An End-to-End Steel Surface Defect Detection Approach via Fusing Multiple Hierarchical Features
Open accessThe DOI matches its Crossref record; 43 of the 53 references it cites were found in a registry or an index.
Abstract
A complete defect detection task aims to achieve the specific class and precise location of each defect in an image, which makes it still challenging for applying this task in practice. The defect detection is a composite task of classification and location, leading to related methods is often hard to take into account the accuracy of both. The implementation of defect detection depends on a special detection data set that contains expensive manual annotations. In this paper, we proposed a novel defect detection system based on deep learning and focused on a practical industrial application: steel plate defect inspection. In order to achieve strong classification ability, this system employs a baseline convolution neural network (CNN) to generate feature maps at each stage, and then the proposed multilevel feature fusion network (MFN) combines multiple hierarchical features into one feature, which can include more location details of defects. Based on these multilevel features, a region proposal network (RPN) is adopted to generate regions of interest (ROIs). For each ROI, a detector, consisting of a classifier and a bounding box regressor, produces the final detection results. Finally, we set up a defect detection data set NEU-DET for training and evaluating our method. On the NEU-DET, our method achieves 74.8/82.3 mAP with baseline networks ResNet34/50 by using 300 proposals. In addition, by using only 50 proposals, our method can detect at 20 ft/s on a single GPU and reach…
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
“There are some shortcomings in manual testing, such as low confidence and high labor intensity.”
From A New Steel Defect Detection Algorithm Based on Deep Learning (2021) · cites this paper as [5]
In the citing paper
“most classification tasks can only find the defects with the highest confidence level in the defect category, resulting in imprecise classification results [3].”
From Steel Surface Defect Recognition: A Survey (2022) · cites this paper as [3]
In the citing paper
“there are many types of steel surface defects and some of these defects may overlap”
From Steel Surface Defect Recognition: A Survey (2022) · cites this paper as [3]
In the citing paper
“Using manual inspection of steel surface defects, it is difficult to distinguish the type of defect and accurately detect the location of the defect [2].”
From A deep learning model for steel surface defect detection (2024) · cites this paper as [2]
In the citing paper
“There are many categories of steel surface defects and the size of each defect category varies widely. The average size of each category in the NEU-DET [34] dataset is shown in Table1.”
From A deep learning model for steel surface defect detection (2024) · cites this paper as [34]
1 more claim about this paper
Create a free account to see every claim other papers make about this one.
Citations
Cites
Deep Residual Learning for Image Recognition
He, K., Zhang, X., Ren, S., & Sun, J. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.90 · Cited by 235,474 · cited 1×
“The third type is ResNet using residual blocks to make networks deeper without overfitting [23].”
MethodVery Deep Convolutional Networks for Large-Scale Image Recognition
Simonyan, K., & Zisserman, A. (2014) · arXiv · doi:10.48550/arxiv.1409.1556 · Cited by 114,148
ImageNet classification with deep convolutional neural networks
Krizhevsky, A., Sutskever, I., & Hinton, G. E. (2017) · Communications of the ACM · doi:10.1145/3065386 · Cited by 112,155 · cited 1×
“cessful networks are based on the sequential pipeline architecture [25], which establish the basic structure of CNN and prove the importance of depth of networks.”
BackgroundLeCun, Y., Bengio, Y., & Hinton, G. (2015) · Nature · doi:10.1038/nature14539 · Cited by 84,126
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472 · cited 6×
“Next, RPN [14] is employed to generate region proposals”
BackgroundMethodInfluentialFaster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2015) · arXiv · doi:10.48550/arxiv.1506.01497 · Cited by 70,812 · cited 6×
“Next, RPN [14] is employed to generate region proposals”
BackgroundMethodInfluentialImageNet: A large-scale hierarchical image database
Deng, J., Dong, W., Socher, R., Li, L.-J., Li, K., & Fei-Fei, L. (2009) · 2009 IEEE Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2009.5206848 · Cited by 65,333 · cited 1×
“As current research in transfer learning [15], the key to drive large networks is pretraining on ImageNet [22].”
BackgroundGradient-based learning applied to document recognition
Lecun, Y., Bottou, L., Bengio, Y., & Haffner, P. (1998) · Proceedings of the IEEE · doi:10.1109/5.726791 · Cited by 58,805 · cited 1×
“combined? There are two essential conditions: nonadjacent, because adjacent layers have highly local correlation [32], and”
BackgroundYou Only Look Once: Unified, Real-Time Object Detection
Redmon, J., Divvala, S., Girshick, R., & Farhadi, A. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.91 · Cited by 53,803 · cited 1×
“The representative direct regression methods are YOLO [29] and SSD [30].”
