{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,2,26]],"date-time":"2026-02-26T06:40:19Z","timestamp":1772088019092,"version":"3.50.1"},"reference-count":44,"publisher":"MDPI AG","issue":"5","license":[{"start":{"date-parts":[[2026,2,24]],"date-time":"2026-02-24T00:00:00Z","timestamp":1771891200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"DOI":"10.13039\/501100001809","name":"the National Natural Science Foundation of China","doi-asserted-by":"crossref","award":["Grant No. 52575581"],"award-info":[{"award-number":["Grant No. 52575581"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"crossref"}]},{"name":"the Qingdao Natural Science Foundation","award":["Grant No. 24-8-4-zrjj-4-jch"],"award-info":[{"award-number":["Grant No. 24-8-4-zrjj-4-jch"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["Sensors"],"abstract":"<jats:p>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.<\/jats:p>","DOI":"10.3390\/s26051413","type":"journal-article","created":{"date-parts":[[2026,2,24]],"date-time":"2026-02-24T13:53:20Z","timestamp":1771941200000},"page":"1413","update-policy":"https:\/\/doi.org\/10.3390\/mdpi_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["SD-IDD: Selective Distillation for Incremental Defect Detection"],"prefix":"10.3390","volume":"26","author":[{"given":"Jing","family":"Li","sequence":"first","affiliation":[{"name":"School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266000, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Chenggang","family":"Dai","sequence":"additional","affiliation":[{"name":"School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266000, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xiaobin","family":"Wang","sequence":"additional","affiliation":[{"name":"School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266000, China"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3185-1062","authenticated-orcid":false,"given":"Chengjun","family":"Chen","sequence":"additional","affiliation":[{"name":"School of Mechanical and Automotive Engineering, Qingdao University of Technology, Qingdao 266000, China"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"1968","published-online":{"date-parts":[[2026,2,24]]},"reference":[{"key":"ref_1","doi-asserted-by":"crossref","first-page":"046106","DOI":"10.1088\/1361-6501\/adbf86","article-title":"Efficient progressive aggregation enhancement network for defect detection","volume":"36","author":"Yu","year":"2025","journal-title":"Meas. Sci. Technol."},{"key":"ref_2","doi-asserted-by":"crossref","unstructured":"Saberironaghi, A., Ren, J., and El-Gindy, M. (2023). Defect detection methods for industrial products using deep learning techniques: A review. Algorithms, 16.","DOI":"10.3390\/a16020095"},{"key":"ref_3","doi-asserted-by":"crossref","first-page":"22136","DOI":"10.1038\/s41598-023-49359-9","article-title":"Surface defect detection of industrial components based on vision","volume":"13","author":"Chen","year":"2023","journal-title":"Sci. Rep."},{"key":"ref_4","doi-asserted-by":"crossref","first-page":"085022","DOI":"10.1088\/1361-6501\/adfba2","article-title":"An enhanced YOLOv11 model combined with dual-path scale network for high-accuracy PCB defect detection","volume":"36","author":"Liu","year":"2025","journal-title":"Meas. Sci. Technol."},{"key":"ref_5","first-page":"5016613","article-title":"A Domain Incremental Learning Framework for PCB Continuous Defect Detection","volume":"74","author":"Yan","year":"2025","journal-title":"IEEE Trans. Instrum. Meas."