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.”
BackgroundA Review of Recent Advances in Surface Defect Detection using Texture analysis Techniques
Xie, X. (2008) · ELCVIA Electronic Letters on Computer Vision and Image Analysis · doi:10.5565/rev/elcvia.268 · Cited by 500 · cited 1×
“1) statistical methods; 2) structural methods; 3) filter-based methods; 4) model-based methods [7].”
A Generic Deep-Learning-Based Approach for Automated Surface Inspection
Ren, R., Hung, T., & Tan, K. C. (2018) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2017.2668395 · Cited by 485 · cited 2×
“[22] proposed a supervised convolutional neural network (CNN) architecture for image”
BackgroundLiou, C.-Y., Cheng, W.-C., Liou, J.-W., & Liou, D.-R. (2014) · Neurocomputing · doi:10.1016/j.neucom.2013.09.055 · Cited by 467 · cited 1×
“1) Why CDAE: The CAE network is commonly used for unsupervised learning of efficient codings [26].”
BackgroundAdaptive Stochastic Gradient Descent Optimisation for Image Registration
Klein, S., Pluim, J. P. W., Staring, M., & Viergever, M. A. (2009) · International Journal of Computer Vision · doi:10.1007/s11263-008-0168-y · Cited by 387 · cited 1×
“Stochastic gradient descent algorithm [30] can be easily applied for optimization of CAE and CDAE models.”
MethodMinimizing Computation in Convolutional Neural Networks
Cong, J., & Xiao, B. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-11179-7_36 · Cited by 372 · cited 2×
“But the practical computational workload is more inclined to the hardware accelerators [46].”
BackgroundMachine learning-based imaging system for surface defect inspection
Park, J.-K., Kwon, B.-K., Park, J.-H., & Kang, D.-J. (2016) · International Journal of Precision Engineering and Manufacturing-Green Technology · doi:10.1007/s40684-016-0039-x · Cited by 311 · cited 1×
“[24] also proposed novel deep CNN architectures, which obviously improve the accuracy of automated defect detection.”
BackgroundDeformable Patterned Fabric Defect Detection With Fisher Criterion-Based Deep Learning
Li, Y., Zhao, W., & Pan, J. (2017) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2016.2520955 · Cited by 265 · cited 1×
“[25] proposed a Fisher criterion-based stacked denoising autoencoder (AE) model for deformable patterned fabric defect detection.”
BackgroundTEXTURE ANALYSIS WITH LOCAL BINARY PATTERNS
Mäenpää, T., & Pietikäinen, M. (2005) · Handbook of Pattern Recognition and Computer Vision · doi:10.1142/9789812775320_0011 · Cited by 244 · cited 1×
“Gray-level co-occurrence matrix-based [8] and local descriptor-based [9] methods are typical statistical texture analysis means.”
MethodIllumination Normalization Based on Weber's Law With Application to Face Recognition
Wang, B., Li, W., Yang, W., & Liao, Q. (2011) · IEEE Signal Processing Letters · doi:10.1109/lsp.2011.2158998 · Cited by 239 · cited 1×
“In order to reduce the rate of false detection, an illumination normalization process based on Weber’s law [31] is conducted first of all.”
MethodAutomated surface inspection for statistical textures
Tsai, D.-M., & Huang, T.-Y. (2003) · Image and Vision Computing · doi:10.1016/s0262-8856(03)00007-6 · Cited by 217
Wood inspection with non-supervised clustering
Silv�n, O., Niskanen, M., & Kauppinen, H. (2003) · Machine Vision and Applications · doi:10.1007/s00138-002-0084-z · Cited by 164
Assessment of surface quality on textured FDM prototypes
Armillotta, A. (2006) · Rapid Prototyping Journal · doi:10.1108/13552540610637255 · Cited by 163 · cited 1×
“uses machine vision techniques for industrial defect inspection and has drawn much attention in recent years, especially for textured surface inspection [4]–[6].”
BackgroundSharma, N., Ray, A., Sharma, S., Shukla, K., Pradhan, S., & Aggarwal, L. (2008) · Journal of Medical Physics · doi:10.4103/0971-6203.42763 · Cited by 161 · cited 1×
“[12] proposed a semisupervised approach for segmentation and classification of medical images and achieved superior performance.”
