Weak Micro-Scratch Detection Based on Deep Convolutional Neural Network
Open accessThe DOI matches its Crossref record; 22 of the 22 references it cites were found in a registry or an index.
Abstract
Metal component surfaces are random textured and non-smooth. There are many stains on the surface of metal component that are similar to the gray scale of the scratches. The scratches have non-uniform gray distribution, various shapes, and low contrast in their background, posing challenges in accurate scratch detection. This paper presents a method for detecting weak scratches on metal component surfaces based on deep convolutional neural networks (DCNNs). First, a DCNN is trained using labeled scratch images. Then, the scratches and some faults are detected by the trained DCNN, and most of the faults can be removed through properly thresholding based on the size of connected regions. Finally, the scratch length united in the number of pixels is obtained by the skeleton extraction. The experimental results show that the proposed method can effectively deal with background noise, thereby achieving accurate scratch detection.
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, Song et al. [108] realized precise detection of weak scratches on metal surface by confusing deep CNN and skeleton extraction, the experimental results indicate its strong robustness to background noises.”
From Automated Visual Defect Detection for Flat Steel Surface: A Survey (2020) · cites this paper as [108]
Citations
Cites
U-Net: Convolutional Networks for Biomedical Image Segmentation
Ronneberger, O., Fischer, P., & Brox, T. (2015) · Lecture notes in computer science · doi:10.1007/978-3-319-24574-4_28 · Cited by 99,404
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
Delving Deep into Rectifiers: Surpassing Human-Level Performance on ImageNet Classification
He, K., Zhang, X., Ren, S., & Sun, J. (2015) · 2015 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2015.123 · Cited by 19,741 · cited 1×
“The experiment uses the Msra initialization method [21] to initialize the parameters.”
MethodVisualizing and understanding convolutional networks
Zeiler, M. D., & Fergus, R. (2014) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-10590-1_53 · Cited by 15,839 · cited 1×
“wider receiving domain and therefore capture global information [13].”
BackgroundFully convolutional networks for semantic segmentation
Shelhamer, E., Long, J., & Darrell, T. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2572683 · Cited by 11,898
Deep Learning‐Based Crack Damage Detection Using Convolutional Neural Networks
Cha, Y., Choi, W., & Büyüköztürk, O. (2017) · Computer-Aided Civil and Infrastructure Engineering · doi:10.1111/mice.12263 · Cited by 3,295 · cited 2×
“[15] combined deep convolutional neural network with sliding window technology to identify and detect cracks on concrete surface, and proved that the method is better than traditional Canny and Sobel edge detection algorithms.”
BackgroundMethodA fast parallel algorithm for thinning digital patterns
Zhang, T. Y., & Suen, C. Y. (1984) · Communications of the ACM · doi:10.1145/357994.358023 · Cited by 2,756 · cited 2×
“[19] T. Y. Zhang and C. Y. Suen, ‘‘A fast parallel algorithm for thinning digital patterns,’’ Commun.”
MethodLaplacian Pyramid Reconstruction and Refinement for Semantic Segmentation
Ghiasi, G., & Fowlkes, C. C. (2016) · Lecture Notes in Computer Science · doi:10.1007/978-3-319-46487-9_32 · Cited by 502
Steel defect classification with Max-Pooling Convolutional Neural Networks
Masci, J., Meier, U., Ciresan, D., Schmidhuber, J., & Fricout, G. (2012) · The 2012 International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn.2012.6252468 · Cited by 345 · cited 1×
“[14] proposed a maximum pooled convolutional neural network method for supervising the classification of steel defects.”
MethodTexture segmentation using wavelet transform
Arivazhagan, S., & Ganesan, L. (2003) · Pattern Recognition Letters · doi:10.1016/j.patrec.2003.08.005 · Cited by 198 · cited 1×
“Texture features extracted from wavelet decomposition images are widely used for texture classification and segmentation [8], [9].”
Method
Research works printed but not linked: 1
- truong, 2017, Automatic image thresholding using Otsu’s method and entropy weighting scheme for surface defect detection, Soft Comput, 2, 4197
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
“[108] realized precise detection of weak scratches on metal surface by confusing deep CNN and skeleton extraction, the experimental results indicate its strong robustness to background noises.”
