Surface defects on strip steel have characteristics such as low contrast, large intra⁃class differences, and inter⁃class similarity, which bring significant challenges to the accuracy of model detection. To address these challenges, a salient object detection model based on a triple⁃attention mechanism was proposed to achieve more accurate detection of surface defects on strip steel. A feature refinement module, consisting of three branches, was constructed for the coarse defect feature maps extracted from the model backbone network. Each branch was designed to focus on the channel and spatial features of surface defects on the strip steel from different perspectives. Additionally, to refine the edge details of salient targets, a local attention feature filter module was built and embedded into the feature refine module as a plugin, which enhances the model’s sensitivity in defect areas. The experimental results indicated that the proposed model had significant advantages over existing methods in salient object detection and outperforms similar methods in evaluation indicators such as mean absolute error and weighted F-measure, thereby verifying the effectiveness and robustness of the proposed method.
TANGJun, WANGZidong, ZHANGHongyi, et al. A lightweight surface defect detection framework combined with dual-domain attention mechanism[J]. Expert Systems with Applications, 2024, 238(Part A): 121726.
[2]
LEITianyang, YANGKewei, LIJichao, et al. Multichannel spatial⁃temporal graph convolution network based on spectrum decomposition for traffic prediction[J]. Expert Systems with Applications, 2024, 238(Part F): 122281.
[3]
VIEIRA E SILVAA L, SIMÕESF, KOWERKOD, et al. Attention modules improve image‑level anomaly detection for industrial inspection: A DifferNet case study[C]//2024 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), Waikoloa, 2024: 8231⁃8240.
[4]
DE BOERP T, KROESED P, MANNORS, et al. A tutorial on the cross‑entropy method[J]. Annals of Operations Research, 2005, 134(1): 19⁃67.
[5]
RAHMANM A, WANGYang. Optimizing intersection‑over‑union in deep neural networks for image segmentation[C]//Advances in Visual Computing, Cham, 2016: 234⁃244.
[6]
SONGGuorong, SONGKechen, YANYunhui. Saliency detection for strip steel surface defects using multiple constraints and improved texture features[J]. Optics and Lasers in Engineering, 2020, 128: 106000.
[7]
DENGJia, DONGWei, SOCHERR, et al. ImageNet: A large⁃scale hierarchical image database[C]//2009 IEEE Conference on Computer Vision and Pattern Recognition, Miami, 2009: 248⁃255.
[8]
KINGMAD, BAJ. Adam: A method for stochastic optimization[J]. Computer Science, 2014, 6:980-991.
[9]
ACHANTAR, HEMAMIS, ESTRADAF, et al. Frequency tuned salient region detection[C]//2009 IEEE Conference on Computer Vision and Pattern Recognition, Miami, 2009: 1597⁃1604.
[10]
PERAZZIF, KRÄHENBÜHLP, PRITCHY, et al. Saliency filters: Contrast based filtering for salient region detection[C]//2012 IEEE Conference on Computer Vision and Pattern Recognition, Providence, 2012: 733⁃740.
[11]
CHENGMingming, FANDengping. Structure‑measure: A new way to evaluate foreground maps[J]. International Journal of Computer Vision, 2021, 129(9): 2622⁃2638.
[12]
YUANYuchen, LIChangyang, KIMJ, et al. Reversion correction and regularized random walk ranking for saliency detection[J]. IEEE Transactions on Image Processing, 2018, 27(3): 1311⁃1322.
[13]
ZHOULi, YANGZhaohui, ZHOUZongtan, et al. Salient region detection using diffusion process on a two layer sparse graph[J]. IEEE Transactions on Image Processing, 2017, 26(12): 5882⁃5894.
[14]
ZHUWangjiang, LIANGShuang, WEIYichen, et al. Saliency optimization from robust background detection[C]//2014 IEEE Conference on Computer Vision and Pattern Recognition, Columbus, 2014: 2814⁃2821.
[15]
ZHAOTing, WUXiangqian. Pyramid feature attention network for saliency detection[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 2019: 3080⁃3089.
[16]
LUOZhiming, MISHRAA, ACHKARA, et al. Non-local deep features for salient object detection[C]//2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Honolulu, 2017: 6593⁃6601.
[17]
DENGZijun, HUXiaowei, ZHULei, et al. R3net: Recurrent residual refinement network for saliency detection[C]//Proceedings of the Twenty‑Seventh International Joint Conference on Artificial Intelligence, Stockholm, 2018: 684‑690.
[18]
LIUNian, HANJunwei, YANGM H. PiCANet: Learning pixel wise contextual attention for saliency detection[C]//2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, Salt Lake City, 2018: 3089⁃3098.
[19]
ZHOUXiaofei, FANGHao, LIUZhi, et al. Dense attention⁃guided cascaded network for salient object detection of strip steel surface defects[J]. IEEE Transactions on Instrumentation and Measurement, 2022, 71: 1⁃14.
[20]
QINXuebin, ZHANGZichen, HUANGChenyang, et al. BASNet: Boundary⁃aware salient object detection[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 2019: 7471⁃7481.
[21]
WUZhe, SULi, HUANGQingming. Cascaded partial decoder for fast and accurate salient object detection[C]//2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), Long Beach, 2019: 3902⁃3911.
[22]
SONGGuorong, SONGKechen, YANYunhui. EDRNet: Encoder-decoder residual network for salient object detection of strip steel surface defects[J]. IEEE Transactions on Instrumentation and Measurement, 2020, 69(12): 9709⁃9719.