},{"key":"ref_6","doi-asserted-by":"crossref","unstructured":"Mohandas, R., Southern, M., O\u2019Connell, E., and Hayes, M. (2024). A survey of incremental deep learning for defect detection in manufacturing. Big Data Cogn. Comput., 8.","DOI":"10.3390\/bdcc8010007"},{"key":"ref_7","doi-asserted-by":"crossref","first-page":"046208","DOI":"10.1088\/1361-6501\/adbf34","article-title":"Incremental rotating machinery fault diagnosis method based on multi-scale knowledge distillation and label smoothing","volume":"36","author":"Xia","year":"2025","journal-title":"Meas. Sci. Technol."},{"key":"ref_8","unstructured":"Hinton, G., Vinyals, O., and Dean, J. (2015). Distilling the knowledge in a neural network. arXiv."},{"key":"ref_9","doi-asserted-by":"crossref","first-page":"109337","DOI":"10.1016\/j.compeleceng.2024.109337","article-title":"Adaptive dual teacher incremental learning for defect detection of flexible printed circuit","volume":"118","author":"Xiong","year":"2024","journal-title":"Comput. Electr. Eng."},{"key":"ref_10","doi-asserted-by":"crossref","first-page":"e1583","DOI":"10.7717\/peerj-cs.1583","article-title":"Class incremental learning of remote sensing images based on class similarity distillation","volume":"9","author":"Shen","year":"2023","journal-title":"PeerJ Comput. Sci."},{"key":"ref_11","doi-asserted-by":"crossref","first-page":"046001","DOI":"10.1088\/1361-6501\/adbc10","article-title":"A lightweight convolutional neural network for surface defect detection in strip steel","volume":"36","author":"Yang","year":"2025","journal-title":"Meas. Sci. Technol."},{"key":"ref_12","doi-asserted-by":"crossref","unstructured":"Redmon, J., Divvala, S., Girshick, R., and Farhadi, A. (2016, January 27\u201330). You only look once: Unified, real-time object detection. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Las Vegas, NV, USA.","DOI":"10.1109\/CVPR.2016.91"},{"key":"ref_13","doi-asserted-by":"crossref","unstructured":"Redmon, J., and Farhadi, A. (2017, January 21\u201326). YOLO9000: Better, faster, stronger. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Honolulu, HI, USA.","DOI":"10.1109\/CVPR.2017.690"},{"key":"ref_14","unstructured":"Redmon, J., and Farhadi, A. (2018). Yolov3: An incremental improvement. arXiv."},{"key":"ref_15","doi-asserted-by":"crossref","unstructured":"Wang, C.Y., Bochkovskiy, A., and Liao, H.Y.M. (2023, January 17\u201324). YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, BC, Canada.","DOI":"10.1109\/CVPR52729.2023.00721"},{"key":"ref_16","doi-asserted-by":"crossref","unstructured":"Liu, W., Anguelov, D., Erhan, D., Szegedy, C., Reed, S., Fu, C.Y., and Berg, A.C. (2016). SSD: Single shot multibox detector. European Conference on Computer Vision, Springer International Publishing.","DOI":"10.1007\/978-3-319-46448-0_2"},{"key":"ref_17","doi-asserted-by":"crossref","unstructured":"Lin, T.Y., Goyal, P., Girshick, R., He, K., and Doll\u00e1r, P. (2017, January 22\u201329). Focal loss for dense object detection. Proceedings of the IEEE International Conference on Computer Vision, Venice, Italy.","DOI":"10.1109\/ICCV.2017.324"},{"key":"ref_18","first-page":"21002","article-title":"Generalized focal loss: Learning qualified and distributed bounding boxes for dense object detection","volume":"33","author":"Li","year":"2020","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"ref_19","doi-asserted-by":"crossref","unstructured":"Girshick, R., Donahue, J., Darrell, T., and Malik, J. (2014, January 23\u201328). Rich feature hierarchies for accurate object detection and semantic segmentation. Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, Columbus, OH, USA.","DOI":"10.1109\/CVPR.2014.81"},{"key":"ref_20","doi-asserted-by":"crossref","unstructured":"Girshick, R. (2015, January 13\u201316). Fast R-CNN. Proceedings of the IEEE International Conference on Computer Vision, Santiago, Chile.","DOI":"10.1109\/ICCV.2015.169"},{"key":"ref_21","doi-asserted-by":"crossref","first-page":"1137","DOI":"10.1109\/TPAMI.2016.2577031","article-title":"Faster R-CNN: Towards real-time object detection with region proposal networks","volume":"39","author":"Ren","year":"2016","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_22","doi-asserted-by":"crossref","first-page":"37643","DOI":"10.1109\/ACCESS.2024.3374869","article-title":"Steel surface defect detection method based on improved YOLOX","volume":"12","author":"Li","year":"2024","journal-title":"IEEE Access"},{"key":"ref_23","doi-asserted-by":"crossref","first-page":"115060","DOI":"10.1016\/j.measurement.2024.115060","article-title":"WSS-YOLO: An improved industrial defect detection network for steel surface defects","volume":"236","author":"Lu","year":"2024","journal-title":"Measurement"},{"key":"ref_24","doi-asserted-by":"crossref","unstructured":"Li, Z., Li, A., Li, W., Kong, X., and Zhang, Y. (July, January 30). HSD-YOLO: A lightweight and accurate method for PCB defect detection. Proceedings of the 2024 International Joint Conference on Neural Networks (IJCNN), Yokohama, Japan.","DOI":"10.1109\/IJCNN60899.2024.10650691"},{"key":"ref_25","doi-asserted-by":"crossref","first-page":"38","DOI":"10.1016\/j.patrec.2024.06.027","article-title":"EHIR: Energy-based hierarchical iterative image registration for accurate PCB defect detection","volume":"185","author":"Deng","year":"2024","journal-title":"Pattern Recognit. Lett."},{"key":"ref_26","doi-asserted-by":"crossref","unstructured":"Li, C., Yu, C., and Lin, H. (2023, January 26\u201328). DesNet: PCB defect detection network based on deformable convolution. Proceedings of the 2023 IEEE 3rd International Conference on Electronic Technology, Communication and Information (ICETCI), Changchun, China.","DOI":"10.1109\/ICETCI57876.2023.10176845"},{"key":"ref_27","doi-asserted-by":"crossref","first-page":"119","DOI":"10.5121\/ijdkp.2013.3408","article-title":"Methods for incremental learning: A survey","volume":"3","author":"Ade","year":"2013","journal-title":"Int. J. Data Min. Knowl. Manag. Process"},{"key":"ref_28","doi-asserted-by":"crossref","first-page":"5513","DOI":"10.1109\/TPAMI.2022.3213473","article-title":"Class-incremental learning: Survey and performance evaluation on image classification","volume":"45","author":"Masana","year":"2022","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."},{"key":"ref_29","unstructured":"Zhou, W., Chang, S., Sosa, N., Hamann, H., and Cox, D. (2020). Lifelong object detection. arXiv."},{"key":"ref_30","doi-asserted-by":"crossref","first-page":"3521","DOI":"10.1073\/pnas.1611835114","article-title":"Overcoming catastrophic forgetting in neural networks","volume":"114","author":"Kirkpatrick","year":"2017","journal-title":"Proc. Natl. Acad. Sci."},{"key":"ref_31","doi-asserted-by":"crossref","first-page":"046204","DOI":"10.1088\/1361-6501\/adbc0b","article-title":"An intelligent fault diagnosis of rotating machinery in class-incremental scenarios","volume":"36","author":"Han","year":"2025","journal-title":"Meas. Sci. Technol."},{"key":"ref_32","doi-asserted-by":"crossref","unstructured":"Mallya, A., and Lazebnik, S. (2018, January 18\u201322). Packnet: Adding multiple tasks to a single network by iterative pruning. Proceedings of the IEEE conference on Computer Vision and Pattern Recognition, Salt Lake City, UT, USA.","DOI":"10.1109\/CVPR.2018.00810"},{"key":"ref_33","doi-asserted-by":"crossref","first-page":"476","DOI":"10.1016\/j.neunet.2023.01.041","article-title":"Continual object detection: A review of definitions, strategies, and challenges","volume":"161","author":"Menezes","year":"2023","journal-title":"Neural Netw."