An improved you only look once model for the multi-scale steel surface defect detection with multi-level alignment and cross-layer redistribution features

Abstract: Steel surface defects can significantly compromise the safety, durability, and cost efficiency of steel structures; however, vision-based automated inspection remains challenging due to complex backgrounds, scale variation, and low defect-background contrast. To address these challenges, we propose YOLOv8-FE (You Only Look Once v8—Feature Enhancement), an enhanced detector built on YOLOv8. The proposed model comprises three modules: multi-scale feature enhancement (MFE), multi-scale feature fusion (MFF), and convolution-attention aggregation (CAA). Together, these modules enhance defect representation and contextual modeling. Experiments on the NEU-DET and GC10-DET datasets demonstrate consistent performance gains. On NEU-DET, YOLOv8-FE achieves 78.0% mAP50, 57.3% recall, and 87.2% precision, corresponding to absolute gains of 2.8%, 0.7%, and 4.4%, respectively. On GC10-DET, it achieves 72.5% mAP50, 55.4% recall, and 88.1% precision, with absolute gains of 4.4%, 2.9%, and 1.5%, respectively. These results indicate that YOLOv8-FE enhances the accuracy and robustness of steel surface defect detection, providing an effective solution for related industrial visual inspection tasks.

In order to address the challenges of deployment difficulties and low small-object detection efficiency in current deep learning-based defect detection models on terminal devices with limited computational capacity, this paper proposes a lightweight steel surface defect detection model, Pyramid-based Small-target Fusion YOLO (PSF-YOLO), based on an improved YOLOv11n object detection framework. The model employs a low-parameter Ghost convolution (GhostConv) to substantially reduce the required computational resources. Additionally, the traditional feature pyramid network structure is replaced with a Multi-Dimensional-Fusion neck (MDF-Neck) to enhance small-object perception and reduce the number of model parameters. Moreover, to achieve multi-dimensional integration in the neck, a Virtual Fusion Head is utilized, and the design of an Attention Concat module further improves target feature extraction, thereby significantly enhancing overall detection performance. Experimental results on the GC10-DET+ dataset demonstrate that PSF-YOLO reduces model parameters by 25% while achieving improvements of 3.2% and 3.3% in mAP_{50} and mAP_{50-95}, respectively, compared to the baseline model. This approach offers valuable insights and practical applicability for deploying defect detection models on terminal devices with limited computational resources.

Surface defect detection is essential for ensuring the quality and safety of steel products. While Transformer-based methods have achieved state-of-the-art performance, they face several limitations, including high computational costs due to the quadratic complexity of the attention mechanism, inadequate detection accuracy for small-scale defects due to substantial downsampling, inconsistencies between classification scores and localization confidence, and feature resolution loss caused by simple upsampling and downsampling strategies. To address these challenges, we propose the HCT-Det model, which incorporates a window-based self-attention residual (WSA-R) block structure. This structure combines window-based self-attention (WSA) blocks to reduce computational overhead and parallel residual convolutional (Res) blocks to enhance local feature continuity. The model's backbone generates three cross-scale features as encoder inputs, which undergo Intra-Scale Feature Interaction (ISFI) and Cross-Scale Feature Interaction (CSFI) to improve detection accuracy for targets of various sizes. A Soft IoU-Aware mechanism ensures alignment between classification scores and intersection-over-union (IoU) metrics during training. Additionally, Hybrid Downsampling (HDownsample) and Hybrid Upsampling (HUpsample) modules minimize feature degradation. Our experiments demonstrate that HCT-Det achieved a mean average precision (mAP@0.5) of 0.795 on the NEU-DET dataset and 0.733 on the GC10-DET dataset, outperforming other state-of-the-art approaches. These results highlight the model's effectiveness in improving computational efficiency and detection accuracy for steel surface defect detection.

Surface defects in industrial production are complex and diverse. Therefore, deep learning-based defect detection models must consistently adapt to newly emerging defect categories. The trained models generally suffer from catastrophic forgetting as they learn new defect categories. To address this issue, we propose a selective distillation for incremental defect detection (SD-IDD) model based on GFLv1. Specifically, three selective distillation strategies are proposed, including high-confidence classification distillation, dual-stage cascaded regression distillation, and Intersection over Union (IoU)-driven difficulty-aware feature distillation. The high-confidence classification distillation aims to preserve critical discriminative knowledge of old categories within semantic confusion regions of the classification head, reducing interference from low-value regions. Dual-stage cascaded regression distillation focuses on high-quality anchors through geometric prior coarse filtering and statistical fine filtering, utilizing IoU-weighted KL divergence distillation loss to accurately transfer localization knowledge. IoU-driven difficulty-aware feature distillation adaptively allocates distillation resources, prioritizing features of high-difficulty targets. These selective distillation strategies significantly mitigate catastrophic forgetting while enhancing the detection accuracy of new classes, without requiring access to old training samples. Experimental results demonstrate that SD-IDD achieves superior performance, with mAP_old of 58.2% and 99.3%, mAP_new of 69.0% and 97.3%, and mAP_all of 63.6% and 98.3% on the NEU-DET and DeepPCB datasets, respectively, surpassing existing incremental detection methods.

