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
Items 201–300 of 323
Zhang, R., & Hu, H. (2025) · IEEE Transactions on Image Processing · doi:10.1109/tip.2025.3607638 · Cited by 6
“…The widely adopted methods for surface anomaly detection are based on the image reconstruction, generally utilizing three types of network architectures: auto-encoders (AE) [1], [2], [4], [7], [8], [14], [15], variational auto-encoders (VAE) [19], and generative adversarial networks (GAN) [20].”
Defect detection method for curved and reflected paint surfaces of rearview mirrors
Liu, P., Yuan, X., Ma, Y., Han, Q., Lv, J., Zhao, D., Zhang, J., & Duan, W. (2024) · Measurement · doi:10.1016/j.measurement.2024.115193 · Cited by 6
Masked contrastive generative adversarial network for defect detection of yarn-dyed fabric
Zhang, H., Lu, Z., Chen, X., Lu, S., & Yao, L. (2024) · The Journal of Supercomputing · doi:10.1007/s11227-024-06711-8 · Cited by 6
“Six unsupervised color fabric defect detection models: MSCDAE [7], DCAE [16], VAE-SSIM [27], UDCAE [17], DCGAN [28] and the proposed MCGAN were tested quantitatively and qualitatively on two datasets YDFDID and ZJU-Leaper, and the results were analyzed.”
Surface Defect Inspection Under a Small Training Set Condition
Yu, W., Zhang, Y., & Shi, H. (2019) · Lecture notes in computer science · doi:10.1007/978-3-030-27538-9_44 · Cited by 6
Liu, Q., Wang, M., Wang, G., Gao, R., & HANAJIMA, N. (2020) · Journal of Robotics Networking and Artificial Life · doi:10.2991/jrnal.k.200909.015 · Cited by 6
“/ Journal of Robotics, Networking and Artificial Life 7(3) 212–216 213”
BackgroundInfluentialWang, J., Xu, G., Li, C., Gao, G., & Cheng, Y. (2026) · Signal Image and Video Processing · doi:10.1007/s11760-026-05448-6 · Cited by 6
“To introduce more image information into the reconstruction process, some attempts have been made to execute reconstruction in the space of pre-trained features [37], [45], image pyramids [39], multi-frequency [46], [47] or texture–structure features [48].”
Defect Segmentation in Surfaces using Deep Learning
Mittal, S., Chopra, C., Trivedi, A., & Chanak, P. (2019) · International Conference on Issues and Challenges in Intelligent Computing Techniques · doi:10.1109/icict46931.2019.8977696 · Cited by 6
“Some of the methods in this domain include fourier transform, wavelet transform and Gabor transform (6).”
MethodInfluentialSplit-Brain Autoencoder Approach for Surface Defect Detection
Ulutaş, T., Öz, M. A. N., Mercımek, M., & Kaymakçı, Ö. T. (2020) · International Conference on Electrical, Communication, and Computer Engineering (ICECCE) · doi:10.1109/icecce49384.2020.9179311 · Cited by 6
“A better approach, Multiscale Denoising Convolutional Autoencoders was proposed by Mei et al. [8] and Yang et al. [9].”
MethodDual-task guided network: hybrid supervised learning for surface defect detection
Zou, J., & Lv, J. (2025) · Measurement Science and Technology · doi:10.1088/1361-6501/adb6cb · Cited by 5
“In unsupervised learning, Mei et al [11] pioneered the use of multi-level convolutional denoising autoencoders for anomaly detection in surface inspection, demonstrating the potential of reconstruction-based approaches for unsupervised defect detection on diverse surface textures.”
Reconstruction-Based Anomaly Localization via Knowledge-Informed Self-Training
Cheng, Q., Lao, X., & Li, C. (2025) · IEEE Transactions on Neural Networks and Learning Systems · doi:10.1109/tnnls.2025.3542229 · Cited by 5
“Some researchers utilize other NNs for industrial anomaly localization [10], [11], [12], [13], [14].”
Developing the Smart Sorting Screw System Based on Deep Learning Approaches
Lin, W.-J., Chen, J., Young, H., Hung, C., & Li, K. (2021) · Applied Sciences · doi:10.3390/app11209751 · Cited by 5
“[20] proposed the convolutional denoising autoencoder networks to detect and localize the defects at the same time.”