MethodFully convolutional networks for semantic segmentation
Long, J., Shelhamer, E., & Darrell, T. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298965 · Cited by 38,834 · cited 2×
“The most famous region-based detectors are the “R-CNN family” [27] [28] [14].”
BackgroundMethod
Cited by
Items 601–700 of 1,338
A Lightweight guided and shared feature fusion model for aircraft skin defect detection
Li, J., Huang, C., Li, H., Han, W., & Fan, J. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.111078 · Cited by 14
Li, B., Yuan, X., & Shi, M. (2020) · Journal of Electronic Imaging · doi:10.1117/1.jei.29.1.013016 · Cited by 9
Steel Defect Classification Using Machine Learning
Arshad, S. R., Obaid, I., Gull, R., & Shahzad, M. (2022) · International Conference on Ubiquitous Information Management and Communication (IMCOM) · doi:10.1109/imcom53663.2022.9721728 · Cited by 11
A Four-Stage Product Appearance Defect Detection Method With Small Samples
Xie, X., Zhang, R., Peng, L., & Peng, S. (2022) · IEEE Access · doi:10.1109/access.2022.3196935 · Cited by 7
Active Learning-Based Complex Pipeline Weld Defect Detection with Lightweight Neural Network
Zuo, F., Liu, J., Wang, L., Qu, F., & Fu, M. (2022) · IEEE Data Driven Control and Learning Systems Conference (DDCLS) · doi:10.1109/ddcls55054.2022.9858378 · Cited by 9
“In [11], feature fusion network of multi-stage feature maps in the process of neural network feature extraction is proposed to improve the location information of the defect.”
BackgroundSAGE: S tealthy A ttack Ge neration in cyber-physical systems
Biehler, M., Zhong, Z., & Shi, J. (2023) · IISE Transactions · doi:10.1080/24725854.2023.2184004 · Cited by 8
“These accuracy results utilizing transfer learning outperform recently published results from He et al. (2019) on a ResNet50 trained from scratch on the dataset (99.67% accuracy).”
InfluentialAutomatic Visual Pit Detection System for Bottom Surface of Cylindrical Lithium Battery
Guo, S., & Xing, Z. (2023) · IEEE Access · doi:10.1109/access.2023.3283576 · Cited by 9
“In [11], this system employs a baseline convolution neural network (CNN) to generate feature maps at each stage, and then the proposed multilevel feature fusion network (MFN) combines multiple hierarchical features into one feature, which can include more location details of defects.”
MethodSteel Surface Defect Detection Based on an Improved YOLOv5 Model
Zhong, H., Wu, B., Zhang, X., Hu, Y., Tianming, Z., & Zhang, Y. (2023) · International Conference on Intelligent Control, Measurement and Signal Processing (ICMSP) · doi:10.1109/icmsp58539.2023.10170867 · Cited by 9
“Dataset and evaluation metrics In this paper, we evaluate the performance of our proposed algorithm using the NEU-DET dataset created by Northeastern University [10].”
MethodYang, J., Peng, J., Li, Y., Xie, Q., Wu, Q., & Wang, J. (2021) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2021.3089761 · Cited by 7
Niu, Y., Zheng, Y.-L., Shi, S.-J., et al. (2025) · Engineering applications of artificial intelligence · doi:10.1016/j.engappai.2025.112754 · Cited by 9
MTGGAN: generate type-controllable magnetic tile defect data to improve defect detection performance
Jiang, C., Xu, B., Li, Z., Zhang, L., Sun, Q. S., & Zhang, D. (2024) · Nondestructive Testing And Evaluation · doi:10.1080/10589759.2024.2419895 · Cited by 6
“The hot-rolled steel data comes from the NEU dataset [43], which includes six types of defects: crazing (Cr), inclusion (In), patches (Pa), pitted surface (PS), rolled-in scale between different types.”
Mobile-YOLO-SDD: A Lightweight YOLO for Real-time Steel Defect Detection
Luo, S., Xu, Y., Zhu, M., Zhang, C., Kong, C., Jin, J., Li, T., Jiang, X., & Guo, B. (2024) · Procedia CIRP · doi:10.1016/j.procir.2024.10.040 · Cited by 8
Song, J., Tian, Y., & Wan, X. (2024) · Sensors · doi:10.3390/s24247914 · Cited by 7
Luo, X., Wang, J., Dong, T. X., et al. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3568956 · Cited by 6
“He et al. [13] used CNNs to generate feature maps and employed a multilevel feature fusion network (MFN) to integrate features from di ff erent levels.”
Jin, S., Cao, Z., & Yu, C. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.111190 · Cited by 9
Li, Z., Lai, X., Lou, S.-T., & Li, L. (2025) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2025.3575135 · Cited by 7
“2) Neu-Det: This dataset [14], comprising 1800 images each with a size of 200 × 200 in total, poses a challenging task for the inspection of strip steel surface defects.”
