Automatic Visual Defect Detection Using Texture Prior and Low-Rank Representation
Open accessThe DOI matches its Crossref record; 26 of the 26 references it cites were found in a registry or an index.
Abstract
Automatic surface detection for quality control has largely employed image processing techniques, for example in steel and fabric defect inspection. There are rising demands in the quality control industry for defective image analysis to fulfill its vital role in visual inspection. In this paper, we introduce an unsupervised method using a low-rank representation based on texture prior for detection of defects on natural surfaces and formulate the detection process as a novel weighted low-rank reconstruction model. The first step of the proposed method estimates the texture prior to a given image by constructing a texture prior map where higher values indicate a higher probability of abnormality. The second step of the proposed method detects the defect via low-rank decomposition with the help of the texture prior. Experiments on synthetic and real images show that the proposed method is superior in terms of detection accuracy and competitive in computational efficiency with respect to the state-of-the-art methods in surface defect detection research. This contribution is of particular interest for manufacturers (e.g., steel and fabric) for which defect detection largely relies on manual inspection.
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
“Huangpeng et al. [98] proposed a novel weighted low-rank reconstruction model for automatic visual defect detection”
From Automated Visual Defect Detection for Flat Steel Surface: A Survey (2020) · cites this paper as [98]
In the citing paper
“These approaches are reported to perform well.”
From Automated Visual Defect Detection for Flat Steel Surface: A Survey (2020) · cites this paper as [98]
Citations
Cites
What energy functions can be minimized via graph cuts?
Kolmogorov, V., & Zabih, R. (2004) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2004.1262177 · Cited by 3,127
Textons, the elements of texture perception, and their interactions
Julesz, B. (1981) · Nature · doi:10.1038/290091a0 · Cited by 1,703
Regularization methods for learning incomplete matrices
Mazumder, R., Hastie, T. J., & Tibshirani, R. J. (2009) · Journal of Machine Learning Research · doi:10.48550/arxiv.0906.2034 · Cited by 1,640
Road crack detection using deep convolutional neural network
Zhang, L., Yang, F., Daniel Zhang, Y., & Zhu, Y. J. (2016) · 2016 IEEE International Conference on Image Processing (ICIP) · doi:10.1109/icip.2016.7533052 · Cited by 1,496 · cited 1×
“Reference [15] presented a supervised deep convolutional neural network that is trained to classify each image patch, and the defective regions are identified by dilation and thresholding.”
Markov Random Field Modeling in Image Analysis
(2009) · Advances in Pattern Recognition · doi:10.1007/978-1-84800-279-1 · Cited by 823
Computer-Vision-Based Fabric Defect Detection: A Survey
Kumar, A. (2008) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.1930.896476 · Cited by 725 · cited 1×
“The literature for surface defect detection based on computer vision is vast, and the conventional methods can be divided into three categories, namely: statistical, spectral, and model-based [3].”
BackgroundSpectral Regularization Algorithms for Learning Large Incomplete Matrices
MazumderRahul, HastieTrevor, & TibshiraniRobert (2010) · Journal of Machine Learning Research · doi:10.5555/1756006.1859931 · Cited by 671 · cited 1×
“The optimization of B can be computed exactly by softimpute algorithm as described in [20], and the optimal solution to (13) can be obtained by iteratively using:”
MethodMoving Object Detection by Detecting Contiguous Outliers in the Low-Rank Representation
Zhou, X., Yang, C., & Yu, W. (2013) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2012.132 · Cited by 648 · cited 3×
“According to [18], we initialize α to be the second largest singular value of I”
BackgroundMethodInfluentialA 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 1×
“Reference [14] presented a generic”
Deformable 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 2×
“In many applications, such as [13], the patch size is chosen as 8× 8.”
BackgroundMethod
Cited by
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
“[98] proposed a novel weighted low-rank reconstruction model for automatic visual defect detection, and Zhou et al.”
MethodEbayyeh, A. A. R. M. A., & Mousavi, A. (2020) · IEEE Access · doi:10.1109/access.2020.3029127 · Cited by 310
“Furthermore, this method can only solve LCD images with simple background [352].”