BackgroundA Multiscale Convolutional Registration Network for Defect Inspection on Periodic Lace Surfaces
Ding, X., Lu, B., & Huang, B. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3150581 · Cited by 5
Self-Supervised Surface Defect Detection Methods for Industrial Products
Lv, G., Mei, Y., Sun, Y., Gu, Y., & Nian, F. (2022) · China Automation Congress (CAC) · doi:10.1109/cac57257.2022.10055459 · Cited by 5
“Self-encoder-based train models to learn patterns of normal instances and determine anomalies when the data reconstruction error is high [8].”
BackgroundChen, J. W., Lin, W.-J., Li, K., Mrozek, D., & Hung, C. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3548207 · Cited by 5
“Conversely, numerous studies have augmented the input image by introducing various types of noise, such as pepper noise [3], [4], [28] or Gaussian noise [4], to make the distribution of training and testing images closer to that of the original images.”
Multi-Class Object Learning with Application to Fabric Defects Detection
Wei, B., Gao, L., Tang, X., & Hao, K. (2021) · AATCC Journal of Research · doi:10.14504/ajr.8.s1.20 · Cited by 5
Cavaliere, G. (2024) · Libera Università di Bolzano · Cited by 5
Zhou, Y., Ge, Q., Zhang, M., He, X., & Wang, K. (2026) · Neurocomputing · doi:10.1016/j.neucom.2025.132598 · Cited by 4
Unsupervised Anomaly Detection of Industrial Images Based on Dual Generator Reconstruction Networks
Gu, C., Ren, S., & Duan, Q. (2024) · Information Technology And Control · doi:10.5755/j01.itc.53.2.36018 · Cited by 4
“Mei et al. [12] reconstructed images by dividing them into blocks and used denoising AE to locate anomalies in texture images.”
MethodDetection of Pin Defects in Transmission Lines Based on Dynamic Receptive Field
Zhang, Z., Du, J., Qian, S., & Xie, C. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-18916-6_32 · Cited by 4
Patch-Wise Auto-Encoder for Visual Anomaly Detection
Cui, Y., Liu, Z., & Lian, S. (2023) · Proceedings - International Conference on Image Processing · doi:10.1109/icip49359.2023.10222183 · Cited by 4
“Among the methods based on reconstruction, one strategy to address anomaly detection problem is to exploit convolutional auto-encoders [9] [10].”
MethodUnsupervised Anomaly Detection via Normal Feature-Enhanced Reverse Teacher–Student Distillation
Wang, X., Fan, J., Yan, F., Hu, H., Zeng, Z., Wu, P., Huang, H., & Zhang, H. (2024) · Electronics · doi:10.3390/electronics13204125 · Cited by 4
Progressive mask‐oriented unsupervised fabric defect detection under background repair
Shancheng, T., Jin, Z., Dai, F., Zhang, Y., Liang, S., & Lu, J. (2023) · Coloration Technology · doi:10.1111/cote.12719 · Cited by 4
Fabric image inspection using deep learning approach
Sizyakin, R., Voronin, V., Gapon, N., Semenishchev, E. A., Zelensky, A. A., & Ilyukhin, Y. V. (2022) · Optical Metrology and Inspection for Industrial Applications · doi:10.1117/12.2646485 · Cited by 4
“In [13], an approach is presented for reconstructing image fragments using convolutional networks of an automatic encoder with noise reduction at different levels of the Gaussian pyramid and synthesizing detection results from these channels with different resolutions.”
MethodHuang, Y., Sang, M., Xing, L., Hu, H., & Liu, T. (2021) · Optics Express · doi:10.1364/oe.438564 · Cited by 4
Ma, Q., Zhang, E., Chen, Y., & Duan, J. (2025) · Journal of Real-Time Image Processing · doi:10.1007/s11554-025-01751-6 · Cited by 3
“To enhance the performance of the encoder, reference [38] employed a Gaussian pyramid to reconstruct image sub-blocks.”
