{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,30]],"date-time":"2026-04-30T22:26:49Z","timestamp":1777588009492,"version":"3.51.4"},"reference-count":41,"publisher":"Frontiers Media SA","license":[{"start":{"date-parts":[[2025,6,10]],"date-time":"2025-06-10T00:00:00Z","timestamp":1749513600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["frontiersin.org"],"crossmark-restriction":true},"short-container-title":["Front. Neurorobot."],"abstract":"<jats:p>Small object detection is a critical task in applications like autonomous driving and ship black smoke detection. While Deformable DETR has advanced small object detection, it faces limitations due to its reliance on CNNs for feature extraction, which restricts global context understanding and results in suboptimal feature representation. Additionally, it struggles with detecting small objects that occupy only a few pixels due to significant size disparities. To overcome these challenges, we propose the Context-Aware Enhanced Feature Refinement Deformable DETR, an improved Deformable DETR network. Our approach introduces Mask Attention in the backbone to improve feature extraction while effectively suppressing irrelevant background information. Furthermore, we propose a Context-Aware Enhanced Feature Refinement Encoder to address the issue of small objects with limited pixel representation. Experimental results demonstrate that our method outperforms the baseline, achieving a 2.1% improvement in mAP.<\/jats:p>","DOI":"10.3389\/fnbot.2025.1588565","type":"journal-article","created":{"date-parts":[[2025,6,10]],"date-time":"2025-06-10T05:23:19Z","timestamp":1749532999000},"update-policy":"https:\/\/doi.org\/10.3389\/crossmark-policy","source":"Crossref","is-referenced-by-count":7,"title":["Context-Aware Enhanced Feature Refinement for small object detection with Deformable DETR"],"prefix":"10.3389","volume":"19","author":[{"given":"Donghao","family":"Shi","sequence":"first","affiliation":[]},{"given":"Cunbin","family":"Zhao","sequence":"additional","affiliation":[]},{"given":"Jianwen","family":"Shao","sequence":"additional","affiliation":[]},{"given":"Minjie","family":"Feng","sequence":"additional","affiliation":[]},{"given":"Lei","family":"Luo","sequence":"additional","affiliation":[]},{"given":"Bing","family":"Ouyang","sequence":"additional","affiliation":[]},{"given":"Jiamin","family":"Huang","sequence":"additional","affiliation":[]}],"member":"1965","published-online":{"date-parts":[[2025,6,10]]},"reference":[{"key":"ref1","doi-asserted-by":"publisher","first-page":"2004.10934","DOI":"10.48550\/arXiv.2004.10934","article-title":"Yolov4: optimal speed and accuracy of object detection","author":"Bochkovskiy","year":"2020","journal-title":"Arxiv"},{"key":"ref2","doi-asserted-by":"crossref","DOI":"10.1109\/CVPRW56347.2022.00495","article-title":"Anomaly detection in autonomous driving: a survey","author":"Bogdoll","year":"2022"},{"key":"ref3","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2018.00644","article-title":"Cascade r-cnn: delving into high quality object detection","author":"Cai","year":"2018"},{"key":"ref4","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-030-58452-8_13","article-title":"End-to-end object detection with transformers","author":"Carion","year":"2020"},{"key":"ref5","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV48922.2021.00951","article-title":"Emerging properties in self-supervised vision transformers","author":"Caron","year":"2021"},{"key":"ref6","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV51070.2023.00610","article-title":"Group detr: fast detr training with group-wise one-to-many assignment","author":"Chen","year":"2023"},{"key":"ref7","doi-asserted-by":"publisher","first-page":"1906.07155","DOI":"10.48550\/arXiv.1906.07155","article-title":"Mmdetection: open mmlab detection toolbox and benchmark","author":"Chen","year":"2019","journal-title":"Arxiv"},{"key":"ref8","doi-asserted-by":"publisher","first-page":"1143","DOI":"10.1109\/tcsvt.2016.2527340","article-title":"Higher order linear dynamical systems for smoke detection in video surveillance applications","volume":"27","author":"Dimitropoulos","year":"2016","journal-title":"IEEE Trans. Circuits Syst. Video Technol."