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 901–1,000 of 1,338
Industrial machine tool component surface defect dataset
Schlagenhauf, T., & Landwehr, M. (2021) · Data in Brief · doi:10.1016/j.dib.2021.107643 · Cited by 52
An Improved Two-Stage Network for Steel Surface Defect Detection
Zhang, Y., Zou, L., Wei, Y., & Dai, W. (2023) · International Conference on Computing and Artificial Intelligence (ICCAI) · doi:10.1145/3594315.3594653 · Cited by 1
“An end-to-end strip surface defect detection network with multi-level features was proposed by He et al [14].”
MethodShi, W., Zhao, T. G., & Zhang, J. W. (2023) · China Semiconductor Technology International Conference (CSTIC) · doi:10.1109/cstic58779.2023.10219404 · Cited by 1
“In the first stage, the region proposal, we use the ResNet34/50 backbone to extract feature maps from the input images[3].”
MethodSurface Defect Detection of Strip Steel Based on YOLOv5
Wang, D., & Liu, X. (2023) · Academic Journal of Science and Technology · doi:10.54097/ajst.v7i2.11778 · Cited by 1
Simpler large margin distribution machine via weighted linear loss for large-scale classification
Chu, M., Liu, L., Liu, L., & Gong, R. (2023) · International Journal of Machine Learning and Cybernetics · doi:10.1007/s13042-023-02028-9 · Cited by 1
“To evaluate our proposed SWLLDM, including classification accuracy and running efficiency, some experiments are conducted on the UCI benchmark datasets [19], David Musicant’s NDC datasets [20] and an application dataset [21, 22].”
Few-shot Industrial Defect Image Classification Based on Lightweight Model with Attention Mechanism
Tu, M., Qi, Z., Yu, L., & Zhang, L. (2023) · Asia Service Sciences and Software Engineering Conference (ASSE) · doi:10.1145/3634814.3634841 · Cited by 1
“3.1.2 NEU surface defect database [19].”
InfluentialStudy on Analysis of Defect Identification Methods in Manufacturing Industry
Pal, V. K., & Mudholkar, P. (2024) · Communications in computer and information science · doi:10.1007/978-3-031-59097-9_35 · Cited by 1
Visual defect inspection of touch screens using multi-angle filtering in curvelet domain
Lin, H.-D., Li, J.-M., & Lin, C.-H. (2024) · Heliyon · doi:10.1016/j.heliyon.2024.e33607 · Cited by 1
Wang, S., Zhang, G., & Guan, B. (2024) · Conference on Fully Actuated System Theory and Applications (FASTA) · doi:10.1109/fasta61401.2024.10595174 · Cited by 1
“He et al. detect steel surface defects by adding multi-level feature fusion to improve network performance [9].”
BackgroundWang, X.-Y., Du, F., Zhao, D., & Liu, C. (2024) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-024-02475-3 · Cited by 3
“…digital model (Tabernik et al., 2020), which needs to label the precise location of cracks pixel by pixel. b Surface defect detection of steel (He et al., 2020), which can be correctly distinguished and labeled by only few professionals due to the high similarity of surface defects such as…”
Ma, Y., Hou, W. S., Xie, Y., Huang, Y., & Han, R. (2024) · IEEE Sensors Journal · doi:10.1109/jsen.2024.3441237 · Cited by 1
“However, the main limitation of these algorithms is that the handcrafted features are generally selected according to the specific dataset and lack of strong feature extraction ability [5].”
BackgroundMethodInfluentialResearch on Steel Surface Defect Identification based on YOLO
Li, S., Yang, B., Kang, J., Ma, B., Pan, Z., & Chang, T. (2024) · International Conference on Computer Information Science and Application Technology (CISAT) · doi:10.1109/cisat62382.2024.10695177 · Cited by 2
“In addition, Yu He [14] has also achieved remarkable results in improving the Faster R-CNN network, by using ResNet to replace the original VGG-16 structure, and combining deformable convolution and dense structure for feature extraction, which enhances the semantic information processing ability…”
MethodA new Dubois et Prade Transform based surface defect categorization
Elleuch, J. F., Mehdi, M. Z., & Sellami, D. (2024) · International Conference on Control, Decision and Information Technologies (CoDIT) · doi:10.1109/codit62066.2024.10708397 · Cited by 1
“We have conducted comparison with both deep learning (Faster-RCNN [2], YOLO-V3 [4] and the MSC-DNet [3]) and traditional machine learning techniques (HoG [16] and LBP [14]).”