InfluentialLi, L., Hu, S., Hou, Y., Ke, Y., & Chen, Z. (2025) · Measurement Science and Technology · doi:10.1088/1361-6501/ae08dc · Cited by 7
“The detailed process of MSPM is illustrated in equation (10): To comprehensively evaluate the defect detection performance and versatility within the ADMA-YOLO, this study selected two public datasets, NEU-DET [36] and GC10-DET [37], for verification.”
Steel surface defect detection algorithm based on improved YOLOv10
Zhang, L., Wang, Z., Ma, Y., & Li, G. (2025) · Scientific Reports · doi:10.1038/s41598-025-16725-8 · Cited by 15
Optimizing surface defect detection with YOLOv9: the role of advanced backbone models
Zeng, Z., Wang, H., Yao, C., Dong, Z., & Cai, S. (2025) · Frontiers in Artificial Intelligence · doi:10.3389/frai.2025.1675154 · Cited by 7
QRT-DETR: Post-training quantization for real-time detection transformer
Huo, Y., Wu, T., Shen, Y., Li, X., Tao, Z., & Yang, D. (2025) · Neurocomputing · doi:10.1016/j.neucom.2025.131957 · Cited by 9
High-resolution image-based surface defect detection method for hot-rolled strip steel
Feng, X., Gao, X., & Luo, L. (2022) · Journal of Electronic Imaging · doi:10.1117/1.jei.31.3.033020 · Cited by 6
Rtsds:a real-time and efficient method for detecting surface defects in strip steel
Zeng, Q., Wei, D., & Zou, M. (2024) · Journal of Real-Time Image Processing · doi:10.1007/s11554-024-01497-7 · Cited by 7
“To evaluate the comprehensive performance of RTSD, we compare it with several state-of-the-art detection methods, including Faster-RCNN [2], YOLOv3 [5], YOLOv5s [6], YOLOv7 [7], YOLOv8 [8], DCAM-Net [31], DDN [32], RT-DETR [33], Salience DETR [34] and other model [35, 36].”
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
Aşar, E., & Özgür, A. (2024) · Engineering materials · doi:10.1007/978-3-031-57468-9_3 · Cited by 11
CEC-YOLO: An Improved Steel Defect Detection Algorithm Based on YOLOv5
Chen, Z., Wang, Y., & Gu, Q. (2024) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn60899.2024.10651516 · Cited by 7
“NEU-DET dataset is the open defect detection dataset of He et al. [29], six kinds of typical surface defects of the hot-rolled steel strip are collected, rolled-in scale (RS), patches (Pa), crazing (Cr), pitted surface (PS), inclusion (In) and scratches (Sc).”
MethodTransformer-Based Multiscale Reconstruction Network for Defect Detection of Infrared Images
Wei, C., Han, H., Wu, Z., Xia, Y., & Ji, Z. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3481573 · Cited by 10
“Within the domain of industrial defect detection, supervised learning predominantly leverages convolutional neural networks (CNNs) such as you only look once (YOLO) and region-based convolutional neural networks (R-CNNs) to extract meaningful features from images [9], [10], [11].”
BackgroundDHNet: a surface defect detection model utilizing multi-scale convolutional kernels
Zhang, Y., Wang, S., Wang, J., Zhao, Y. H., & Chen, Z. (2025) · Journal of Real-Time Image Processing · doi:10.1007/s11554-025-01623-z · Cited by 6
“These datasets include the NEU [36] dataset, and the DeepPCB dataset, which was obtained from the study [37]. and SMT dataset.”
Optimal 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
Jiao, L., Liu, H., Liang, Z., Chen, P., Wang, R., & Liu, K. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3502719 · Cited by 8
“However, in diverse domains like aerial images [28], [29], [30], [31], scene text [32], [33], [34], [35], and industrial inspection [36], [37], [38], the adoption of oriented BBs has emerged as a preferred approach.”
Multi-defect detection and classification for aluminum alloys with enhanced YOLOv8
Han, Y., Li, X., Cui, G., Song, J., Zhou, F., & Wang, Y. (2025) · PLoS ONE · doi:10.1371/journal.pone.0316817 · Cited by 11
“He et al. [50] proposed a DDN system for steel plate defect detection using deep learning.”
CSLNet: An enhanced YOLOv8-based approach to defect surface foreign objects in Lyophilized powder
Xu, H., Liu, Q., Zhu, J., Dai, H., & Zhang, D. (2025) · Signal Image and Video Processing · doi:10.1007/s11760-025-04335-w · Cited by 6
“He et al. [19] designed a defect detection network (DDN) for steel plates, which is used for discerning the specific category and precise location of a defect through the multilevel features.”