Memory guided dual discrepancy network with domain adaptation for few-shot anomaly detection
Zhang, F., Gong, L., Lin, S., & Zhao, Y. (2026) · Expert Systems with Applications · doi:10.1016/j.eswa.2026.131711 · Cited by 3
Wang, L., Huang, X., Zheng, Z., & Ruan, H. (2023) · PLoS ONE · doi:10.1371/journal.pone.0279035 · Cited by 3
Defect detection of color-patterned fabric based on DenoisingGAN
Zhang, H., Wang, S., Mi, H., Lu, S., Yao, L., & Ge, Z. (2023) · International Journal of Clothing Science and Technology · doi:10.1108/ijcst-03-2022-0032 · Cited by 3
Lightweight Reconstruction Network for Surface Defect Detection Based on Texture Complexity Analysis
Shi, H., Li, G., & Bao, H. (2023) · Electronics · doi:10.3390/electronics12173617 · Cited by 3
Yan, H., Zhan, X., Wu, Z., Cheng, J., Wen, L., & Jia, X. (2024) · Proceedings of the Institution of Mechanical Engineers Part E Journal of Process Mechanical Engineering · doi:10.1177/09544089241258027 · Cited by 3
“3,4 The core principle of AE is to encode and reconstruct normal data, and then take the reconstruction error as an indicator to judge anomalies, so as to achieve anomaly detection.”
A Learning-Based Surface Defect Detection Method for Internal Surfaces of Bellows
Zhang, Z., Xu, J., Wang, S., Li, F., & Ling, Q. (2021) · Chinese Control and Decision Conference (CCDC) · doi:10.1109/ccdc52312.2021.9601703 · Cited by 3
Deep Neural Network Models for the Recognition of Traffic Signs Defects
Nagy, A. M., & Czúni, L. (2021) · IEEE International Conference on Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications (IDAACS) · doi:10.1109/idaacs53288.2021.9660936 · Cited by 3
Yarn-dyed Shirt Piece Defect Detection Based on U-shaped Swin Transformer Auto-encoder
Zhang, H., Xiong, W., Zhang, W., & Lu, S. (2022) · IEEE Data Driven Control and Learning Systems Conference (DDCLS) · doi:10.1109/ddcls55054.2022.9858466 · Cited by 3
基于U型Swin Transformer自编码器的色织物缺陷检测
Yuan-yuan, 黄. H., Wenbo, 熊. X., Hongwei, 张. Z., & Weiwei, 张. Z. (2023) · Laser & Optoelectronics Progress · doi:10.3788/lop220691 · Cited by 3
A Novel Fault Location Method for Power Cables Based on Unsupervised Learning
Cheng, X. F., Xianjun, S., Yongtao, J., Yuli, W., & Li, M. (2020) · IEEE International Conference on High Voltage Engineering and Application (ICHVE) · doi:10.1109/ichve49031.2020.9280039 · Cited by 3
“In the past 20 years, with the development of image processing technology, machine learning methods have been widely used [5-6].”
A Lightweight Reconstruction Network for Surface Defect InspectionPreprint
Hu, C., Yao, J., Wu, W., Qiu, W., & Zhu, L. (2022) · Preprints.org · doi:10.20944/preprints202210.0355.v1 · Cited by 3
Decision Fusion Network with Perception Fine-tuning for Defect ClassificationPreprint
Jiang, X., Tian, S., Zhu, Z., Lu, Y., Liu, H., Chen, L., Li, S., & Xu, M. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2309.12630 · Cited by 3
Manifold-Constrained Dynamic Decoupling Learning for Unsupervised Multiclass Anomaly Detection
Qiu, S., Zhao, G., Wang, X., Yan, F., & Lin, B. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3602566 · Cited by 2
“Reconstruction-driven methods [28], [29], [30], [31], [32], [33] learn the reconstruction rules of normal samples through autoencoders [34], [35], transformers [36], [37], [38], or di ff usion models [39], [40].”
Semi-Supervised Surface Defect Segmentation via Contrastive Learning
Wang, R., He, Z., Huang, H., & Zheng, Y. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3612555 · Cited by 2
“To reduce reliance on annotated samples, researchers have proposed various methods, including semi-supervised learning [6], unsupervised learning [7], and few-shot learning [8].”
Pieressa, A., Lin, C., Lucchetta, G., & Turng, L. (2026) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-026-17386-1 · Cited by 2
Scale-aware modulation network for unsupervised medical anomaly detection
Cao, H., Sun, S., & Yan, H. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.133672 · Cited by 2
Chatterjee, S., McCarthy, É., Nyabadza, A., Bandhu, D., McCann, R., & Brabazon, D. (2026) · Journal of Engineering Research · doi:10.1016/j.jer.2026.04.030 · Cited by 2
PCBClip: Vision-Language Defect Detection Model for Low-Sample Inspection Systems
Chen, Z., Guo, F., Chen, Z., Zuo, L., & Liu, Y. (2026) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2026.3711750 · Cited by 2
Cui, J., Li, X., Zhang, X., Huang, S., & Feng, Y. (2023) · Multimedia Tools and Applications · doi:10.1007/s11042-023-17388-1 · Cited by 2
“(cid:4) For the metadata meta As to details of learning from the structure of autoencoder [31] in anomaly detection, a symmetric convolutional neural network composed of encoding-decoding is constructed as the pretraining network.”