MethodDetecting Defect Copper Parts Based on Machine Vision
Zhenhai, T. A. N., Yang, Y. S., Wang, X., et al. (2024) · IEICE Transactions on Fundamentals of Electronics Communications and Computer Sciences · doi:10.1587/transfun.2024eap1109 · Cited by 1
Research on rail surface defect detection algorithm based on improved yolov8
Bu, T., & He, T. (2024) · doi:10.1109/icbase63199.2024.10762389 · Cited by 1
“Yu et al[3]. designed a defect detection system, which is based on the use of ResNet backbone network and the different levels of feature maps. and fusing different levels of feature maps, and then using RPN network for prediction of the fused complete feature maps to get the surface defect…”
Real-time surface defect detection algorithm on aluminum ingot alloy casting lines
Chen, H., Yan, F., Yang, J., yan, J., Qin, T., & Zhang, J. (2024) · Physica Scripta · doi:10.1088/1402-4896/ad9fab · Cited by 4
A surface defect detection instrument for large aperture spherical optical elements
Li, M., Shi, Y., Zhang, Z., Tao, X., & Shang, X. (2025) · Neural Computing and Applications · doi:10.1007/s00521-024-10889-6 · Cited by 2
“Methods are also improved specifically for some industrial tasks in the scope of surface defect detection [24–27 , 42– Based on the main challenges in the above research, in this paper, a surface defect detection instrument is established, which realizes fast and accurate detection for large…”
Liu, Q., Zhong, Y., Yang, W., Liu, Z., Ruan, J., & Zeng, Z. (2024) · IEEE International Conference on Cognitive Computing and Complex Data (ICCD) · doi:10.1109/iccd62811.2024.10843560 · Cited by 3
“The NEU-DET dataset, developed by Northeastern University (NEU) [18], serves as a resource for identifying defects on steel surfaces.”
Designing a New Deep Convolutional Neural Network for Accurate Steel Surface Defects Classification
Ibrahim, A. A. M. S., & Tapamo, J. (2025) · Smart innovation, systems and technologies · doi:10.1007/978-981-97-8096-9_13 · Cited by 1
Xinquan, L., Li, Z., & Lou, S.-T. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3551034 · Cited by 1
“Furthermore, we compare our method with several SOTA defect detectors, including DDN [43], ETDNet [45], ES-Net [21], DCC-CenterNet [25], and CA-AutoAssign [46], which are optimized for specific industrial defect inspection tasks.”
Large Visual Perception Network With Auxiliary Supervision for Surface Defect Detection
Huang, Q., Mu, Y., Han, Y., & Zhang, Y. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3552394 · Cited by 3
“4) Experimental results demonstrate that the proposed LVPAS-Net under the NEU-DET [24], GC10-DET [25], APDDD, and PV multidefect datasets have better performance than the state-of-the-art methods.”
SN-TOOD: An Improved TOOD-Based Steel Surface Defect Detection
Liu, G., Chu, M., & Gong, R. (2024) · International Annual Conference on Complex Systems and Intelligent Science (CSIS-IAC) · doi:10.1109/csis-iac63491.2024.10919284 · Cited by 1
“It exhibits robust classification capabilities and the promise of real-time detection capabilities [3].”
Chen, Z., Zhang, R., Chen, Z., Yuan, B., Zheng, Y., & Li, K. (2025) · Soft Computing · doi:10.1007/s00500-025-10562-7 · Cited by 1
“In recent years, some state-of-the-art object-detection technologies have been developed in industrial defect detection (He et al. 2020; Nakazawa and Kulkarni 2019), mainly dividing into traditional classification detection and deep convolution object detection.”