Deep learning-based methods for detecting surface defects in steel plates
Patil, S., Miglani, A., Kankar, P. K., & Roy, D. (2022) · Elsevier eBooks · doi:10.1016/b978-0-323-90789-7.00001-4 · Cited by 8
Detection of Defective Steel Surface with Image Segmentation
Gupta, C., Chauhan, G., & Aiyar, H. D. S. (2022) · International Conference on Sustainable Computing and Data Communication Systems (ICSCDS) · doi:10.1109/icscds53736.2022.9760806 · Cited by 6
Üzen, H., Türkoğlu, M., Arı, A., & Hanbay, D. (2022) · Gazi Üniversitesi Mühendislik-Mimarlık Fakültesi Dergisi · doi:10.17341/gazimmfd.1024425 · Cited by 8
Stephen, O., Madanian, S., & Nguyen, M. (2022) · Sensors · doi:10.3390/s22249971 · Cited by 6
Wang, L., Huang, X., & Zheng, Z. (2023) · PLoS ONE · doi:10.1371/journal.pone.0280363 · Cited by 9
“[21] adopts feature fusion strategy to improve the richness of the extracted features by fusing multilevel feature information of the defect target, but fails to solve problems such as small samples and small targets in the field of surface defect detection; Wang et al.”
MethodMulti-label Classification of Steel Surface Defects Using Transfer Learning and Vision Transformer
Komijani, A., Vafaeinezhad, F., Khoramdel, J., Borhani, Y., & Najafi, E. (2022) · International Conference on Information and Knowledge Technology (IKT) · doi:10.1109/ikt57960.2022.10039038 · Cited by 9
“Yu He [13] used object detection, a composite task of localization and classification.”
MethodA Tiny Defect Detection System for Tire Mold Surfaces Based on Consecutive Frames
Lin, Y.-H., & Ruan, S.-J. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3252632 · Cited by 6
“[16] used convolutional neural networks (CNNs) to detect surface defects on the steel product.”
MethodAn Improved YOLOv5-based Method for Surface Defect Detection of Steel Plate
Zhu, L., Zhang, J., & Jia, C. (2022) · 2022 China Automation Congress (CAC) · doi:10.1109/cac57257.2022.10055055 · Cited by 6
“Additionally, a strip surface defect detection network based on Faster R-CNN [9] is proposed [10], which achieves 82.”
MethodZhang, F., & Chen, Z. (2023) · Journal of Industrial Information Integration · doi:10.1016/j.jii.2023.100463 · Cited by 12
GAN-based Data Augmentation for Metal Surface Defect Detection Using Convolutional Neural Networks
Tseng, L.-S., Wu, C., Chen, Y. H., & Tsai, C.-H. (2023) · International Symposium on Computer, Consumer and Control · doi:10.1109/is3c57901.2023.00029 · Cited by 6
Detection of Internal Wire Broken in Mining Wire Ropes Based on WOA–VMD and PSO–LSSVM Algorithms
Li, P., Tian, J., Zhou, Z., & Wang, W. (2023) · Axioms · doi:10.3390/axioms12100995 · Cited by 7
Li, W., Li, Z., Wang, C., et al. (2023) · Journal of Intelligent & Fuzzy Systems · doi:10.3233/jifs-231423 · Cited by 6
Surface Defect Classification of Steels Based on Ensemble of Extreme Learning Machines
Liu, Y., Jin, Y., & Ma, H. (2019) · WRC Symposium on Advanced Robotics and Automation (WRC SARA) · doi:10.1109/wrc-sara.2019.8931807 · Cited by 6
“, pitted surface, rolled-in scale and scratches [1, 2].”
BackgroundTsai, C., & Chen, H. (2020) · Sensors · doi:10.3390/s20164356 · Cited by 7
“[21] proposed a CNN-based surface defect detection approach, which uses a Multilevel-feature Fusion Network (MFN) to combine multiple hierarchical features into a multilevel feature.”
MethodPeng, C.-C. (2021) · IEEE Open Journal of the Industrial Electronics Society · doi:10.1109/ojies.2021.3109906 · Cited by 6
Watermark Detection in CMOS Image Sensors Using Cosine-Convolutional Semantic Networks
Solórzano, C. A. P., & Tsai, D. (2023) · IEEE Transactions on Semiconductor Manufacturing · doi:10.1109/tsm.2023.3245606 · Cited by 6
“CNN models have been used in the manufacturing industry for steel surface defect detection [21], [22], mura defect detection in LCDs [23] and detection of cracks on nuclear reactors by using individual video frames [24].”