Surface defects have a serious detrimental effect on the quality of steel. To address the problems of low efficiency and poor accuracy in the manual inspection process, intelligent detection technology based on machine learning has been gradually applied to the detection of steel surface defects. An improved YOLOv8 steel surface defect detection model called YOLOv8-MGVS is designed to address these challenges. The MLCA mechanism in the C2f module is applied to increase the feature extraction ability in the backbone network. The lightweight GSConv and VovGscsp cross-stage fusion modules are added to the neck network to reduce the loss of semantic information and achieve effective information fusion. The self-attention mechanism is exploited into the detection network to improve the detection ability of small targets. Defect detection experiments were carried out on the NEU-DET dataset. Compared with YOLOv8n from experimental results, the average accuracy, recall rate, and frames per second of the improved model were improved by 5.2%, 10.5%, and 6.4%, respectively, while the number of parameters and computational costs were reduced by 5.8% and 14.8%, respectively. Furthermore, the defect detection generalization experiments on the GC-10 dataset and SDD DET dataset confirmed that the YOLOv8-MGVS model has higher detection accuracy, better lightweight, and speed.

Insulator defect detection is a crucial task in the intelligent inspection of transmission lines. Currently, there are challenges such as insufficient image samples and difficulties in annotation. Artificially generating samples is a feasible solution, but there are discrepancies between artificial and real sample distributions. This article proposes a cross-domain multilevel feature alignment R-CNN network. It uses a large number of artificially labeled images and a small number of unlabeled real images as the source and target domains, respectively. Based on Faster R-CNN, an instance-level feature adaptive alignment module is constructed. Different gradient adaptive training strategies are employed for the source and target domains to better achieve cross-domain instance-level feature alignment. An image-level multiscale local feature aggregation (MLA) module is built to achieve cross-domain image-level local feature alignment. A global feature alignment (GFA) module is also constructed to achieve cross-domain image-level global feature alignment. In the insulator defect detection experiment, the average precision (AP) at the Intersection over Union of 0.5 (AP50) of the proposed method is 6.8% higher than that of the baseline model, 8.6% higher than that of Faster-RCNN trained with artificial samples alone, and 1.1% higher than that of Faster-RCNN trained with real samples alone. Moreover, the proposed method achieves the highest accuracy on both self-built and public datasets, exhibiting an average improvement of 7.7% and 8.6% in AP50, respectively, compared to multiple comparable methods. The code for this article can be found at <uri xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">https://github.com/n-tong/CMFAA-R-CNN</uri>.

Cas-VSwin transformer: A variant swin transformer for surface-defect detection

The speed and load capacity of trains are improving at an accelerating rate, which raises the standards for railway services' security criteria. The track's surface may progressively reveal varying degrees of defects because of the impacts of moisture, temperature, load, and other factors; if the flaws never addressed in a prompt period, the degree of defects would expand, significantly raising the probability of train service. The development of automatic RSDD (Rail Surface Defect Detection) has substantial practical and scientific implications. In this study, a multi-crack detection technique depends on Variational Autoencoder Diffusion Model (VAEDM), The introduction of VAEDM allows for the non-destructive detection of fastener and RSD. In order to produce hidden elements from images that have various cracks, VAEDM combines the multi-layer and a rail surface image encoder was introduced. Zero Shot- Divergence Asynchronous Reinforcement Learning (ZS-DARL), Policy-Gradient is used to establish the relationship between defects and non-defects. DARL transformer encoder is introduced for taking long-range dependences for FE (Feature Extraction) from detected objects images with various patterns and dimensions. The simulation outcomes on the open Railway Track Fault Detection (RTFD) and Rail Surface Defect Datasets (RSDDs) with rail surface defects are collected from rail tracks surface defect detection. Results are measured using the metrics like recall, precision, F-measure, and accuracy.

Nowadays, domain adaptation (DA) is getting more attention in cross-scene hyperspectral image (HSI) classification, and various DA algorithms have been proposed. However, regular convolution indiscriminately extracting features around the center pixel will result in the inaccurate extraction of spatial-spectral features, which significantly affect the subsequent feature alignment. Meanwhile, the method of aligning the category features of source and target domains from a single-level may not cope well with complex HSIs. Therefore, we propose a multilevel feature alignment algorithm based on spatial attention deformable convolution (MFA-SADC), which achieves multilevel feature alignment from feature to feature, feature to cluster-center, and cluster-center to cluster-center. In addition, spatial attention deformable convolution is proposed to compose the feature extraction network of MFA-SADC, which guarantees the purity of spatial-spectral features. Experiments on three HSI datasets indicate MFA-SADC can obtain better classification performance when compared with the seven state-of-the-art methods.