A Low-Complexity and High-Accuracy Defect Detection Network
Zhou, X., Chen, X., Chen, J., et al. (2025) · Journal of Systems Science and Complexity · doi:10.1007/s11424-025-4425-8 · Cited by 1
“To illustrate the generalization performance of CARNet, we benchmark the constructed CARNet on three challenging datasets: A self-collected dataset, SSGD [24] and NEU-Det [25] .”
Wu, H., Li, B., Tian, L., Dong, C., & Liao, W. (2025) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2025.111009 · Cited by 2
Deep Learning for Weak Feature Defect Detection in MLCC Inspection System
Wang, J., Cao, X., Wang, Q., Zhang, D., & Chen, M. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3571167 · Cited by 3
“Deep learning-based defect detection has been widely used in various fields, such as steel [20], Micro-LED [21], and fabrics [22].”
Feature Decoupling-Guided Annotation Framework for Surface Defects on Steel Strips
Yuan, W., & Liu, W. (2025) · Electronics · doi:10.3390/electronics14112304 · Cited by 1
Ye, W., Ren, J., Liang, J., Liu, W., Yi, H., Zhang, A. A., & Lu, C. (2025) · Structural Health Monitoring · doi:10.1177/14759217251338866 · Cited by 2
Improvement of YOLOv7 Algorithm for Steel Surface Defect Detection Research
Liang, L. H., & Yang, D. (2024) · Asian Conference on Artificial Intelligence Technology (ACAIT) · doi:10.1109/acait63902.2024.11022319 · Cited by 1
“We selected the NEU-DET dataset [1-3] of steel surface defects to conduct our research.”
Narayanasamy, S., & Kulandaivel, R. (2025) · International Journal on Interactive Design and Manufacturing (IJIDeM) · doi:10.1007/s12008-025-02290-x · Cited by 3
“Since GAN uses the similarity index to determine how dissimilar a picture is to the original, the resulting realistic image is easy on the eyes [14–16].”
Tang, H., Zhao, Y., Yang, P., Li, G., & Yan, H. (2025) · International Conference on Neural Networks, Information and Communication Engineering (NNICE) · doi:10.1109/nnice64954.2025.11064388 · Cited by 3
MDC-Net: multimodal defect captioning network for surface steel defects
Chazhoor, A. A. P., Hu, S., Gao, B., & Woo, W. L. (2025) · The Visual Computer · doi:10.1007/s00371-025-04093-3 · Cited by 1
“The graph compares the mAP scores for small, medium, and large objects within the validation dataset, emphasizing the model’s capability in detecting objects of varying sizes across the GC-10 dataset [17, 41] benchmarks, as shown in Table 4.”
InfluentialEdge-FastestDet: a real-time and ultra-lightweight model for detection
Wu, M., Peng, J., Yu, X., et al. (2025) · Journal of Real-Time Image Processing · doi:10.1007/s11554-025-01731-w · Cited by 1
“• In this paper, the accuracy of NEU-DET [19] and GC10 [20] datasets is 49.0% and 33.4% with the number of parameters of 0.29 M.”
InfluentialA Lightweight Steel Surface Defect Detection Algorithm Based on Improved YOLO
Song, J., & Zhang, P. (2025) · Chinese Control and Decision Conference (CCDC) · doi:10.1109/ccdc65474.2025.11091209 · Cited by 1
“The NEU-DET[13] steel defect dataset from Northeastern University is utilized in this experiment.”
Zhang, Z., Peng, L., Dai, G., Wang, M., Gao, Y., Lin, J., Chen, Y., & Cai, W. (2025) · Measurement · doi:10.1016/j.measurement.2025.118610 · Cited by 1
Jiang, H., Li, K., Guo, Z., et al. (2025) · IEEJ Transactions on Electrical and Electronic Engineering · doi:10.1002/tee.70129 · Cited by 1
“For instance, some studies utilize image segmentation techniques to extract location and morphological characteristics from images [14,15].”