BackgroundTEXEMS: Texture Exemplars for Defect Detection on Random Textured Surfaces
Xie, X., & Mirmehdi, M. (2007) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2007.1038 · Cited by 159 · cited 6×
“of the pyramid architecture [20]; therefore, this parameter is fixed throughout our experiments.”
BackgroundResultInfluentialA new texture descriptor using multifractal analysis in multi-orientation wavelet pyramid
Xu, Y., Yang, X., Ling, H., & Ji, H. (2010) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2010.5540217 · Cited by 153 · cited 1×
“This phenomenon is in accordance with that in [20] and [27].”
ResultWavelet Domain Multifractal Analysis for Static and Dynamic Texture Classification
Ji, H., Yang, X., Ling, H., & Xu, Y. (2012) · IEEE Transactions on Image Processing · doi:10.1109/tip.2012.2214040 · Cited by 139 · cited 2×
“In general, these methods can be divided into spatial domain [13], frequency domain [14], and”
MethodNoisy training for deep neural networks in speech recognition
Yin, S., Liu, C., Zhang, Z., Lin, Y., Wang, D., Tejedor, J., Zheng, T. F., & Li, Y. (2015) · EURASIP Journal on Audio, Speech, and Music Processing · doi:10.1186/s13636-014-0047-0 · Cited by 118 · cited 3×
“It should not be used in the testing phase so as to ensure fairness of performance comparison [35].”
BackgroundMethodInfluentialDetection of local defects in textile webs using Gabor filters
Navarro, R. (1998) · Optical Engineering · doi:10.1117/1.601751 · Cited by 118
Rakun, J., Stajnko, D., & Zazula, D. (2011) · Computers and Electronics in Agriculture · doi:10.1016/j.compag.2011.01.007 · Cited by 118
Weakly Supervised Learning for Industrial Optical Inspection
Wieler, M., Hahn, T., & Hamprecht, F. A. (2007) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.8086135 · Cited by 90
Approximate real inversion formulas of the gaussian convolution
Saitoh, S. (2004) · Applicable Analysis · doi:10.1080/00036810410001657198 · Cited by 87 · cited 1×
“I (n+1) = S↓G(I ), n = 1, 2, · · · nl; I (1) = I ′ (6) where S↓ refers to the downsampling process, G(·) denotes the Gaussian convolution [33], nl is the number of pyramid levels, and I ′ is the image after illumination normalization.”
BackgroundThe phase only transform for unsupervised surface defect detection
Aiger, D., & Talbot, H. (2010) · 2010 IEEE Computer Society Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2010.5540198 · Cited by 87 · cited 5×
“12 with the DCT, LCA, PHOT, and our method are exhibited in Table II.”
InfluentialDetecting Defects in Fabric with Laser-Based Morphological Image Processing
Mallik-Goswami, B., & Datta, A. K. (2000) · Textile Research Journal · doi:10.1177/004051750007000902 · Cited by 82 · cited 1×
“This approach is commonly applied to textures with repetitive patterns such as fabrics [10] and bricks [11].”
BackgroundAutomated defect detection in textured surfaces using optimal elliptical Gabor filters
Hu, G. (2015) · Optik · doi:10.1016/j.ijleo.2015.04.017 · Cited by 79
Analysis and Synthesis of Structural Textures
Lefebvre, L., & Poulin, P. (2000) · Proceedings · doi:10.20380/gi2000.12 · Cited by 74
THE PHASE ONLY TRANSFORM FOR UNSUPERVISED SURFACE DEFECT DETECTION
Aiger, D., & Talbot, H. (2011) · Series in Computer Vision · doi:10.1142/9789814343008_0011 · Cited by 73
Improved Canny Edge Detector in Ceramic Tiles Defect Detection
Hocenski, Z., Vasilic, S., & Hocenski, V. (2006) · IECON 2006 - 32nd Annual Conference on IEEE Industrial Electronics · doi:10.1109/iecon.2006.347535 · Cited by 69 · cited 1×
“It commonly occurs in many industrial applications, such as textile inspection [1], wood inspection [2], and ceramic tile inspection [3].”