MethodKlamert, V., Achsel, T., Toker, E., Bublin, M., & Otto, A. (2023) · Applied Sciences · doi:10.3390/app132011273 · Cited by 5
“A steel plate defect inspection was proposed by He et al. [52] for real-time quality control of the manufacturing process.”
BackgroundA Multi-Scale Defect Detection for Steel Surface Based on Improved Faster R-CNN
Shen, M., Cai, F., & Xie, J. (2023) · China Automation Congress (CAC) · doi:10.1109/cac59555.2023.10451783 · Cited by 5
Research on Low Contrast Surface Defect Detection Method Based on Improved YOLOv7
Chen, S., Li, W. X., Yan, X., Liu, W., Chen, C., Liao, J., Chen, X., & Shu, J. (2024) · IEEE Access · doi:10.1109/access.2024.3429283 · Cited by 5
Zhou, Y. X., Wu, H., Wang, Y., Liu, X., Zhai, X., Sun, K., Zheng, Z., Tian, C., Zhao, H., Jia, W., et al. (2025) · Measurement · doi:10.1016/j.measurement.2024.116639 · Cited by 6
A Study of Corrosion-Grade Recognition on Metal Surfaces Based on Improved YOLOv8 Model
Chen, H., Cao, Y. S., Cao, S., & Piao, H. (2025) · Sensors · doi:10.3390/s25082630 · Cited by 5
“Therefore, the direct economic losses caused by corrosion are huge, and in order to prevent the problem of the corrosion of station equipment from becoming more serious, timely detection, assessment, and maintenance by station staff are required [3,4].”
Li, M., Liu, Y., Chen, D., & Li, X. (2025) · The Journal of Supercomputing · doi:10.1007/s11227-025-07397-2 · Cited by 7
“Therefore, deep networks have become an important means to address metal image analysis and detection issues [29–32].”
Multimodal mechanical wear fault diagnosis: Fusion of signal characterization and image information
Zhou, Q., Chai, B., Guo, Y., Li, T., Zhou, S., Wang, K., & Ye, Y. (2025) · Results in Engineering · doi:10.1016/j.rineng.2025.107204 · Cited by 5
Fang, W., Yang, Y. Q., Zhang, W., Wang, T., Feng, J., & Liu, G. (2025) · Measurement Science and Technology · doi:10.1088/1361-6501/ae06bf · Cited by 8
“This study utilizes two steel surface defect datasets: NEU-DET from Northeastern University [36] and the GC10-DET dataset [37], both specifically designed for surface defect detection in steel materials.”
RMC-YOLO: An Object Detection Model for Insulator Defects
Lan, X., Wang, X., & Liu, L. (2025) · Signal Image and Video Processing · doi:10.1007/s11760-025-04568-9 · Cited by 5
Wang, H., Zhang, F., & Yi, R. (2025) · Sensors · doi:10.3390/s25226890 · Cited by 6
“Traditionally, steel surface defect identification [4] was performed using a combination of manual inspection and traditional machine vision methods.”
A lightweight detection network integrating multi-scale semantic refinement for steel strip defects
Bian, B., Chen, L., Han, Z., Wu, S., Li, W., & Chen, H. (2025) · Robotics and Computer-Integrated Manufacturing · doi:10.1016/j.rcim.2025.103188 · Cited by 9
Zhang, Y., Sun, Y.-N., Huang, Q., & Han, Y. (2025) · Measurement Science and Technology · doi:10.1088/1361-6501/ae26a5 · Cited by 10
“(1) NEU-DET [55]: this public dataset was developed by Northeastern University.”
Wu, M., Peng, J., Yu, X., Xu, H., & Sun, H. (2026) · PLoS ONE · doi:10.1371/journal.pone.0339617 · Cited by 5
“To evaluate the effectiveness of our proposed framework, we conduct comprehensive experiments using the NEU-DET [37] dataset for initial module verification, followed by generalization validation on the GC10-DET [38] dataset.”
A Survey of Deep learning in Advancing Steel Industry Standards
Aghapanah, H., Rad, A. S., & Rasti, R. (2024) · CSI International Symposium on Artificial Intelligence and Signal Processing · doi:10.1109/aisp61396.2024.10475283 · Cited by 9
“The Northeastern University (NEU) dataset is also related to steel surface defect detection with 1800 images [41][42].”
Sun, M., Dong, F., Huang, Z., & Luo, J. (2024) · Tsinghua Science & Technology · doi:10.26599/tst.2024.9010039 · Cited by 6
TAANet: A Task-Aware Attention Network for Weak Surface Defect Detection
Cui, L., Xie, S., Chen, E., Jiang, X., Wang, Z., Guo, X., & Xu, M. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3457936 · Cited by 8
“We evaluate the proposed methods on two publicly available steel surface defect detection datasets: NEU-DET [45] and GC10-DET [46].”