Industrial Defect Detection Model Based on Multi-Scale Feature Extraction and Fusion
Wang, S. (2025) · Asia-Pacific Conference on Communications Technology and Computer Science (ACCTCS) · doi:10.1109/acctcs66275.2025.00144 · Cited by 1
“Conventional manual inspection approaches are not only labor-intensive but also susceptible to errors caused by human factors, which compromises the reliability of detection outcomes and leads to increased costs and resource wastage [1].”
Dual-Branch Discriminative Feature Cross-Attention Network for Lumber Surface Defect Segmentation
Zhang, S., Ling, Z., Peng, P., et al. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3617405 · Cited by 1
“The experimental results on two open-source sawn lumber surface defect datasets, as well as a public dataset, including Crack500 [14] and NEU-Seg [15], show that our proposed method significantly improves the performance of defect semantic segmentation.”
CMH-Net: a structured and optimized network for real-time steel surface defect detection
Li, J., Liang-fu, L., & Ai, L. (2025) · Journal of Real-Time Image Processing · doi:10.1007/s11554-025-01786-9 · Cited by 1
“The experiments evaluate the proposed method on two publicly available steel surface defect datasets: NEU-DET [29] and GC10-DET [30].”
Saliency-Guided Diffusion Model for Copper Electrolysis Short-Circuit Detection
Liu, H., Zhang, Y., Zhou, C., Zhang, L., & Hu, Y. (2025) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2025.3614848 · Cited by 1
“In this section, we perform additional experiments on the zinc electrolysis dataset [43] and NEU-DET dataset [69] and compare our SDM with other SOTA detectors to evaluate its generalization capability.”
DDSPNet: A Two-stage defect detection and severity prediction model for industrial products
Tian, K., Yang, S., Chen, L., Wang, K., Ke, Y., Miao, Z., Hu, J., Sun, C., & Zhang, X. (2025) · Applied Intelligence · doi:10.1007/s10489-025-06924-1 · Cited by 4
Yan, Z., Yin, C., Huang, X., et al. (2025) · Nondestructive Testing And Evaluation · doi:10.1080/10589759.2025.2566771 · Cited by 1
“These debris travel at extreme velocities of 7-8 km = s [5], posing significant threats to spacecraft [6–8].”
Research on X‐DR Image Detection Method for Defects in Basin Insulators Based on DSPN‐Transformer
Luo, B., Wang, T., & Xiao, W. (2025) · IET Electric Power Applications · doi:10.1049/elp2.70113 · Cited by 1
CIFormer: Cross Interaction TransFormer for surface defect detection in ceramic tile
Luo, F. L., Zhang, W., Wang, X., Sun, Z., & Zhang, Q. (2025) · Array · doi:10.1016/j.array.2025.100553 · Cited by 2
SD-INPFormer: an unsupervised anomaly detection model for steel surface defect
Liu, G., Chu, M., Gong, R., & Zong, P. (2025) · Nondestructive Testing And Evaluation · doi:10.1080/10589759.2025.2577840 · Cited by 1
Intra-inter modality learning for copper electrolysis short circuit detection
Liu, H., Chen, Y., Zhou, C., Zhang, L., & Hu, Y. (2025) · Advanced Engineering Informatics · doi:10.1016/j.aei.2025.104020 · Cited by 1
ADFE-DET: An Adaptive Dynamic Feature Enhancement Algorithm for Weld Defect Detection
Wu, X., Liu, C., Zhang, H., & Xu, P. (2025) · Applied Sciences · doi:10.3390/app152111595 · Cited by 3
High-precision steel pipe size inspection using deep learning and geometric fitting
Li, L., Gao, Y., Li, K., et al. (2025) · Measurement · doi:10.1016/j.measurement.2025.119556 · Cited by 1
Khanam, R., Hussain, M., & Hill, R. (2025) · Discover Artificial Intelligence · doi:10.1007/s44163-025-00542-z · Cited by 3
“He et al. [27] introduced a multilevel feature fusion network (MFN) with a region proposal network (RPN) for steel surface defect detection, achieving 99.67% classification accuracy and 82.3% mAP on the NEU-DET dataset at 20 fps.”