MethodRecent advances in surface defect inspection of industrial products using deep learning techniques
Zheng, X., Zheng, S., Kong, Y., & Chen, J. (2021) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-021-06592-8 · Cited by 287
Fabric Defect Detection Using Computer Vision Techniques: A Comprehensive Review
Rasheed, A., Zafar, B., Rasheed, A., Ali, N., Sajid, M., Dar, S. H., Habib, U., Shehryar, T., & Mahmood, M. T. (2020) · Mathematical Problems in Engineering · doi:10.1155/2020/8189403 · Cited by 205
“Accuracy (Acc), F-measure, true-positive rate, false-positive rate, and positive predictive value have been used to quantify the detection accuracy [58].”
MethodInfluentialReview of vision-based defect detection research and its perspectives for printed circuit board
Zhou, Y., Yuan, M., Zhang, J., Ding, G., & Qin, S. (2023) · Journal of Manufacturing Systems · doi:10.1016/j.jmsy.2023.08.019 · Cited by 193
Research Progress of Automated Visual Surface Defect Detection for Industrial Metal Planar Materials
Fang, X., Luo, Q., Zhou, B., Li, C., & Tian, L. (2020) · Sensors · doi:10.3390/s20185136 · Cited by 160
“FPR = 0.010 FNR = 0.004 Time (sec) = 0.195 [106] Low-rank sparse reconstruction model The method detects the defect via low-rank decomposition with the help of the texture prior, which is estimated by constructing a texture prior map on the given images where higher values indicate a higher…”
MethodInfluentialSurface Defect Detection via Entity Sparsity Pursuit With Intrinsic Priors
Wang, J., Li, Q., Gan, J., Yu, H., & Yang, X. (2020) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2019.2917522 · Cited by 141
“Specifically, images in both [13] and [25] are divided into uniform-sized patches rather than varisized superpixels.”
BackgroundMethodResultInfluentialJeyaraj, P. R., & Nadar, E. R. S. (2019) · International Journal of Clothing Science and Technology · doi:10.1108/ijcst-11-2018-0135 · Cited by 112
Automatic fabric defect detection using a wide-and-light network
Wu, J., Le, J., Xiao, Z., Zhang, F., Geng, L., Liu, Y., & Wang, W. (2021) · Applied Intelligence · doi:10.1007/s10489-020-02084-6 · Cited by 87
“In recent years, machine vision-based object detection methods have been widely used in the field of fabric defect detection [3, 4].”
MethodSurface Defect Detection Using Image Pyramid
Xiao, L., Wu, B., & Hu, Y. (2020) · IEEE Sensors Journal · doi:10.1109/jsen.2020.2977366 · Cited by 88
“These methods can be divided into two parts, modern image processing methods [7]–[9] and machine learning techniques [10]–[14].”
MethodZhou, X., Fang, H., Fei, X., Shi, R., & Zhang, J. (2021) · IEEE Access · doi:10.1109/access.2021.3124814 · Cited by 70
Shi, B., Liang, J., Lan, D., Chen, C., & Hou, Z. (2020) · Information Sciences · doi:10.1016/j.ins.2020.08.100 · Cited by 60
Weighted Double-Low-Rank Decomposition With Application to Fabric Defect Detection
Mo, D., Wong, W. K., Lai, Z., & Zhou, J. (2020) · IEEE Transactions on Automation Science and Engineering · doi:10.1109/tase.2020.2997718 · Cited by 59
Fabric defect detection based on information entropy and frequency domain saliency
Liu, G., & Zheng, X. (2020) · The Visual Computer · doi:10.1007/s00371-020-01820-w · Cited by 60
Xu, C., & Zhu, G. (2020) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-020-01570-5 · Cited by 50
“Huangpeng et al. (2018) proposed a low-rank representation method based on texture prior to the detection of natural surface defects and described the detection process as a novel weighted low-rank reconstruction model.”
MethodMaeda, K., Takahashi, S., Ogawa, T., & Haseyama, M. (2019) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.12451 · Cited by 42
“Although cracks are one of the most important types of distress, methods for detecting other types of distress have also been studied (Poskus et al., 2018; Huangpeng et al., 2018).”