InfluentialI2DKPCN: an unsupervised deep learning network
Zhao, R., & Shi, F. (2022) · Applied Intelligence · doi:10.1007/s10489-021-03007-9 · Cited by 2
“Recently, increasing attempts have been made on unsupervised deep learning [10–13], and the unsupervised learning systems have the ability to utilize the unlabeled data.”
BackgroundA convolutional neural network method for damage location based on raw guided Lamb wave technique
Yu, S., Fan, C., Chen, Q., Gao, B., & Guo, J. (2021) · IEEE Far East NDT New Technology & Application Forum (FENDT) · doi:10.1109/fendt54151.2021.9749662 · Cited by 2
“In construction, industrial production, aerospace and other fields, some scholars have studied the application of ultrasonic nondestructive testing based on deep learning, Haixing Zhu studied the classification of welding surface defects based on deep learning, The defects were identified by convolutional neural netwo…”
BackgroundZhang, H., Liu, S., Meng, L., Lu, S., & Li, P. (2023) · IEEE Data Driven Control and Learning Systems Conference (DDCLS) · doi:10.1109/ddcls58216.2023.10166540 · Cited by 2
“Mei et al. [5] proposed an unsupervised fabric defect detection method based on a multi-scale convolutional denoising autoencoder (MSCDAE), which achieved fabric defect detection by comparing the pixel-level differences between the multi-scale reconstruction map output by the model and the input…”
MethodAn improved Yolov5-based algorithm for pavement disease detection in airport flight areas
Wang, S., Hao, S., Zhang, P., Cao, T., Shao, L., & Li, Y. (2025) · doi:10.1117/12.3045297 · Cited by 2
“Based on this method, the robustness and accuracy of the model were effectively improved, and at the same time, it had certain advantages in the detection of uniform and irregular texture surfaces [6].”
How Incompletely Segmented Information Affects Multi-Object Tracking and Segmentation (MOTS)
Chou, Y.-S., Wang, C.-Y., Lin, S.-D., & Liao, H.-Y. M. (2020) · Proceedings - International Conference on Image Processing · doi:10.1109/icip40778.2020.9190802 · Cited by 2
“For example, it can be applied to successfully solve the problems of object detection [1, 2, 3, 4], pattern matching [5, 6], visual inspection [7], etc.”
Multi-Region Two-Stream Deep Architecture for Visual Power Monitoring Systems
Gan, J., Jiang, W., Zhao, T., Wu, P., Zhang, G., & Zhang, Z. (2021) · IEEE Access · doi:10.1109/access.2021.3061084 · Cited by 2
“Generally, abnormalities can be automatically identified or located through various computer algorithms or machine learning models [2]–[8].”
BackgroundPM-AE: Pyramid Memory Autoencoder for Unsupervised Textured Surface Defect Detection
Yao, H., Li, D., Zhu, Y., & Yu, W. (2020) · International Conference on Mechanical, Control and Computer Engineering (ICMCCE) · doi:10.1109/icmcce51767.2020.00290 · Cited by 2
“The conventional AEs and their extension usually suffer from high computational cost [10] and the ability in generalization [12].”
BackgroundMethodInfluentialAutomated Shell Abrasion Detection for Cell Phone Based on GraphFPN With Learnable Sparse Priors
Wu, H., Guo, X., Xiao, F., Yan, W., & Wen, Z. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3150591 · Cited by 2
Fast and Few-Shot Unsupervised Defect Detection on Flexible Packaging Surfaces
Li, X., & Li, S. (2022) · IEEE Journal of Radio Frequency Identification · doi:10.1109/jrfid.2022.3208586 · Cited by 2
“Patch SVDD [4]uses patch-level features to improve performances, but the inference time increases linearly with more training samples.”