InfluentialIF-DETR: Incremental Few-Shot Detection Transformer for Surface Defect Detection
Yang, X., Li, Z., Lin, X., Sheng, J., Hei, N., Dai, L., & Qi, L. (2025) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn64981.2025.11229065 · Cited by 1
“NEU-DET [36] dataset contains 1,800 steel surface defect images (200 × 200 pixels), covering six defect types: rolled-in scale (RS), patches (Pa), crazing (Cr), pitted surface (PS), inclusion (In), and scratches (Sc).”
Li, L., & Zhou, W. (2025) · Digital Signal Processing · doi:10.1016/j.dsp.2025.105776 · Cited by 4
FE-YOLO11n: a feature-enhanced detection model for multi-class surface damage of wind turbine blades
Chen, Y., Zeng, X., Zhang, J., et al. (2025) · Nondestructive Testing And Evaluation · doi:10.1080/10589759.2025.2593429 · Cited by 1
Attention-guided YOLOv5s-SDF model for accurate detection of strip steel surface defects
Liu, K., Chang, X., Ma, H., Lv, L., & Chen, H. (2025) · Scientific Reports · doi:10.1038/s41598-025-26313-5 · Cited by 3
“In order to verify the effectiveness of the proposed model, the NEU-DET dataset for surface defects of hot-rolled steel strips published by Northeastern University 18–20 is used, which includes six types of defects: Rolled-in scale, Patches, Crazing, Pitted_surface, Inclusion and Scratches.”
DFRD-Net: dual-domain feature representation dynamic network for steel surface defect detection
Wang, Y., Yang, X., Zhang, X., & Tao, Z. (2025) · Measurement Science and Technology · doi:10.1088/1361-6501/ae2b94 · Cited by 1
“To thoroughly validate the effectiveness and generalization performance of the proposed method, we used two popular public datasets, namely the NEU-DET [54] and GC10-DET [55] datasets, for steel surface defect detection.”
UnfoldDet: Advancing Surface Defect Detection With Dual Feature Separation and Relation Reasoning
Hou, X., Liu, M., & Du, S. (2025) · IEEE Transactions on Circuits and Systems for Video Technology · doi:10.1109/tcsvt.2025.3634684 · Cited by 1
“Some detailed settings on each dataset are given as follows: • NEU-DET : Following common practice [1], [19], we split the 1800 images at a ratio of 9:1 for training and validation.”
InfluentialZheng, T., Yu, L., Shi, Y., & Niu, F. (2025) · Journal of Computational Design and Engineering · doi:10.1093/jcde/qwaf138 · Cited by 1
A synergistic perception defect detector for complex spiral weld images
Cui, W., Song, K., Zhang, Y., et al. (2025) · Journal of Manufacturing Processes · doi:10.1016/j.jmapro.2025.12.052 · Cited by 1
Self-citationA Cross-Scale Feature Fusion Method for Effectively Enhancing Small Object Detection Performance
Kang, Y., ZHANG, Y., Ren, Y., et al. (2025) · Information · doi:10.3390/info17010025 · Cited by 1
Chen, Z., Fan, W., & Guo, B. Q. (2025) · IEEE Access · doi:10.1109/access.2025.3649757 · Cited by 1
“The NEU-DET dataset [43] is collected by Northeast-ern University and contains images of various steel surface defects.”
Yin, Z., Zhang, L., Liu, H., Du, Q., & Yu, C. (2026) · Frontiers in Computer Science · doi:10.3389/fcomp.2025.1700167 · Cited by 1
Jiang, Z., Shi, D., Pan, D., Yu, H., & Wang, Y. (2026) · IEEE Sensors Journal · doi:10.1109/jsen.2026.3651494 · Cited by 2
“This study is mainly based on the NEU-DET dataset [22], which provides rich data samples and accurate bounding boxes for defect regions.”