MethodA deep learning-based approach for the automated surface inspection of copper clad laminate images
Zheng, X., Chen, J., Wang, H., Zheng, S., & Kong, Y. (2020) · Applied Intelligence · doi:10.1007/s10489-020-01877-z · Cited by 36
Fabric Defect Detection via Low-Rank Decomposition With Gradient Information
Shi, B., Liang, J., Lan, D., Chen, C., & Hou, Z. (2019) · IEEE Access · doi:10.1109/access.2019.2939843 · Cited by 26
“[29] introduced a weighted low rank reconstruction model called W-LRR based on GHOG, where the defect prior texture features are constructed to guide model decomposition.”
MethodAutomated Product Boundary Defect Detection Based on Image Moment Feature Anomaly
Peng, Y., Ruan, S., Cao, G., Huang, S., Kwok, N. M., & Zhou, S. (2019) · IEEE Access · doi:10.1109/access.2019.2911358 · Cited by 25
“To this end, machine vision technology has been widely adopted for automatic surface defect detection [5]–[8].”
MethodTextile defect detection using multilevel and attentional deep learning network (MLMA-Net)
Wei, B., Xu, B., Hao, K., & Gao, L. (2022) · Textile Research Journal · doi:10.1177/00405175211073773 · Cited by 26
Texture Representation Through Overlapped Multi-Oriented Tri-Scale Local Binary Pattern
Fawad, F., Khan, M. J., Riaz, M. A., Shahid, H., Khan, M. S., Amin, Y., Loo, J., & Tenhunen, H. (2019) · IEEE Access · doi:10.1109/access.2019.2918004 · Cited by 24
“Agricultural produce such as herbs [3], fruits [4], leaves [5], rocks [6], mud [7], and wood [8], as well as industrial products including marble [9], and fabric [10] possess unique textural characteristics that provide basis for categorization into respective classes.”
BackgroundMeeradevi, T., Sasikala, S., Gomathi, S., & Prabakaran, K. J. G. (2022) · Multimedia Tools and Applications · doi:10.1007/s11042-022-13575-8 · Cited by 21
“Huangpeng et al. (2018) have proposed an automatic identification of visual defects using prior and low ranking textures with a detection accuracy of 99%.”
BackgroundMethodInfluentialDefective texture classification using optimized neural network structure
Mohammed, K. M. C., Kumar, S. S., & Prasad, G. S. (2020) · Pattern Recognition Letters · doi:10.1016/j.patrec.2020.04.017 · Cited by 21
Object Detection Algorithm for Surface Defects Based on a Novel YOLOv3 Model
Lv, N., Xiao, J. M., & Qiao, Y. (2022) · Processes · doi:10.3390/pr10040701 · Cited by 24
Real-time defect detection and classification on wood surfaces using deep learning
Mohsin, M., Balogun, O. S., Haataja, K., & Toivanen, P. (2022) · Electronic Imaging · doi:10.2352/ei.2022.34.10.ipas-382 · Cited by 19
Anami, B. S., & Elemmi, M. C. (2021) · Journal of the Textile Institute · doi:10.1080/00405000.2021.1915559 · Cited by 19
Tao, X., Yan, S., Gong, X., & Adak, C. (2022) · IEEE Transactions on Artificial Intelligence · doi:10.1109/tai.2022.3227142 · Cited by 14
“An increasing number of automated visual inspection methods are applied to the defect detection of textured surfaces, such as wooden surfaces [51], textured fabrics [52], ceramic tiles [53], photovoltaic cells [54], steel surfaces [55].”
MethodFast and efficient computing for deep learning-based defect detection models in lightweight devices
Fișne, A., Kalay, A., & Eken, S. (2024) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-024-02487-z · Cited by 15
“Huangpeng et al. (2018) develop a weighted model with low-rank reconstruction for the unsupervised defect detection process on natural surfaces.”
Defective and nondefective classif ication of fabric images using shallow and deep networks
Elemmi, M. C., Anami, B. S., & Malvade, N. N. (2021) · International Journal of Intelligent Systems · doi:10.1002/int.22774 · Cited by 13
“Qizi et al.7 have presented a paper on automatic fabric defect detection using texture and low‐rank representation.”