Due to the complexity of steel processing environments, surface defects inevitably occur during production. Detecting these defects is critical for ensuring product quality and industrial safety. Traditional manual inspection methods suffer from inefficiency and subjectivity, while existing algorithms struggle with feature extraction in complex scenarios. We propose a novel steel surface defect detection model YOLODDIM-DWConv-C3 based on YOLOv9, which enhances feature extraction capabilities while significantly reducing computational complexity. To address the scarcity of original data, we employ the Denoising Diffusion Implicit Model DDIM for data augmentation. The proposed YOLO based defect detection model minimizes computational demands, enabling seamless deployment on edge devices for real-time defect monitoring. Experimental results on the NEU-DET dataset demonstrate that YOLO-DDC outperforms existing methods in both detection accuracy and computational efficiency.We have published the complete project at https: github.com zhzhzsword YOLO-DDC.

Currently, object detection-based rail fastener defect detection methods still face challenges such as limited detection categories, insufficient accuracy, and high computational complexity. To this end, the YOLOv8n-FDD, an advanced multi-category fastener defect detection model designed upon the YOLOv8n with comprehensive optimizations is developed in this paper. Concretely, by introducing the CUT-based style transfer model to generate diverse defect samples, the concern due to imbalanced distribution of sample categories is effectively alleviated. The CA mechanism is incorporated to enhance the feature extraction capability, and the bounding box loss function is further upgraded to improve the model’s generalization performance. With respect to efficiency, the Conv and c2f modules of the YOLOv8n model are, respectively, replaced with the GSConv and VoVGSPCP modules, accordingly achieving a lightweight design. Comparative experimental results demonstrate that the presented YOLOv8n-FDD model outperforms several classic object detection models in terms of detection accuracy, detection speed, model size, and computational complexity.

Increasing the robustness of material-specific deep learning models for crack detection across different materials

Ensuring industrial quality critically depends on reliable steel surface defect detection. Although deep learning-based object detection has achieved notable progress, challenges such as substantial computational overhead, limited real-time adaptability, and suboptimal feature extraction efficiency persist. To address these issues, we propose a lightweight detection model, MCS-YOLOv11. The model adopts MobileNetv3 as the backbone, in which the original Squeeze-and-Excitation (SE) module is replaced by the proposed Depthwise Convolution Attention Module (DSAM) to enhance both channel and spatial feature representation. In addition, a Space-to-Depth (SPD) module is integrated to replace conventional down-sampling, thereby reducing information loss and strengthening multi-scale feature extraction. Experimental results on the NEU-DET dataset demonstrate that MCS-YOLOv11 reduces parameters by 47.1% and GFLOPs by 84.8%, while improving mAP accuracy by 2.8 percentage points compared with the baseline model, achieving a more favorable balance between computational efficiency and detection accuracy.

Software defects and quality assurance are crucial aspects of software development that should be considered during the software development cycle. To ensure high-quality software, it is essential to have a robust quality assurance process in place. System reliability and quality are very key components that must be considered during software development, and this can only be achieved when software undergoes a thorough test process for errors, anomalies, defects, omissions, and bugs. Early software defect prediction and detection play an essential role in ensuring the reliability and quality of software systems, ensuring that software companies discover errors or defects early enough and allocate more resources to defect-prone modules. This study proposes the development of an enhanced classifier model for software defect prediction and detection. The aim is to harness the collective intelligence of selected base classifiers like Support Vector Machine, Logistic regression, Decision Trees, Random Forest, AdaBoost, Gradient Boosting, K-Nearest Neighbor, GaussianNB, and Multi-Layer Perception to improve accuracy, robustness, and generalization in identifying potential defects using a soft voting ensemble technique. The ensemble model leveraged the confidence probability of the soft voting technique and the generalization advantage of cross-validation leading to a more robust and dynamic model. The performance of the model with existing classifiers was evaluated using accuracy, F1 score, Precision, and area under the ROC curve (ROC- AUC) as the evaluation metrics. The results of the experiment revealed that the Proposed Classifier produced an overall Accuracy rate of 93%, and ROC AUC of 98%. The results demonstrate the effectiveness of our enhanced ensemble classifier in software defect detection and prediction. By harnessing the strengths of diverse base classifiers, our approach provides a robust and adaptive solution to the challenges of early detection and mitigating defects in software systems. This research contributes to the advancement of reliable software development practices and lays the foundation for future enhancements in ensemble-based defect detection methodologies.

Automatic sleep staging is essential for sleep assessment and disorder diagnosis. Most existing methods depend on one specific dataset and are limited to be generalized to other unseen datasets, for which the training data and testing data are from the same dataset. In this paper, we introduce domain generalization into automatic sleep staging and propose the task of generalizable sleep staging which aims to improve the model generalization ability to unseen datasets. Inspired by existing domain generalization methods, we adopt the feature alignment idea and propose a framework called SleepDG to solve it. Considering both of local salient features and sequential features are important for sleep staging, we propose a Multi-level Feature Alignment combining epoch-level and sequence-level feature alignment to learn domain-invariant feature representations. Specifically, we design an Epoch-level Feature Alignment to align the feature distribution of each single sleep epoch among different domains, and a Sequence-level Feature Alignment to minimize the discrepancy of sequential features among different domains. SleepDG is validated on five public datasets, achieving the state-of-the-art performance.