Jie, 徐. X., Yuang, 贾. J., & Zhifeng, 汪. W. (2026) · Laser & Optoelectronics Progress · doi:10.3788/lop251315 · Cited by 1
Lu, Q., Gong, M., & Yin, L. (2026) · Applied Sciences · doi:10.3390/app16031181 · Cited by 1
Jian, Z., Liu, S., Liu, J., Hu, J., Huang, J., & Liang, Y. (2026) · IEEE Transactions on Industrial Informatics · doi:10.1109/tii.2025.3650258 · Cited by 1
“ExperimentsonDeepPCB[10]andNEU-DET[11]showsu-periorsamplequality,diversity,andenhanceddetectionperfor-manceunderlow-samplesettings.”
Zhao, Z., Kang, J., Xu, Y., Liang, J., Feng, L., Ren, Y., Xue, T., & Wu, B. (2026) · Expert Systems with Applications · doi:10.1016/j.eswa.2026.131341 · Cited by 1
Chen, J., Cheng, Y., Wen, G., & Liu, X. (2026) · Neurocomputing · doi:10.1016/j.neucom.2026.132859 · Cited by 3
Zhang, Y., Xu, B., & Yang, G. (2026) · PeerJ Computer Science · doi:10.7717/peerj-cs.3485 · Cited by 1
MCD-RTDETR: a multi-category defect detection algorithm for steel surface
Huang, X., Qiu, Y., Luo, K., et al. (2026) · The Journal of Supercomputing · doi:10.1007/s11227-025-08202-w · Cited by 1
“The NEU-DET [33] and DeepPCB [34] datasets are used for cross-dataset generalization evaluation, containing 1800 and 1500 images, respectively, with 8:1:1 train/validation/test splits.”
A YOLOv8-Based Network for Surface Defect Detection in Strip Steel
Jiang, J., Su, Y., & Song, J. (2026) · Journal of Computer and Communications · doi:10.4236/jcc.2026.142006 · Cited by 1
SCS-YOLOv9: a novel YOLOv9 framework for small object detection via edge computing
Liu, X., Wu, G., Zhou, T., Wang, Z., Wang, K., Li, J., Xie, C., & Liu, T. (2026) · International Journal of Machine Learning and Cybernetics · doi:10.1007/s13042-025-02848-x · Cited by 1
“Ren et al. [20] improved RCNN’s efficiency using depthwise separable convolutions [21], while He et al. [22] utilized multi-level features and a region proposal network to achieve high accuracy.”
DSA-DAMB: A Dynamic Sparse Attention and Defect-Aware Memory Approach for Rare Industrial Defects
Wei, M., Guo, W., Gao, L., Tu, Z., & Liu, Y. (2025) · International Conference on Artificial Intelligence and Computer Engineering (ICAICE) · doi:10.1109/icaice68195.2025.11382445 · Cited by 1
“Each image has a resolution of 200 × 200 , and the dataset was divided into a training set and a testing set following the original NEU-DET protocol [19].”
Xu, W., Wen, Z., Tao, G., Wu, X., Yang, C., Xie, Y., Cai, W., Liu, J., Chi, M., & Liang, S. (2026) · Measurement · doi:10.1016/j.measurement.2026.120847 · Cited by 1
Benchmarking Few-Shot Learning Techniques for Steel Surface Defect Detection
Ghali, R., Benhafid, Z., & Selouani, S. (2025) · IEEE Smart World Congress (SWC) · doi:10.1109/swc65939.2025.00031 · Cited by 1
“We focus on approaches evaluated using the widely adopted Northeastern University Surface Defect Database (NEU-DET), a standard benchmark in this field [5], [8].”
MCD-YOLO: An improved method for detecting mixed-class defects on steel plate surfaces
Li, J., Chu, M., Zong, P., & Ma, S. (2026) · Engineering Applications of Artificial Intelligence · doi:10.1016/j.engappai.2026.114282 · Cited by 3
SDCM-YOLO: a cross-scale, texture-aware framework for metal surface defect detection
Deng, F., Chen, W., & Fu, L. (2026) · Engineering Research Express · doi:10.1088/2631-8695/ae51e4 · Cited by 1
Efficient Detection Algorithm of Steel Surface Defects Based on YOLOv8
Liu, S., Miao, X., & Zhou, Y. (2026) · JOM · doi:10.1007/s11837-026-08174-z · Cited by 1
Gao, S., Xia, T., Hu, R., Si, W., Zhu, Y., & Xi, L. (2026) · Ocean Engineering · doi:10.1016/j.oceaneng.2026.125137 · Cited by 3
Tong, J., Zhang, X., Lu, X., Cao, H., Yao, L., & Gao, B. (2026) · Sensors · doi:10.3390/s26072132 · Cited by 1
“Regarding the two-stage method, He et al. [4] constructed a steel surface defect detection framework based on Faster R-CNN, and proposed a multi-level feature fusion network (MFN) to integrate features from different stages of the backbone network across layers.”