BackgroundFine-Grained Material Classification Using Micro-geometry and Reflectance
Kampouris, C., Zafeiriou, S. P., Ghosh, A., & Malassiotis, S. (2016) · Lecture notes in computer science · doi:10.1007/978-3-319-46454-1_47 · Cited by 65 · cited 2×
“In these data, the used fabric samples mainly come from the Fabrics [38] and the KTH-TIPS2 [39] data sets.”
MethodWeakly Supervised Learning for Industrial Optical Inspection
Wieler, M., Hahn, T., & Hamprecht, F. A. (2007) · Zenodo (CERN European Organization for Nuclear Research) · doi:10.5281/zenodo.8086136 · Cited by 64
Vujasinović, T., Pribic, J., Kanjer, K., Milošević, N. T., Tomasevic, Z. I., Milovanović, Z., Nikolić‐Vukosavljević, D., & Radulović, M. (2015) · Microscopy and Microanalysis · doi:10.1017/s1431927615000379 · Cited by 59 · cited 1×
“Gray-level co-occurrence matrix-based [8] and local descriptor-based [9] methods are typical statistical texture analysis means.”
MethodUnsupervised-Learning-Based Feature-Level Fusion Method for Mura Defect Recognition
Mei, S., Yang, H., & Yin, Z. (2017) · IEEE Transactions on Semiconductor Manufacturing · doi:10.1109/tsm.2017.2648856 · Cited by 56 · cited 2×
“The used LCD samples come from [36].”
MethodSelf-citationGabor wavelets combined with volumetric fractal dimension applied to texture analysis
Zuñiga, A. G., Florindo, J. B., & Bruno, O. M. (2014) · Pattern Recognition Letters · doi:10.1016/j.patrec.2013.09.023 · Cited by 54 · cited 1×
“Fourier analysis [14], wavelet transform [16], and Gabor transform [17] are commonly used”
MethodMalik, F., & Baharudin, B. (2013) · Cited by 45
Blind weave detection for woven fabrics
Schneider, D., & Merhof, D. (2014) · Pattern Analysis and Applications · doi:10.1007/s10044-014-0403-9 · Cited by 43 · cited 2×
“However, these methods are usually used together with other approaches to specific issues [16], [18], [19].”
MethodGao, B., Zhang, H., Woo, W. L., Tian, G. Y., Bai, L., & Yin, A. (2014) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2013.2287126 · Cited by 41 · cited 2×
“[21] developed a spatial-frequency feature extraction algorithm for defect detection using smooth”
MethodLin, H.-D. (2007) · Image and Vision Computing · doi:10.1016/j.imavis.2007.07.009 · Cited by 32 · cited 6×
“As can be seen that the DCT method (the third row) shows good performance mainly in obvious periodic textures, e.g., samples in Fig.”
BackgroundInfluentialA Texture-Based Image Processing Approach for the Description of Human Oocyte Cytoplasm
Basile, T. M. A., Caponetti, L., Castellano, G., & Sforza, G. (2010) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2010.2057552 · Cited by 32 · cited 1×
“However, these methods are usually used together with other approaches to specific issues [16], [18], [19].”
MethodTsai, D., & Kuo, C.-C. (2007) · Machine Vision and Applications · doi:10.1007/s00138-007-0073-3 · Cited by 21 · cited 1×
“uses machine vision techniques for industrial defect inspection and has drawn much attention in recent years, especially for textured surface inspection [4]–[6].”
BackgroundStochastic Texture Analysis for Measuring Sheet Formation Variability in the Industry
Scharcanski, J. (2006) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2006.881036 · Cited by 16 · cited 2×
“However, these methods are usually used together with other approaches to specific issues [16], [18], [19].”
MethodTotal variation flow-based multiscale framework for unsupervised surface defect segmentation
Kim, W., & Kim, C. (2012) · Optical Engineering · doi:10.1117/1.oe.51.12.127201 · Cited by 9 · cited 2×
“It commonly occurs in many industrial applications, such as textile inspection [1], wood inspection [2], and ceramic tile inspection [3].”
BackgroundMethod
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
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.