MethodSteel Surface Defect Detection Based on Denoising Diffusion Implicit Models with Data Augmentation
Hong, Y., Wang, Z., Wu, W., et al. (2024) · International Conference on Imaging, Signal Processing and Communications (ICISPC) · doi:10.1109/icispc63824.2024.00010 · Cited by 5
“1: Some random samples from the NET-DET [1] dataset: from left to right, crazing, pitted surface, patches, rolled-in scales, inclusion, scratches. to degradation of product performance, shortening of product life, and can easily lead to accidents.”
BackgroundMethodInfluentialChen, X., Chang, W., Li, Y., et al. (2024) · Electronic Research Archive · doi:10.3934/era.2024292 · Cited by 5
“It is widely used for a variety of vision tasks, and competitive results can be obtained with only a few parameters [28].”
EC-PFN: a multiscale woven fusion network for industrial product surface defect detection
Liu, S., & Li, J. (2024) · Complex & Intelligent Systems · doi:10.1007/s40747-024-01699-3 · Cited by 9
“In the Northeastern University (NEU) surface defect database [41, 42] contains six typical surface defects found on hot-rolled steel strips, including rolled-in scale (RS), patches (Pa), crazing (Cr), pitted surface (PS), inclusion (In), and scratches (Sc).”
FP-YOLOv8: Surface Defect Detection Algorithm for Brake Pipe Ends Based on Improved YOLOv8n
Rao, K., Zhao, F., & Shi, T. (2024) · Sensors · doi:10.3390/s24248220 · Cited by 6
Defect attention-based lightweight real-time adaptable surface defect analysis
Delibaşoğlu, İ., DAŞ, D. B., & Daş, O. (2025) · Ironmaking & Steelmaking Processes Products and Applications · doi:10.1177/03019233241296959 · Cited by 7
“They validated their approach by considering two datasets called NEU-DET [36] and their dataset GC10-DET.”
InfluentialJiang, S., Liu, M., Liu, Y., Ma, Y., & Wang, Y. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3547122 · Cited by 6
“…the above two scenarios, it can be seen that when the clustering analysis in CSS performs a binary classification task, the mean MBA is 4.54% , 1.34% , and 4.50% higher than when performing a nonbinary clustering task on the NEU-DET [3], GC10-DET [53], and Magnetic-Tile [54] dataset, respectively.”
InfluentialMSCM-Net: Rail Surface Defect Detection Based on a Multi-Scale Cross-Modal Network
Wen, X., Zheng, X., & He, Y. (2025) · Computers, materials & continua/Computers, materials & continua (Print) · doi:10.32604/cmc.2025.060661 · Cited by 6
“He et al. [16] focused on the challenge of feature extraction, designing a multi-level feature fusion mechanism based on ResNet combined with an RPN (Region Proposal Network) network to construct a complete end-to-end detection system, thereby enhancing the algorithm’s ability to recognize…”
Self-citationA novel steel surface defect detector with wavelet upsampling and frequency-domain attention
Gan, Y., Ren, X., Liu, H., Lin, P., & Chen, Y. (2025) · Nondestructive Testing And Evaluation · doi:10.1080/10589759.2025.2566787 · Cited by 8
“To validate the effectiveness of the improved model, the publicly available NEU-DET dataset [38] developed by Northeastern University is used.”
Flaw Detection in White Porcelain Wine Bottles Based on Improved YOLOv4 Algorithm
Gong, G., Huang, J. J., & Wang, H. (2022) · Frontiers in Bioengineering and Biotechnology · doi:10.3389/fbioe.2022.928900 · Cited by 5
Jiang, Z., Yu, J., Liu, J., Chen, Z., Gui, W., & Cao, T. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3216662 · Cited by 5
“Aligned with the sustainable development of industrial information and image technology, detection methods based on machine vision have the advantages of noncontact, high accuracy, and low cost [7].”
BackgroundQin, Y., Wang, S., Xi, D., & Chen, L. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3250305 · Cited by 5
“Convolutional neural networks (CNNs) have shown excellent feature extraction ability and good segmentation performance in recent years, which are gradually applied to different fields [7], [9], such as automatic drive, face recognition, and medical image segmentation [10].”
BackgroundBiSPD-YOLO: Surface Defect Detection Method for Small Features and Low-resolution Images
Yan, S., Chen, G., Gao, A., Liu, C., & Xiong, Z. (2023) · IEEE/ASME International Conference on Advanced Intelligent Mechatronics · doi:10.1109/aim46323.2023.10196194 · Cited by 6
“This model is based on Faster R-CNN, combines multiple hierarchical features into one feature through multi-layer feature fusion network (MFN), and achieves higher detection accuracy on NEU-DET dataset.”