BackgroundVisual-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
An efficient lightweight convolutional neural network for industrial surface defect detection
Zhang, D., Hao, X., Wang, D., Qin, C., Zhao, B., Liang, L., & Liu, W. (2023) · Artificial Intelligence Review · doi:10.1007/s10462-023-10438-y · Cited by 208
Steel Surface Defect Recognition: A Survey
Wen, X., Shan, J., He, Y., & Song, K. (2022) · Coatings · doi:10.3390/coatings13010017 · Cited by 210
A Survey on Unsupervised Anomaly Detection Algorithms for Industrial Images
Cui, Y., Liu, Z., & Lian, S. (2023) · IEEE Access · doi:10.1109/access.2023.3282993 · Cited by 207
A Survey on Unsupervised Anomaly Detection Algorithms for Industrial Images
Cui, Y., Liu, Z., & Lian, S. (2022) · arXiv · doi:10.48550/arxiv.2204.11161 · Cited by 154
A novel deep convolutional neural network algorithm for surface defect detection
Zhang, D., Hao, X., Liang, L., Liu, W., & Qin, C. (2022) · Journal of Computational Design and Engineering · doi:10.1093/jcde/qwac071 · Cited by 134
TLU-Net: A Deep Learning Approach for Automatic Steel Surface Defect Detection
Damacharla, P., V., A. R. M., Ringenberg, J., & Javaid, A. Y. (2021) · International Conference on Applied Artificial Intelligence (ICAPAI) · doi:10.1109/icapai49758.2021.9462060 · Cited by 125
CenterNet-based defect detection for additive manufacturing
Wang, R., & Cheung, C. F. (2021) · Expert Systems with Applications · doi:10.1016/j.eswa.2021.116000 · Cited by 80
Inspection of Imprint Defects in Stamped Metal Surfaces Using Deep Learning and Tracking
Block, S. A. B., da Silva, R. D., Dorini, L. B., & Minetto, R. (2020) · IEEE Transactions on Industrial Electronics · doi:10.1109/tie.2020.2984453 · Cited by 84
A Review on Industrial Surface Defect Detection Based on Deep Learning Technology
Qi, S., Yang, J., & Zhong, Z. (2020) · International Conference on Machine Learning and Machine Intelligence (MLMI) · doi:10.1145/3426826.3426832 · Cited by 86
“In addition to the above applications, from the current relevant literature, deep learning has also been applied to electronic component defect inspection [41], diode chip defect inspection [42], optical fiber defect inspection [43], metal surface defect inspection [44], wheel hub surface defect inspection [45], bottl…”
BackgroundA Light-Weight Deep-Learning Model with Multi-Scale Features for Steel Surface Defect Classification
Liu, Y., Yuan, Y., Balta, C., & Liu, J. (2020) · Materials · doi:10.3390/ma13204629 · Cited by 65
“Song et al. [22] proposed a method using a CNN to detect weak scratches and achieved high-performance with noisy images.”
BackgroundMethodImage data assessment approach for deep learning-based metal surface defect-detection systems
Lin, H.-I., & Wibowo, F. S. (2021) · IEEE Access · doi:10.1109/access.2021.3068256 · Cited by 43
“Song et al. [23] developed an FCN based on a U-net framework to recognize minute surface defects with uneven grayscale distributions on metal plates and used two white LED bars as the lighting system.”
MethodAboulhosn, Z., Musamih, A., Salah, K., Jayaraman, R., Omar, M., & Aung, Z. (2024) · IEEE Access · doi:10.1109/access.2024.3430113 · Cited by 43
“Song et al. [16] presented work that focuses on image segmentation of weak scratches using a Deep CNN DU-Net, with an Intersection over Union (IoU) score of 0.8125.”
MethodInfluentialIndustrial Weak Scratches Inspection Based on Multifeature Fusion Network
Tao, X., Zhang, D., Hou, W. S., Ma, W., & Xu, D. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.3025642 · Cited by 30
“Song et al. [8] designed an approach for inspecting weak scratches on the metal component surface based on a modified U-net [18], which doubles the feature channel and simplifies the original network convolution structure.”
BackgroundMethodCombining Prior Knowledge With CNN for Weak Scratch Inspection of Optical Components
Hou, W. S., Tao, X., & Xu, D. (2020) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2020.3011299 · Cited by 27
“Song et al. [16] used a UNet to detect weak scratches on metal component surfaces, and precise results are obtained via the skeleton extraction.”
BackgroundMethodA Lightweight Network for Defect Detection in Nickel-Plated Punched Steel Strip Images
Liang, Y., Li, J., Zhu, J., Du, R., Wu, X., & Bing-quan, C. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3237844 · Cited by 32
“Moreover, CNNs have been successfully applied for defect detection in steel strips [4], [5], [6], [7], [8], [9], [10], [11], [12], [13], health prediction of power batteries [14], and estimation of the state-of-charge for lithium-ion batteries [15], [16].”