BackgroundOut-of-Distribution with Text-to-Image Diffusion Models
Tong, J., & Dai, L. (2023) · Lecture notes in computer science · doi:10.1007/978-981-99-8552-4_22 · Cited by 2
Self-Attention Mechanism based Visual Detection for Transmission Line Pins
Ding, S., Wang, Y., Wang, T., & Wang, H. (2023) · International Conference on Deep Learning Technologies (ICDLT) · doi:10.1145/3613330.3613335 · Cited by 1
Surface Defect Detection of Medical Cables Based on Anomaly Detection
Yue, J., Cao, G., Hu, Z., Yu, L., & Wang, H. (2023) · Annual Conference of the IEEE Industrial Electronics Society (IECON) · doi:10.1109/iecon51785.2023.10311868 · Cited by 1
“Unsupervised learning is suitable for defect detection scenarios that have a large number of positive samples but lack negative samples, such as Mei et al. [9] scaled images to different resolutions added random noise, and input them into an autoencoder (AE) for restoration.”
BackgroundHuang, Y., Wang, C., & Peng, T.-T. (2025) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-025-15601-z · Cited by 1
“Mei et al. proposed a method for detecting and localizing defects by using only data corresponding to defect-free conditions [8].”
Metal Surface Defect Detection Based on Transformer Merging Edge Information
Ling, J., Tan, C., & Cui, L. (2025) · International Symposium on System Security, Safety, and Reliability (ISSSR) · doi:10.1109/isssr65654.2025.00024 · Cited by 1
Liu, Q., Wang, Y., Yang, C., An, J., & Cheung, Y. (2025) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2025.3591089 · Cited by 1
“Due to the ability to fit complex data distributions [30], generative model-based reconstruction methods have become popular for unsupervised learning-based anomaly detection [31], [32], [33].”
Xiong, J., Zhang, Q., Wang, Q., Yang, J., Dong, X., & Liang, Q. (2025) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-025-16431-9 · Cited by 1
“Mei et al. [128] initially used convolutional denoising autoencoder networks to reconstruct image patches at various levels of the Gaussian pyramid.”
AnomNet: A Dual-Stage Centroid Optimization Framework for Unsupervised Anomaly Detection
Gao, Y., Wang, Y., Tu, X., & Shen, J. (2025) · Journal of Imaging · doi:10.3390/jimaging11090301 · Cited by 1
“Mei et al. [12] designed a Multi-Scale Convolutional Denoising Autoencoder (MSCDAE), integrating multi-pyramid layers and multimodal reconstruction results to improve detection accuracy across various material types.”
Towards Realistic Industrial Anomaly Detection: MADE-Net Framework and ManuDefect-21 Benchmark
Yang, J., Cao, J., & Duan, C. (2025) · Applied Sciences · doi:10.3390/app152010885 · Cited by 1
Surface Defect Detection Method Based on a Small Number of Labeled Images
Zakharov, A., Shamshin, M., & Львович, Ж. А. (2025) · Pattern Recognition and Image Analysis · doi:10.1134/s1054661825700282 · Cited by 1
Wang, H., Zhang, S., Zhang, H., & Lu, S. (2026) · Coloration Technology · doi:10.1111/cote.70052 · Cited by 1
A two-stage weakly supervised framework for endface defect detection in shaft metal parts
Liao, Y., Suo, X., Zhou, F., Liu, J., & Yin, S. (2026) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-025-02785-0 · Cited by 1
YOLO-DCF: dual distillation and context-aware fusion for defect detection
Xing, H., & Yang, Z. (2026) · Scientific Reports · doi:10.1038/s41598-026-46602-x · Cited by 1
“Mei et al. 20 proposed a multi-scale convolutional denoising autoencoder with high accuracy and robustness, capable of synthesizing results across multiple pyramid levels.”
Optical boundary discrimination in spatial posture measurement
Wu, Z., Wang, S., & Liu, T. (2021) · Applied Optics · doi:10.1364/ao.421790 · Cited by 1
A New Semi-Supervised Deep Learning Approach for Intelligent Defects Recognition
Gao, Y., Gao, L., & Li, X. (2020) · IEEE International Conference on Networking, Sensing and Control (ICNSC) · doi:10.1109/icnsc48988.2020.9238100 · Cited by 1
Wong, Z.-H., Thong, C.-M., Loh, W. M. E., & Wong, C. J. (2019) · International Journal of Engineering & Technology · doi:10.14419/ijet.v7i4.14.27693 · Cited by 1
“Non-referential approaches do not use any referential image for learning or comparison [7], [8].”