A lightweight end-to-end fabric defect detection network with dynamic multi-scale fusion
Liang, J., Wang, N., Cao, Y., et al. (2026) · Journal of Intelligent Manufacturing · doi:10.1007/s10845-026-02868-6 · Cited by 1
A review of deep learning-based surface defect detection techniques for strip steel
Shang, F., Chen, H., Sun, B., Miao, R., Zhou, X., Xing, C., & Zhang, P. (2026) · Measurement · doi:10.1016/j.measurement.2026.121905 · Cited by 7
Ruan, J., Cheng, Z., & Zeng, M. (2026) · Journal of Real-Time Image Processing · doi:10.1007/s11554-026-01920-1 · Cited by 3
“NEU-DET [36] provides 1800 grayscale images of six common hot-rolled steel defects (Cr, In, Pc, Ps, Rs, Sc) at 200 × 200 resolution, split 7:1:2 for training, validation, and testing.”
Sun, Y., Wang, M., Xu, C., Guo, H., Zhang, L., Yue, H., & Feng, Y. (2026) · Journal of Imaging · doi:10.3390/jimaging12070312 · Cited by 2
“…losses use cross-entropy, and the regression losses adopt the smooth L1 loss: To evaluate the effectiveness of the proposed A 2 S 2 C-Det framework, three publicly available steel strip surface defect datasets are utilized, namely NEU-DET [40,41], GC10-DET [42], and SD-saliency-900 [43].”
FSWD-YOLO: a lightweight attention-enhanced YOLO framework for FSW surface defect detection
Zhou, Y., Zhu, Y., Wang, L., et al. (2026) · Welding in the World · doi:10.1007/s40194-026-02556-2 · Cited by 1
Automobile airbag defect detection algorithm based on improved Faster RCNN
Luo, L., Deng, C., Wu, Z., et al. (2021) · International Conference on Computer Engineering and Artificial Intelligence (ICCEAI) · doi:10.1109/icceai52939.2021.00056 · Cited by 1
Derinlemesine Özellik Piramit Ağı Kullanarak Yüzey Hata Tespiti
Üzen, H., İlhami, S. E. L., Türkoğlu, M., & Hanbay, D. (2021) · Computer Science · doi:10.53070/bbd.990950 · Cited by 1
Automated Shell Abrasion Detection for Cell Phone Based on GraphFPN With Learnable Sparse Priors
Wu, H., Guo, X., Xiao, F., Yan, W., & Wen, Z. (2022) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2022.3150591 · Cited by 2
Substation Defect Detection Network Based on the Attention Model
Wang, Z., Wang, P., Wang, Y., et al. (2022) · Journal of Physics Conference Series · doi:10.1088/1742-6596/2213/1/012019 · Cited by 1
“[15] proposed an end-to-end steel surface defect detection method and achieved good results, but it needed a lot of manual marking.”
BackgroundDetection of the Steel Faults Based on Deep Learning
Karaduman, G., Aydın, İ., Akın, E., & Ozdemir, S. (2022) · International Conference on Innovations in Intelligent Systems and Applications (INISTA) · doi:10.1109/inista55318.2022.9894131 · Cited by 1
“He et al.[8] performed an application by combining CNN and multilevel feature fusion network (MFN) methods to detect only the surface deteriorations in steel plates.”