MethodDeep Learning Based Automated Inspection of Weak Microscratches in Optical Fiber Connector End-Face
Mei, S., Cai, Q., Gao, Z., Hu, H., & Wen, G. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3059105 · Cited by 24
“Song et al. [5] realized the scratch detection on metal surface using the improved U-Net [31].”
BackgroundMethodWang, C., Zhou, Z., & Chen, Z. (2022) · IEEE Access · doi:10.1109/access.2022.3203198 · Cited by 28
Zhao, Z., Hu, B.-T., Feng, Y.-X., Zhao, B., Yang, C., Hong, Z.-X., & Tan, J.-R. (2022) · International Journal of Production Research · doi:10.1080/00207543.2022.2138613 · Cited by 22
“As one of the most widely used deep learning-based methods, the convolutional neural network (CNN) provides an end-to-end structure by integrating feature extraction and image classification (Song et al. 2019).”
MethodMeDERT: A Metal Surface Defect Detection Model
Wang, C., & Xie, H. (2023) · IEEE Access · doi:10.1109/access.2023.3262264 · Cited by 21
“3) deep learning [7], [8], [9].”
Yang, Z., Zhang, M., Li, C., Meng, Z., Li, Y., Chen, Y., & Liu, L. (2022) · IEEE Access · doi:10.1109/access.2022.3178380 · Cited by 18
“CNN has been widely used in machine vision, especially in the fields of image recognition and image classification [4].”
MethodDevelopment of Machine Vision System for Off-Line Inspection of Fine Defects on Glass Screen Surface
Yang, W., Zhang, Y., Dong, Y., Xu, D., & Pan, T. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3190052 · Cited by 15
Üzen, H., Türkoğlu, M., & Hanbay, D. (2022) · Neural Computing and Applications · doi:10.1007/s00521-022-07885-z · Cited by 17
“The proposed SFEDNet and MFE-DEDNet were compared with the Unet [29, 30, 58] and Mobile-Net [59, 60] networks commonly used in the literature.”
MethodResultFeng, Q., Li, F., Li, H., Liu, X., Fei, J., Xu, S., Lu, C. B., & Yang, Q. (2023) · The Visual Computer · doi:10.1007/s00371-023-03056-w · Cited by 14
“Image-segmentation methods have been applied to the field of defect detection [4, 28–31].”
Hao, X., Dong, T., & Zhang, D. (2021) · International Conference on Robotics and Automation Engineering (ICRAE) · doi:10.1109/icrae53653.2021.9657815 · Cited by 11
A High-Accuracy Calibration Method for a Telecentric Structured Light System
Chen, C., Kong, Y., Wang, H., & Zhang, Z. (2022) · Sensors · doi:10.3390/s22176370 · Cited by 11
“Comprehensively considering the pros and cons, we deduce that the proposed method is quite suitable for high-accuracy 3D measurements of small-scale objects [22] and has great potential in 3D shape measurement of micro-parts with a size on the order of millimeters, used in micro-parts mechanical system (MEMS) [23].”
MethodAutofeeding System for Assembling the CBCs on Automobile Engine Based on 3-D Vision Guidance
Li, J., Du, R., Zhang, J., Zhu, J., Xu, H., & Cai, M. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3096278 · Cited by 12
A Novel Scratch Detection and Measurement Method for Automotive Stamping Parts
Liu, C., An, B., Hou, Y., Wang, H., & Liu, Y. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3193970 · Cited by 8
Nguyêñ, H. T., Tsao, Y.-M., & Wang, H. (2022) · Applied Optics · doi:10.1364/ao.461380 · Cited by 7
Application of machine vision image feature recognition in 3D map construction
Wang, P., Liu, N., & Qiao, J. (2022) · Alexandria Engineering Journal · doi:10.1016/j.aej.2022.11.007 · Cited by 7
Defect Detection for Mobile Phone Cases Based on Improved Yolo Model
Wen, S., Tao, Y., & Chen, J. (2021) · International Conference on Computing and Artificial Intelligence (ICCAI) · doi:10.1145/3467707.3467712 · Cited by 6
An Image Quality Assessment Method for Surface Defect Inspection
Lin, H.-I., & Lin, P.-Y. (2020) · IEEE International Conference on Artificial Intelligence Testing (AITest) · doi:10.1109/aitest49225.2020.00008 · Cited by 6
“In recent years, methods directly recognizing and locating defects have been proposed, such as AE (Autoencoder) [10], FCN (Fully Convolutional Networks) [11], [12], etc.”