BackgroundAnomaly Detection and Segmentation Based on Defect Repaired Image Resynthesis
Dai, W., Erdt, M., & Sourin, A. (2021) · International Conference on Cyberworlds (CW) · doi:10.1109/cw52790.2021.00023 · Cited by 1
Presswork defect inspection using only defect-free high-resolution images
Guan, Z., Wang, Z., Zhu, Y., & Liu, G. (2022) · The Visual Computer · doi:10.1007/s00371-022-02403-7 · Cited by 1
“In [10,11], they add the salt-and-pepper noise to defect-free pictures and then use these pictures to generate defect-free pictures so that they can get the defect area through the difference between the input image and output image.”
BackgroundMethodResearch on Multi-View Image Reconstruction Technology Based on Auto-Encoding Learning
Zhang, T., Gu, S., Niu, J., & Cao, Y. (2022) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2022.027079 · Cited by 1
“This function focuses on the global difference of the image and does not consider local texture features, therefore the inpainting model with MSE as the loss function performs better on regular texture samples than on the irregular texture.”
MethodHierarchical Defect Detection Based On Reinforcement Learning
Fang, F., Xu, Q., & Lim, J. (2022) · Proceedings - International Conference on Image Processing · doi:10.1109/icip46576.2022.9897947 · Cited by 1
“Particularly, CNN-based methods [4, 5, 6, 7, 8, 9] have been applied and achieved remarkable performance on irregular”
MethodDefect Detection in Multiple Product Variants Using Hammering Test with Machine Learning
Yamashita, Y., Yoshida, K., Kishita, Y., & Umeda, Y. (2022) · International Journal of Automation Technology · doi:10.20965/ijat.2022.p0783 · Cited by 1
An Unsupervised Surface Anomaly Detection Method Based on Attention and ASPP
Huang, Y., Xie, X., Ning, W., Wu, D., Li, Z., & Yang, H. (2023) · Lecture notes in computer science · doi:10.1007/978-3-031-20099-1_16 · Cited by 1
Self-citationUnsupervised Time Series Anomaly Detection under Data Contamination
Lin, X., Li, Z., Huang, X., Chen, X., & Fan, H. (2022) · International Conference on Information Technology in Medicine and Education · doi:10.1109/itme56794.2022.00062 · Cited by 1
“Recently, with the rapid development of deep learning, methods based on deep convolutional neural network [6], [7], [8], [9] have been applied to the time series anomaly detection domain.”
BackgroundMethodA new multiscale texture surface defect detection method based on convolutional neural network
Li, K., Dong, M., & Li, D. (2022) · IEEE International Conference on Tools with Artificial Intelligence (ICTAI) · doi:10.1109/ictai56018.2022.00196 · Cited by 1
Research of Neural Network Methods of Visual Defectoscopy of Metal Structures
Chernikov, A., Smirnova, E., & Zhavoronkov, O. (2023) · International Conference on Industrial Engineering, Applications and Manufacturing (ICIEAM) · doi:10.1109/icieam57311.2023.10139006 · Cited by 1
“Usually, generative-adversarial networks [23] or a noise-canceling autoencoder [24][25] are used inunsupervised learning, which require only samples of images without defects.”
Chen, X., Lin, X., Li, Z., & Fan, H. (2023) · International Conference on Control, Automation and Robotics (ICCAR) · doi:10.1109/iccar57134.2023.10151775 · Cited by 1
“For example, Mei et al. [9] divide the data into fragments and use a multi-scale strategy to improve the accuracy of anomaly localization.”
MethodA Novel Denoising Autoencoder Method for Surface Defect Detection of Screw Products
Chen, J., Lin, W., & Hung, C. (2023) · Journal of Physics Conference Series · doi:10.1088/1742-6596/2609/1/012007 · Cited by 1
“It is not possible to implement supervised learning due to the requirement of a large number of defective samples for model training [2-6].”
BackgroundMulti-Channel Knowledge Distillation via Efficient Feature Fusion Module for Anomaly Segmentation
Zhou, Y., Huang, Z., Zeng, D., & Wu, Z. (2023) · China Automation Congress (CAC) · doi:10.1109/cac59555.2023.10451795 · Cited by 1
“Among these, reconstruction methods based on Autoencoders (AE) [10], [11] and Generative Adversarial Networks (GANs) [12], [13] have been widely applied.”