BackgroundMethodInfluentialSurface defect detection of steel based on improved YOLOv5 algorithm
Jiang, Y. (2023) · Mathematical Biosciences & Engineering · doi:10.3934/mbe.2023879 · Cited by 7
“He et al. [13] generated feature maps through CNNs for steel plate defect detection applications and used multilevel feature fusion networks to combine multiple levels of features into one feature, achieving improved defect detection performance.”
MethodDevelopment of an Algorithm for Detecting Real-Time Defects in Steel
Yu, J., Wang, C., Xi, T., Ju, H., Qu, Y., Kong, Y., & Chen, X. (2023) · Electronics · doi:10.3390/electronics12214422 · Cited by 5
SOIF-DN: Preserving Small Object Information Flow With Improved Deep Learning Model
Joo, I., Kim, S.-H., Kim, G., & Yoo, K. (2023) · IEEE Access · doi:10.1109/access.2023.3333942 · Cited by 5
“To address the issues, He et al. [17] proposed a feature pyramid network (FPN).”
BackgroundHe, Z., Zhao, J., & Zhao, X. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3420348 · Cited by 9
“More effective approaches enable simultaneous training of a detector and a bounding box regressor under similar network configurations [17], [18], [19].”
MethodSu, X., Feng, X., Gong, S., & Ansari, F. (2024) · Tunnelling and Underground Space Technology · doi:10.1016/j.tust.2024.106085 · Cited by 6
CFFDist: Cross-Scale Feature Fusion Distillation Network for Industrial Anomaly Localization
Zhi, H., Qin, H., Zhang, L., Guo, J., & Song, B. (2024) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2024.3502773 · Cited by 6
A small defect detection technique for industrial product surfaces based on the EA-YOLO model
Li, B., Wang, B., Hu, X., Zhai, J., & Ji, C. (2025) · The Journal of Supercomputing · doi:10.1007/s11227-025-06929-0 · Cited by 7
YOLOv8-SC: improving the YOLOv8 network for real-time detection of automotive coated surface defects
Ling, L., Zhu, C., Liu, M., Hu, J., Zhang, X., & Ge, M. (2025) · Measurement Science and Technology · doi:10.1088/1361-6501/adb05b · Cited by 8
“In order to enhance the generalizability of the proposed method, the public industrial data set NEU-DET [43] is selected for supplementary verification.”
Background-Weaken Generalization Network for Few-Shot Industrial Metal Defect Segmentation
Yu, R., Li, H., Guo, B., & Zhao, Z. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3544335 · Cited by 12
“Multilevel feature fusion meth-ods [27], [28], [29] are proposed to deal with different level feature, but often inefficient and lack robustness.”
Fault Diagnosis of Rolling Bearing with Imbalanced Small Sample Scenarios
Guan, Y., Meng, Z., & Sun, D. (2021) · Global Reliability and Prognostics and Health Management (PHM) · doi:10.1109/phm-nanjing52125.2021.9612860 · Cited by 5
Steel Frame Structure Defect Detection Using Image Processing and Artificial Intelligence
Baskaran, R., & Fernando, P. (2021) · International Conference on Smart Generation Computing, Communication and Networking (SMART GENCON) · doi:10.1109/smartgencon51891.2021.9645845 · Cited by 7
Elanangai, V., & Vasanth, K. (2022) · Journal of Intelligent & Fuzzy Systems · doi:10.3233/jifs-213031 · Cited by 5
“Yu He et al. [20] have highlighted a steel plate defect inspection technique using deep learning.”
Simulation of Fatigue Fracture Detection of Bridge Steel Structures under Cyclic Loads
Yang, D., Yao, L., & Pang, Q. (2022) · Computational Intelligence and Neuroscience · doi:10.1155/2022/8534824 · Cited by 5
Robust Learning Against Label Noise Based on Activation Trend Tracking
Wang, Y., Zhang, Y., Jiang, Z., Li, Z., Chen, J., & Lu, J. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3219493 · Cited by 6
“Methods that use deep learning have achieved impressive performance in many tasks such as defect classifi-cation [9], defect location [10], and segmentation [11].”
MethodAn Improved Faster RCNN based on Swin Transformer for Surface Defect Detection of Metal Workpieces
Xu, S., & Shao, Z. (2022) · International Conference on Electronic Information Technology and Computer Engineering (EITCE) · doi:10.1145/3573428.3573774 · Cited by 5
“The data set used in this study is NEU-DET [11], with 1800 images in total, containing six categories of defects commonly found on steel surfaces, namely: crazing, inclusion, patches, pitted surface, rolled-in scale and scratches.”