MethodXiao, J., Zhang, F., Qiu, F., & Wang, J. (2025) · Journal of Magnetism and Magnetic Materials · doi:10.1016/j.jmmm.2025.173413 · Cited by 4
Innovative UNET-Based Steel Defect Detection Using 5 Pretrained Models
Yulianto, R., Faqihudin, F., Rusli, M. S., Karno, A. S. B., Hastomo, W., Kardian, A. R., Terisia, V., & Surawan, T. (2023) · Evergreen · doi:10.5109/7160923 · Cited by 11
Nam, C., & Lee, S. J. (2026) · Measurement · doi:10.1016/j.measurement.2026.121432 · Cited by 3
LAD-YOLO: A Lightweight YOLOv5 Network for Surface Defect Detection on Aluminum Profiles
Zhao, D., Liu, S., Chen, Y., Chen, D., Hu, Z., & Tang, L. (2023) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2023.0140924 · Cited by 3
Machine Learning‐Based Automatic Scratch Performance Quantification on Polymeric Surfaces
Khatri, S., Li, Z., & Sue, H. (2025) · Journal of Applied Polymer Science · doi:10.1002/app.57456 · Cited by 3
GFU-Net: A Deep Learning Approach for Automatic Metal Crack Detection
Zhang, Y., Li, X., Qiu, J., Zhai, X., & Wei, M. (2021) · Communications in computer and information science · doi:10.1007/978-981-16-5188-5_27 · Cited by 4
Relational-based transfer learning for automatic optical inspection based on domain discrepancy
Salgado, E. I. V., Yan, H., Hong, Y., Zhu, P., Zhu, S., Liao, C., Wen, Y., Li, X., Qian, X., Wang, X., et al. (2022) · Optoelectronic Imaging and Multimedia Technology · doi:10.1117/12.2644087 · Cited by 3
Xiao, Y., Cai, Z., & Yuan, X. (2019) · IEEE Access · doi:10.1109/access.2019.2949269 · Cited by 3
Metal Surface Defect Detection Based on Weighted Fusion
Wang, C., Xu, J., Liang, X., & Yin, D. (2020) · International Conference on Virtual Reality and Visualization · doi:10.1109/icvrv51359.2020.00044 · Cited by 3
Zero-sample surface defect detection and classification based on semantic feedback neural networkPreprint
Guo, Y., Fan, Y., Xiang, Z., Wang, H., Meng, W., & Xu, M. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2106.07959 · Cited by 5
“improved the U-Net framework and proposed DU-Net[14] based on the Deep Convolutional Neural Network (DCNN), using two white light sources as the lighting system, and deconvolution instead of upsampling so that the model can extract more features.”
MethodWTAYOLO: A Detection Model for Weak Scratches on Metal Surfaces
Xie, W., Pan, W. J., Li, S., Zeng, Q., & Li, H. (2024) · International Conference on Sensors, Signal and Image Processing (SSIP) · doi:10.1145/3725949.3725956 · Cited by 2
“This paper compares the proposed method with five distinct scratch detection algorithms, including the Otsu [24] method, UNet [9] network, and YOLO network.”
Unsupervised machine learning to analyze corneal tissue surfaces
Rickert, C. A., Henkel, F., & Lieleg, O. (2023) · APL Machine Learning · doi:10.1063/5.0159502 · Cited by 2
Steel Surface Defects Detection Based on Improved Faster R-CNN
Xu, Y., Yang, S., Zhang, X., & Xie, Z. (2022) · International Conference on Control and Robotics (ICCR) · doi:10.1109/iccr55715.2022.10053878 · Cited by 5
A deep transfer learning-based algorithm for concrete surface defect detection
Jin, Z., & Wang, L. (2024) · Engineering Research Express · doi:10.1088/2631-8695/ad5e34 · Cited by 1
Detection of weak micro-scratches by semantic segmentation
Nguyêñ, H. T., Tsao, Y.-M., & Wang, H. (2022) · International Conference on Engineering and Emerging Technologies (ICEET) · doi:10.1109/iceet56468.2022.10007384 · Cited by 1
“[6] proposed a designed model based on the U-net structure.”
MethodResearch 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
“At first, feed forward networks were used in research [1], but soon more attention was paid to convolutional networks [2][3][4], which are more suitable for image processing, since they significantly reduce the computational load.”