MethodZhou, Y., Li, C., Wang, S., Peng, G., Ma, S., Yang, Z., & Feng, Y. (2024) · KSCE Journal of Civil Engineering · doi:10.1007/s12205-024-0188-3 · Cited by 1
U-Shape Network for Chip Surface Defect Detection
Ruan, R., Zou, W., Li, Z., Qin, Z., Wu, Q., & Huang, S. (2024) · IEEE International Conference on e-Business Engineering (ICEBE) · doi:10.1109/icebe62490.2024.00053 · Cited by 1
“For example, [6] uses an autoencoder network to learn the representations of these local anomalies and fi nd common features between different defects.”
Weng, W., He, Z., Jiang, J., Zheng, G., Wan, A., & Cheng, X. (2025) · Engineering Research Express · doi:10.1088/2631-8695/ade034 · Cited by 1
Ameen, M., Alrahmawy, M. F., Aboueleneen, A., & Tolba, A. (2020) · Journal of Intelligent & Fuzzy Systems · doi:10.3233/jifs-192071
“Generally, structural approaches are applied to textures with repetitive patterns like fabrics and bricks [4].”
Mura Defect Detection on Compact Camera Module (CCM) Using Metric Learning
Kim, Y. G., & Park, T. H. (2020) · IEEE International Conference on Automation Science and Engineering (CASE) · doi:10.1109/case48305.2020.9216886
MASS DISTRIBUTION OF DARK MATTER HALO AND SCALE EVOLUTION OF EARLY TYPE GALAXIES
", M. O. N. S. I. T. D. O., Kairatkyzy, D., Astronomy", P. D. S. I. T. S. "., Astronomy", P. A., kairatkyzy_dina90@mail.ru, & Quevedo, H. C. (2020) · NEWS OF THE NATIONAL ACADEMY OF SCIENCES OF THE REPUBLIC OF KAZAKHSTAN · doi:10.32014/2020.2518-1726.102
Surface Defect Recognition Classification Based on Multi-Scale Convolutional Neural Network
Wang, Z., Xiao, Z., & He, Z. (2021) · Advances in intelligent systems and computing · doi:10.1007/978-981-16-1726-3_129
Multiple Paired Pixel Consistency Model for Robust Defect Detection in Printed Logotypes
Xiang, S. (2021) · Institutional Repositories DataBase (IRDB) · doi:10.14943/doctoral.k14596
Using deep learning for quality control in cyber-manufacturing
Dai, W. (2022) · DR-NTU (Nanyang Technological University) · doi:10.32657/10356/155249
2018 IndexIEEE Transactions on Instrumentation and MeasurementVol. 67
(2018) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2018.2881625
Fabric Defect Defection based on Lightweight Convolutional Denoising Auto-Encoder
Dong, J., Li, W., Sun, Z., & Li, X. (2022) · International Conference on Computing and Artificial Intelligence (ICCAI) · doi:10.1145/3532213.3532285
“Mei et al. [6] first counted the mean and variance of residual values of normal samples with different sampling scales in MSCDAE, and then calculated the threshold of the scale through formula 10 and binarized by the threshold, finally fused the segmentation results of each scale through formula 11.”
MethodInfluentialWang, X.-T., & Liu, J. (2022) · Lecture notes in computer science · doi:10.1007/978-3-031-15934-3_11
Ming, 梁. L., Minglu, 张. Z., & Lü, 吕. X. (2023) · Laser & Optoelectronics Progress · doi:10.3788/lop212927
Defect detection algorithm for image reconstruction based on the self-attentive mechanism
Fu, H. H., & Gong, J. (2023) · Frontiers in Computing and Intelligent Systems · doi:10.54097/fcis.v3i2.6485
Image Segmentation of Cell Blue Film Based on Resnet50
陶, 佳. (2023) · Software Engineering and Applications · doi:10.12677/sea.2023.122030
An Incremental Surface Defect Detection Method by Fused Unsupervised and Supervised Methods
Deng, W., Wang, W., Jie, J., & Wu, D. (2022) · doi:10.1145/3573942.3574107
“Some methods have enhanced the reconstruction ability by the way of multi-scale [1] and loss function [2].”
MethodWang, X., & Li, W. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4511134
Mtddpm: Visual Anomaly Detection with Multiscale Denoising Diffusion Probabilistic ModelPreprint
Wang, X., & Li, W. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4632278
Mtddpm: Visual Anomaly Detection with Multiscale Denoising Diffusion Probabilistic ModelPreprint
Wang, X., & Li, W. (2023) · SSRN Electronic Journal · doi:10.2139/ssrn.4654306
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.