MethodA Method for Detection of Steel Coil End Face Defects Based on YX-FR Ensemble Model
Li, Y., Lin, S. Z., & Xue, H. (2022) · International Conference on Pattern Recognition and Artificial Intelligence (PRAI) · doi:10.1109/prai55851.2022.9904226 · Cited by 1
“Among them, He et al. used the improved Faster RCNN network to achieve a mAP of 82.3% on the strip steel surface defect dataset [9].”
MethodInfluentialΨ-Net is an Efficient Tiny Defect Detector
Wang, B., Zhou, H., Luo, W., Li, C., Li, Z., & Tian, Z. (2022) · Proceedings - International Conference on Image Processing · doi:10.1109/icip46576.2022.9897524 · Cited by 1
“Comparison result of our model Ψ -Net, BYEC [17], OVERFEAT [18], AECLBP [8], Decaf [19], ResNet34 [1], ResNet-50 [1] and ResNet-50+MFN [7] on NEU-CLS dataset.”
BackgroundDeep Learning-Based Quality Inspection System for Steel Sheet
Sambath, M., Reddy, C. S. B., Reddy, Y. K., Reddy, M. M. S., Kathiravan, M., & Ravi, S. (2022) · Algorithms for intelligent systems · doi:10.1007/978-981-19-6004-8_16 · Cited by 1
A Cascaded Twin Dense Network Based Metallic Surface Defect Detection Algorithm
Wang, D., Liu, Y., Du, W., & Zhu, R. (2022) · Journal of Computer-Aided Design & Computer Graphics · doi:10.3724/sp.j.1089.2022.19056 · Cited by 4
“The International Journal of Advanced Manufacturing Technology, 2017, 90(5-8): 1665-1678 [4] He Y, Song K C, Meng Q G, et al.”
BackgroundSharma, M., & Maity, T. (2022) · IEEE Conference on Information and Communication Technology (CICT) · doi:10.1109/cict56698.2022.9997894 · Cited by 1
“In [11], the authors proposed a multisensory fusion model to make reliable decisions”
BackgroundZou, M., You, M., Kageyama, Y., & Akashi, T. (2023) · IEEJ Transactions on Electrical and Electronic Engineering · doi:10.1002/tee.23776 · Cited by 1
Song, K., Wang, H., Guo, X., Sun, M., Shao, Y., Xue, S., Zhang, H., & Zhang, T. (2023) · Bioengineering · doi:10.3390/bioengineering10040450 · Cited by 2
“At this stage, with the rapid development of deep learning technology, many achievements have been made in the following fields: computer vision [9,10], object detection [11], natural language processing (NLP) [12], and speech recognition [13,14].”
BackgroundInfluentialSelf-citationA Deep Learning Based End-to-end Surface Defect Detection Method for Industrial Scenes
Lan, Y., & Huang, C. (2023) · International Conference on Bioinformatics and Intelligent Computing (BIC) · doi:10.1145/3592686.3592695 · Cited by 2
“Defect detection networks based on detection methods can be structurally divided into two stage networks, represented by Faster R-CNN [13, 14], and one stage networks, represented by SSD or YOLO.”
MethodChen, W. R., Zou, B., Zheng, Q., Huang, C., Li, L., & Liu, J. (2023) · The International Journal of Advanced Manufacturing Technology · doi:10.1007/s00170-023-11762-x · Cited by 1
“In addition, other scholars have done some research on data fusion [11, 12].”
BackgroundXiang, Y. H., Hu, Y., Xuzhou, F., Gao, J., & Zhiqiang, L. (2023) · Lecture notes in computer science · doi:10.1007/978-981-99-4761-4_16 · Cited by 1
Bi, C., & He, S. (2023) · IEEE International Conference on Electronics Information and Emergency Communication · doi:10.1109/iceiec58029.2023.10199610 · Cited by 1
“We consider a steel surface defect dataset (NEU-DET) [9].”
MethodInfluentialSmall target particleboard defect detection algorithm based on improved YOLO
Cheng, J., Chen, X., Hou, Q., Zhang, J., & Liu, T. (2023) · Chinese Control Conference (CCC) · doi:10.23919/ccc58697.2023.10240623 · Cited by 1
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.