BackgroundMethodInfluentialIngot oxide slag detection using two-stage UNet network based on mixed supervised learning
Wu, J., Xu, D., Yang, C., & Gui, W. (2023) · Neural Computing and Applications · doi:10.1007/s00521-023-08600-2 · Cited by 5
“10, the watershed algorithm and FCN are unable to identify oxide slag regions, and the training of the [28] and [29] networks lacks a large number of pixel-level labels, resulting in overfitting of the model and poor robustness of detection.”
BackgroundMethodInfluentialAn Anchor-Free Pipeline MFL Image Detection Method
Han, F., Lang, X., & Liu, M. (2023) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2023.3304688 · Cited by 8
“[12] used a baseline CNN for feature extraction while introducing a multilayer feature fusion method (MFN) to enhance the accuracy of”
MethodXie, W., Liu, W., Xiao, Y., Dai, Y., Dong, J., & Liu, R. (2023) · IEEE Sensors Journal · doi:10.1109/jsen.2023.3318998 · Cited by 6
“For example, He et al. [1] propose an end-to-end steel surface defect detection method by fusing multiple hierarchical features.”
MethodApplication of Task-Aligned Model Based on Defect Detection
Hung, M.-H., Ku, C.-H., & Chen, K.-Y. (2023) · Automation · doi:10.3390/automation4040019 · Cited by 5
“[7] proposed the defect detection network (DDN) architecture, capable of generating more feature information, but it performs poorly on multiple sca tt ered defects. feature information, but it performs poorly on multiple scattered defects.”
BackgroundCasting defect-detection method based on multi-image feature of fusion artificial intelligence model
Yang, D., Jiang, H., He, Y., Wang, Y., Liu, Z., Zhang, X., & Cheng, H. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06429-x · Cited by 5
“To further assess the detection capability and generalization of the proposed model, we conducted experiments using the publicly available hot-rolled strip steel surface defect data-set, NEU-DET [51].”
Self-citationSteel Surface Defect Detection Algorithm Based on Improved YOLOv8 Modeling
Peng, M., Shunke, B. A. I., & Lu, Y. (2025) · Applied Sciences · doi:10.3390/app15158759 · Cited by 7
CTC-YOLO: an improved YOLOv11 algorithm for steel surface defect detection
Yang, L., Li, Z., Hu, X., et al. (2025) · Engineering Research Express · doi:10.1088/2631-8695/adf9bf · Cited by 4
Wang, X., Luo, K., Liu, M., Zheng, R., & Li, Y. (2025) · Advanced Engineering Informatics · doi:10.1016/j.aei.2025.104011 · Cited by 6
An Efficient Lightweight Method for Steel Surface Defect Detection
Zheng, A., Jiang, X., & Liu, W. (2025) · Sensors · doi:10.3390/s25247527 · Cited by 4
“To verify the effectiveness of the proposed LCED-YOLO network, we conduct experimental verification on the NEU-DET [45] and GC10-DET [46] datasets.”
A Cross-Scale Spatial–Semantic Feature Aggregation Network for Strip Steel Surface Defect Detection
Xu, C., Sun, Y., Huang, L., & Guo, H. (2025) · Materials · doi:10.3390/ma18245567 · Cited by 5
“To validate the effectiveness of the proposed method, we conduct extensive experiments on two challenging steel surface defect datasets: NEU-DET [45–47] and GC10-DET [48].”
Metadata
- Authors
- He, Y., Song, K., Meng, Q., & Yan, Y.
- Affiliations
- Northeastern University (CN)
- Loughborough University (GB)
- Venue
- IEEE Transactions on Instrumentation and Measurement
- Published
- 2020-04
- Volume
- 69
- Issue
- 4
- Pages
- 1493-1504
- ISSN
- 0018-9456 (ISSN-L)
- 1557-9662
- OpenAlex
- W2944303778
- Semantic Scholar
- 46d4a2b7ccf38b80b09574cf17544ff2297a8fbf
- Cited by
- 1,341
- References
- 53
- Citations by year
- 2026: 188 · 2025: 304 · 2024: 269 · 2023: 238 · 2022: 186 · 2021: 111 · 2020: 34
- FWCI
- 65.18
- Type
- journal article
- Language
- en
- Open access
- green
- figshare.com
- Licence
- Locations
- IEEE Transactions on Instrumentation and Measurement (publishedVersion)
- Figshare (submittedVersion, cc-by)
- Funders
- National Natural Science Foundation of China (51805078, 51374063)
- National Basic Research Program of China (2017YFB0304200)
- Fundamental Research Funds for the Central Universities (N170304014, N150308001)
- China Scholarship Council (201806085007)
- 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
- multi-level feature fusion, hierarchical features, region proposal network, bounding box regression, NEU-DET dataset, real-time defect detection, industrial visual inspection, object detection
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.