},{"key":"ref_34","doi-asserted-by":"crossref","unstructured":"Chen, L., Yu, C., and Chen, L. (2019, January 14\u201319). A new knowledge distillation for incremental object detection. Proceedings of the 2019 International Joint Conference on Neural Networks (IJCNN), Budapest, Hungary.","DOI":"10.1109\/IJCNN.2019.8851980"},{"key":"ref_35","doi-asserted-by":"crossref","unstructured":"Shmelkov, K., Schmid, C., and Alahari, K. (2017, January 22\u201329). Incremental learning of object detectors without catastrophic forgetting. Proceedings of the IEEE International Conference on Computer Vision, Venice, Italy.","DOI":"10.1109\/ICCV.2017.368"},{"key":"ref_36","doi-asserted-by":"crossref","unstructured":"Hao, Y., Fu, Y., Jiang, Y.G., and Tian, Q. (2019, January 8\u201312). An end-to-end architecture for class-incremental object detection with knowledge distillation. Proceedings of the 2019 IEEE International Conference on Multimedia and Expo (ICME), Shanghai, China.","DOI":"10.1109\/ICME.2019.00009"},{"key":"ref_37","doi-asserted-by":"crossref","first-page":"109","DOI":"10.1016\/j.patrec.2020.09.030","article-title":"Faster ILOD: Incremental learning for object detectors based on faster RCNN","volume":"140","author":"Peng","year":"2020","journal-title":"Pattern Recognit. Lett."},{"key":"ref_38","first-page":"1","article-title":"Incremental detection of remote sensing objects with feature pyramid and knowledge distillation","volume":"60","author":"Chen","year":"2020","journal-title":"IEEE Trans. Geosci. Remote Sens."},{"key":"ref_39","doi-asserted-by":"crossref","unstructured":"Liu, Y., Schiele, B., Vedaldi, A., and Rupprecht, C. (2023, January 18\u201322). Continual detection transformer for incremental object detection. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Vancouver, Canada.","DOI":"10.1109\/CVPR52729.2023.02279"},{"key":"ref_40","doi-asserted-by":"crossref","unstructured":"Feng, T., Wang, M., and Yuan, H. (2022, January 19\u201324). Overcoming catastrophic forgetting in incremental object detection via elastic response distillation. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, New Orleans, LA, USA.","DOI":"10.1109\/CVPR52688.2022.00921"},{"key":"ref_41","doi-asserted-by":"crossref","first-page":"1493","DOI":"10.1109\/TIM.2019.2915404","article-title":"An end-to-end steel surface defect detection approach via fusing multiple hierarchical features","volume":"69","author":"He","year":"2019","journal-title":"IEEE Trans. Instrum. Meas."},{"key":"ref_42","unstructured":"Huang, W., and Wei, P. (2019). A PCB dataset for defects detection and classification. arXiv."},{"key":"ref_43","doi-asserted-by":"crossref","unstructured":"Sun, B., Li, B., Cai, S., Yuan, Y., and Zhang, C. (2021, January 19\u201325). FSCE: Few-shot object detection via contrastive proposal encoding. Proceedings of the IEEE\/CVF Conference on Computer Vision and Pattern Recognition, Virtual.","DOI":"10.1109\/CVPR46437.2021.00727"},{"key":"ref_44","first-page":"1","article-title":"An incremental knowledge learning framework for continuous defect detection","volume":"73","author":"Sun","year":"2023","journal-title":"IEEE Trans. Instrum. Meas."}],"container-title":["Sensors"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/5\/1413\/pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,2,26]],"date-time":"2026-02-26T05:33:58Z","timestamp":1772084038000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.mdpi.com\/1424-8220\/26\/5\/1413"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,2,24]]},"references-count":44,"journal-issue":{"issue":"5","published-online":{"date-parts":[[2026,3]]}},"alternative-id":["s26051413"],"URL":"https:\/\/doi.org\/10.3390\/s26051413","relation":{},"ISSN":["1424-8220"],"issn-type":[{"value":"1424-8220","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,2,24]]}}}