, F., Zhu, H., Cen, Y., Kan, S., Zhang, L., Vadakkepat, P., & Lee, T. H. (2024) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2024.3420818 · Cited by 30
RETRACTED ARTICLE: Efficient utilization of deep learning for the detection of fabric defects
Zahra, A., Amin, M. A., El‐Samie, F. E. A., & Emam, M. (2024) · Neural Computing and Applications · doi:10.1007/s00521-023-09137-0 · Cited by 30
“Mei et al. [4, 5] suggested a Multi-Scale Convolutional Denoising Auto-Encoding (MSCDAE) approach to automatically detect and localize defects in fabric by combining multiple convolutional network models with Gaussian pyramid decomposition.”
RetractedZhang, H., Wang, S., Lu, S., Yao, L., & Hu, Y. (2023) · Textile Research Journal · doi:10.1177/00405175221149450 · Cited by 29
“Mei et al. 28 proposed an unsupervised fabric defect detection method based on the Multi-Scale Convolutional Denoising Autoencoder (MSCDAE).”
MethodSDDiff: Semi-supervised surface defect detection with Diffusion Probabilistic Model
Wang, X., Li, W., Liu, C., & Ouyang, N. (2024) · Measurement · doi:10.1016/j.measurement.2024.115276 · Cited by 29
D4Net: De-deformation defect detection network for non-rigid products with large patterns
Xu, X., Chen, J., Zhang, H., & Ng, W. W. Y. (2020) · Information Sciences · doi:10.1016/j.ins.2020.05.050 · Cited by 29
Template-based Feature Aggregation Network for industrial anomaly detection
Luo, W., Yao, H., & Yu, W. (2024) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2023.107810 · Cited by 28
Computer Vision Analysis on Material Characterization Images
Cheng, D., Sha, W., Zuo, X., Huang, L., Du, Y., Tang, S., Guo, Y., Cao, Y.-C., & Cheng, S. (2021) · Advanced Intelligent Systems · doi:10.1002/aisy.202100158 · Cited by 27
SCL—Segmentation–Classification combined Loss for surface defect detection
Versini, E., Snidaro, L., & Liani, A. (2022) · Expert Systems with Applications · doi:10.1016/j.eswa.2022.116710 · Cited by 27
Zhang, H., Liu, S., Lu, S., Yao, L., & Li, P. (2023) · Coloration Technology · doi:10.1111/cote.12705 · Cited by 27
Multi-Category Decomposition Editing Network for the Accurate Visual Inspection of Texture Defects
Yang, H. S., Zhu, H., Li, J., Chen, J., & Yin, Z. (2023) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2023.3295872 · Cited by 27
“Product surfaces with uniform structures, such as steel [1], nanomaterials [2], and screens [3], are commonly referred to as textured surfaces.”
BackgroundSelf-citationLi, M., Liu, J., Zhu, T., Zhou, W., & Zhou, C. (2019) · Sensors · doi:10.3390/s19092083 · Cited by 27
“In Reference [12], an unsupervised-learning-based approach was proposed for automated defect inspection on textured surfaces, and convolutional denoising autoencoder network was introduced to reconstruct image patches.”
MethodInfluentialAutomatic zipper tape defect detection using two-stage multi-scale convolutional networks
Fang, H., Xia, M., Liu, H., Chang, Y., Wang, L., & Liu, X. (2020) · Neurocomputing · doi:10.1016/j.neucom.2020.09.046 · Cited by 27
Metadata
- Authors
- Mei, S., Yang, H., & Yin, Z.
- Affiliations
- State Key Laboratory of Digital Manufacturing Equipment and Technology (CN)
- Venue
- IEEE Transactions on Instrumentation and Measurement
- Published
- 2018-06
- Volume
- 67
- Issue
- 6
- Pages
- 1266-1277
- ISSN
- 0018-9456 (ISSN-L)
- 1557-9662
- OpenAlex
- W2791709739
- Semantic Scholar
- 38ec86557f805ec27e18e5517a34ca43b2bfa60f
- Cited by
- 308
- References
- 51
- Citations by year
- 2026: 16 · 2025: 41 · 2024: 36 · 2023: 59 · 2022: 63 · 2021: 44 · 2020: 36 · 2019: 11 · 2018: 2
- FWCI
- 19.21
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- IEEE Transactions on Instrumentation and Measurement (publishedVersion)
- Funders
- National Science Foundation of China (51327801, 51475193)
- Major Project Foundation of Hubei Province (2016AAA009)
- Topics
- Physical Sciences › Engineering › Industrial and Manufacturing Engineering › Industrial Vision Systems and Defect Detection
- Physical Sciences › Engineering › Mechanical Engineering › Non-Destructive Testing Techniques
- Physical Sciences › Engineering › Civil and Structural Engineering › Infrastructure Maintenance and Monitoring
- Keywords
- automated defect detection, unsupervised learning, textured surfaces, convolutional denoising autoencoder, Gaussian pyramid, reconstruction residuals, defect localization, multimodal fusion, industrial inspection
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.