MethodGuo, J., Han, C., Ma, Y., & Zhang, C. (2024) · IISE Transactions · doi:10.1080/24725854.2024.2425292 · Cited by 2
“Compared with traditional statistical or machine learning methods, they can capture more complex anomaly features and be applied for more advanced anomaly detection (Masci et al ., 2012; Di et al ., 2019) and localization tasks (Song et al ., 2019).”
ECFN: efficient cross-scale fusion network for fast semantic segmentation of steel surface defects
Wang, C., Wang, M., & Yang, H. (2025) · Engineering Research Express · doi:10.1088/2631-8695/ae28b5 · Cited by 1
A Review of Computer Vision Techniques in the Detection of Metal Failures
Fitzgerald, D., & Fragoudakis, R. (2021) · MATEC Web of Conferences · doi:10.1051/matecconf/202134902021 · Cited by 1
Detection and Recognition of Shallow and Fine Scratches on Mobile Phone Screens via CGM-YOLOv8Preprint
Wang, Y., Wang, T., Xie, G., Zhao, Z., Wang, J.-A., & Hu, X. (2025) · Research Square · doi:10.21203/rs.3.rs-6181822/v1 · Cited by 1
IMPROVED-SDROM filtering for scratches removal from images
Hamdani, M., & Belbachir, M. F. (2020) · Multimedia Tools and Applications · doi:10.1007/s11042-020-09224-7
Dynamic Light Settings as Data Augmentations for Automated Scratch Detection
Valentin, G., Fukuda, O., Yeoh, W. L., Hiroshi, O., & Nobuhiko, Y. (2022) · International Journal of Advanced Computer Science and Applications · doi:10.14569/ijacsa.2022.01312107
“Several studies also aimed at facilitating the examination of surfaces, but their experiments often did not take the orientation of the light into consideration [12], [13], [14].”
BackgroundSurface crack detection method of winding inductor under complex background
Chen, Y., Zhang, Z., & Long, Y. (2024) · International Conference on Computer Vision, Image and Deep Learning (CVIDL) · doi:10.1109/cvidl62147.2024.10604171
Wang, T. (2024) · Civil-comp conferences · doi:10.4203/ccc.7.8.4
Human Interfaces with Machine Learning Recognition Systems
Clarkson, C. (2025) · Cronfa (Swansea University) · doi:10.23889/suthesis.69942
Wu, Y., Zhou, C., Dewulf, W., & Peeters, J. R. (2025) · Procedia CIRP · doi:10.1016/j.procir.2024.12.143
A review of super-resolution in industrial applications: Methods, task challenges, and future trends
Zhao, W., Lee, C. K. M., & Cheung, B. C. F. (2026) · Intelligent Systems with Applications · doi:10.1016/j.iswa.2026.200690
Tang, J., Cao, L., Xu, G., Li, M., Wu, W., & Yuan, H. (2026) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-026-02909-0
Metadata
- Authors
- Song, L., Lin, W., Yang, Y.-G., Zhu, X., Guo, Q., & Xi, J.
- Affiliations
- Tiangong University (CN)
- Key Laboratory of Advanced Electrical Engineering and Energy Technology, Tianjin Polytechnic University, Tianjin, China
- National-Local Joint Engineering Laboratory of Intelligent Manufacturing Oriented Automobile Die & Mould, Tianjin University of Technology and Education, Tianjin, China
- University of Wollongong (AU)
- Venue
- IEEE Access
- Published
- 2019
- Volume
- 7
- Pages
- 27547-27554
- ISSN
- 2169-3536 (ISSN-L)
- OpenAlex
- W2917663261
- Cited by
- 61
- References
- 22
- Citations by year
- 2026: 3 · 2025: 6 · 2024: 6 · 2023: 9 · 2022: 16 · 2021: 10 · 2020: 10 · 2019: 1
- FWCI
- 6.82
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Locations
- IEEE Access (publishedVersion)
- Research Online (University of Wollongong) (submittedVersion)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
- Figshare (submittedVersion)
- Funders
- National Natural Science Foundation of China (61078041, 51806150)
- Natural Science Foundation of Tianjin City (16JCYBJC15400, 18JCQNJC04400)
- National Natural Science Foundation Committee
- Tianjin Research Program of Application Foundation and Advanced Technology
- State Key Laboratory of Precision Measurement Technology and Instruments (PIL1603)
- Tianjin University (TD13-5036)
- 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 › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Keywords
- scratch detection, deep convolutional neural networks, surface defects, background noise suppression, skeleton extraction
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.