BackgroundInfluentialConvolutional neural networks for real-time wood plank detection and defect segmentation
Mohsin, M., Balogun, O. S., Haataja, K., & Toivanen, P. (2023) · F1000Research · doi:10.12688/f1000research.131905.1 · Cited by 11
AI-Driven Toolbox for Efficient and Transferable Visual Quality Inspection in Production
Trampert, P., Mantowsky, S., Schmidt, F., Masiak, T., & Schneider, G. (2025) · SN Computer Science · doi:10.1007/s42979-025-03988-1 · Cited by 10
“Predominantly, VQI relies on specialized camera systems that come with built-in or attached traditional image processing capabilities [9, 15].”
Chen, S., Toyoura, M., Terada, T., Mao, X., & Xu, G. (2020) · Integrated Computer-Aided Engineering · doi:10.3233/ica-200647 · Cited by 10
“[22] constructed a weighted, low-rank, representational model of textures and detected defects.”
BackgroundOptimal surface defect detector design based on deep learning for 3D geometry
Suh, S. (2025) · Scientific Reports · doi:10.1038/s41598-025-88112-2 · Cited by 9
CACFNet: Fabric defect detection via context-aware attention cascaded feedback network
Liu, Z., Tian, B., Li, C., Ding, S., & Xi, J. (2023) · Textile Research Journal · doi:10.1177/00405175231151439 · Cited by 7
Detection of Stain Defects in Textile Industry using State-of-Art Transfer Learning Models
Karegowda, A. G., Pooja, R., Rani, A. L., & Devika, G. (2024) · International Conference on Smart Systems for Applications in Electrical Sciences (ICSSES) · doi:10.1109/icsses62373.2024.10561384 · Cited by 6
Fabric defect detection via low-rank decomposition with multi-priors and visual saliency features
Di, L., Long, H., Shi, B., Xia, Y., & Liang, J. (2024) · Journal of the Franklin Institute · doi:10.1016/j.jfranklin.2024.107150 · Cited by 6
Surface Defect Inspection in Images Using Statistical Patches Fusion and Deeply Learned Features
Chian, Y. T. E., & Tian, J. (2021) · AI · doi:10.3390/ai2010002 · Cited by 6
“A low-rank representation based on texture prior for detection of defects on natural surfaces is developed in [11].”
BackgroundDouble Low-rank Based Matrix Decomposition for Surface Defect Segmentation of Steel Sheet
Zhou, S., Wu, S., Cui, K., & Liu, H. (2021) · ISIJ International · doi:10.2355/isijinternational.isijint-2021-024 · Cited by 6
Zhou, S., Liu, H., Cui, K., & Hao, Z. (2021) · IEEE Access · doi:10.1109/access.2021.3117736 · Cited by 7
“The proposed DLMD is compared with five representative segmentation methods quantitatively and qualitatively, including RPCA [27], SSD [32], PG-LSR [33], W-LRR [34], and ESP [35]. that most of methods can handle simple defect images with relatively homogeneous background (i.e., column 5, and 10).”
BackgroundMethodLi, C., Ban, J., Liu, Z., Dong, Y. X., Tang, S., & Weng, D. (2021) · Journal of Engineered Fibers and Fabrics · doi:10.1177/15589250211008453 · Cited by 8
“Qizi et al. 27 proposed a low-rank decomposition detection method based on image texture prior information.”
MethodResearch of express box defect detection based on machine vision
Wei, L., Zhang, N., Xue, M., & Huo, J. (2020) · International Conference on Multimedia and Image Processing (ICMIP) · doi:10.1145/3381271.3381280 · Cited by 4
“Huang, et al. [9] pointed out a novel reconstruction model to detect the defects.”
BackgroundProgress in Fabric Defect Detection Based on Machine Learning
Wu, Y., Li, D., Guo, P., & Liu, Y. (2025) · Journal of Shanghai Jiaotong University (Science) · doi:10.1007/s12204-025-2804-x · Cited by 3
Schadt, M., Mai, C., & Buettner, R. (2025) · IEEE Access · doi:10.1109/access.2025.3615795 · Cited by 3
A novel representation of periodic pattern and its application to untrained anomaly detection
Ye, P., Tao, C., & Du, J. (2025) · IISE Transactions · doi:10.1080/24725854.2025.2526776 · Cited by 3
“The anomalies in periodic textures can be regarded as sparse matrices (Huangpeng et al., 2018).”
Visual Locating of Reactor in an Industrial Environment Using the Composite Method
Cao, C., Ouyang, Q., Hou, J., & Zhao, L. (2020) · Sensors · doi:10.3390/s20020504 · Cited by 2
“The visual location system is similar to a visual measurement system [3] and a visual detection system [4], and has a higher efficiency and lower error rate than a manual operation.”
MethodDetecting aluminium tube surface defects by using faster region-based convolutional neural networks
Chen, S., Wang, D.-G., & Wang, F.-B. (2022) · Journal of Computational Methods in Sciences and Engineering · doi:10.3233/jcm-226107 · Cited by 2
“Zi et al. [4] proposed an unsupervised natural surface defect detection method on the basis of low-order representations for texture, in which detection process was considered to be a new weighted low-order reconstruction model.”
SACNN: Spatial Adversarial Convolutional Neural Network for Textile Defect Detection
Hou, W. S., Tao, X., Ma, W., & Xu, D. (2020) · Fibres and Textiles in Eastern Europe · doi:10.5604/01.3001.0014.3808 · Cited by 1
“introduced an unsupervised method using a low-rank representation based on texture prior to the detection of defects on natural surfaces and formulated the detection process as a novel weighted low-rank reconstruction model [2].”
MethodMulti-Modal Defect Detection System for Single Color Fabrics in the Apparel Industry
Silva, V. D., Senevirathne, T., Fareed, N., Sandanayake, T. C., & Karunaratne, I. (2023) · International Conference on Information Technology Research (ICITR) · doi:10.1109/icitr61062.2023.10382928 · Cited by 1
“The combination of these technologies allows to develop a robust and efficient defect detection algorithm for the textile industry [7][8].”
BackgroundWu, Q., Ma, Z., Chen, W., Li, C., & Hu, C. (2020) · IEEE Access · doi:10.1109/access.2020.2981371
Research on Steel Surface Defect Algorithm Based on Deep Residual Network
Jin, G., Hong, R., Lin, X., & Liu, Y. (2022) · International Conference on Artificial Intelligence and Advanced Manufacture (AIAM) · doi:10.1109/aiam57466.2022.00068
“used the lowrank representation method to determine the optimal kernel function to improve the accuracy and speed of defect detection [5].”
MethodLiu, F., Cao, W., Ye, Y., Gu, F., Long, S., & Song, Z. P. (2023) · IEEE Access · doi:10.1109/access.2023.3329116
“For the feature domain-based methods, saliency defect detection [8], [9], [10], [11] has outstanding performance as the main representative.”
MethodAnchor selection based deep learning two stage fabric defect localization
Pooja, H., & Soma, S. (2024) · IAES International Journal of Artificial Intelligence · doi:10.11591/ijai.v13.i4.pp4711-4721
“Huangpeng et al. [13] proposed an unsupervised method using texture prior for detection of defects.”
MethodQu, G., Liao, J., Liu, K., Xu, B., & Qian, Y. (2025) · Machines · doi:10.3390/machines13121107
Metadata
- Authors
- Huangpeng, Q., Zhang, H., Zeng, X., & Huang, W.
- Affiliations
- National University of Defense Technology (CN)
- University of Alberta (CA)
- Venue
- IEEE Access
- Published
- 2018
- Volume
- 6
- Pages
- 37965-37976
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W2831321715
- Semantic Scholar
- 6d32ac948e7f3034c83115eaab634e792ed44e9d
- Cited by
- 53
- References
- 25
- Citations by year
- 2025: 6 · 2024: 3 · 2023: 5 · 2022: 7 · 2021: 10 · 2020: 16 · 2019: 6
- FWCI
- 6.61
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- IEEE Access (publishedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Funders
- National Natural Science Foundation of China (61602494)
- National University of Defense Technology (ZK16-03-16)
- Topics
- Physical Sciences › Engineering › Industrial and Manufacturing Engineering › Industrial Vision Systems and Defect Detection
- Physical Sciences › Engineering › Civil and Structural Engineering › Infrastructure Maintenance and Monitoring
- Physical Sciences › Engineering › Mechanical Engineering › Non-Destructive Testing Techniques
- Keywords
- surface defect detection, low-rank representation, unsupervised anomaly detection, natural surfaces, quality control, low-rank decomposition
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.