},{"key":"ref9","doi-asserted-by":"publisher","first-page":"2103.13597","DOI":"10.48550\/arXiv.2103.13597","article-title":"Mask attention networks: rethinking and strengthen transformer","author":"Fan","year":"2021","journal-title":"Arxiv"},{"key":"ref10","doi-asserted-by":"publisher","first-page":"2107.08430","DOI":"10.48550\/arXiv.2107.08430","article-title":"Yolox: exceeding yolo series in 2021","author":"Ge","year":"2021","journal-title":"Arxiv"},{"key":"ref11","first-page":"1440","author":"Girshick","year":"2015"},{"key":"ref12","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2014.81","article-title":"Rich feature hierarchies for accurate object detection and semantic segmentation","author":"Girshick","year":"2014"},{"key":"ref13","first-page":"1160","article-title":"Effective fusion factor in fpn for tiny object detection","author":"Gong","year":"2021"},{"key":"ref14","doi-asserted-by":"publisher","first-page":"244","DOI":"10.1007\/s11063-024-11699-6","article-title":"Feature enhancement based oriented object detection in remote sensing images","volume":"56","author":"Guo","year":"2024","journal-title":"Neural. Process. Lett."},{"key":"ref15","doi-asserted-by":"crossref","DOI":"10.23919\/MVA57639.2023.10215748","article-title":"Ensemble fusion for small object detection","author":"Hou","year":"2023"},{"key":"ref16","article-title":"Ultralytics YOLO","author":"Jocher","year":"2023"},{"key":"ref17","doi-asserted-by":"crossref","DOI":"10.23919\/MVA57639.2023.10215935","article-title":"Mva2023 small object detection challenge for spotting birds: dataset, methods, and results","author":"Kondo","year":"2023"},{"key":"ref18","doi-asserted-by":"publisher","first-page":"6064","DOI":"10.1109\/tce.2024.3412168","article-title":"Small object detection based on lightweight feature pyramid","volume":"70","author":"Li","year":"2024","journal-title":"IEEE Trans. Consum. Electron."},{"key":"ref19","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR52688.2022.01325","article-title":"Dn-detr: accelerate detr training by introducing query denoising","author":"Li","year":"2022"},{"key":"ref20","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2017.106","article-title":"Feature pyramid networks for object detection","author":"Lin","year":"2017"},{"key":"ref21","doi-asserted-by":"crossref","DOI":"10.1007\/978-3-319-46448-0_2","article-title":"Ssd: single shot multibox detector","author":"Liu","year":"2016"},{"key":"ref22","doi-asserted-by":"publisher","first-page":"2201.12329","DOI":"10.48550\/arXiv.2201.12329","article-title":"Dab-detr: dynamic anchor boxes are better queries for detr","author":"Liu","year":"2022","journal-title":"Arxiv"},{"key":"ref23","doi-asserted-by":"publisher","first-page":"2407.17140","DOI":"10.48550\/arXiv.2407.17140","article-title":"Rt-detrv2: improved baseline with bag-of-freebies for real-time detection transformer","author":"Lv","year":"2024","journal-title":"Arxiv"},{"key":"ref24","doi-asserted-by":"publisher","first-page":"2304.07193","DOI":"10.48550\/arXiv.2304.07193","article-title":"Dinov2: learning robust visual features without supervision","author":"Oquab","year":"2023","journal-title":"Arxiv"},{"key":"ref25","doi-asserted-by":"publisher","first-page":"1137","DOI":"10.1109\/TPAMI.2016.2577031","article-title":"Faster r-cnn: towards real-time object detection with region proposal networks","volume":"28","author":"Ren","year":"2015","journal-title":"Adv. Neural Inform. Proces. Syst."},{"key":"ref26","doi-asserted-by":"publisher","first-page":"773","DOI":"10.1007\/s12239-023-0064-z","article-title":"Cnn-based object detection and distance prediction for autonomous driving using stereo images","volume":"24","author":"Song","year":"2023","journal-title":"Int. J. Automot. Technol."},{"key":"ref27","doi-asserted-by":"publisher","first-page":"1380","DOI":"10.1049\/itr2.12212","article-title":"Real-time traffic cone detection for autonomous driving based on yolov4","volume":"16","author":"Su","year":"2022","journal-title":"IET Intell. Transp. Syst."},{"key":"ref28","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR42600.2020.01079","article-title":"Efficientdet: scalable and efficient object detection","author":"Tan","year":"2020"},{"key":"ref29","first-page":"6614","article-title":"Hic-yolov5: improved yolov5 for small object detection","author":"Tang","year":"2024"},{"key":"ref30","doi-asserted-by":"publisher","first-page":"104471","DOI":"10.1016\/j.imavis.2022.104471","article-title":"Deep learning-based detection from the perspective of small or tiny objects: A survey","volume":"123","author":"Tong","year":"2022","journal-title":"Image Vis. Comput."},{"key":"ref31","first-page":"5998","article-title":"Attention is all you need","volume":"30","author":"Vaswani","year":"2017","journal-title":"Adv. Neural Inform. Process. Syst."},{"key":"ref32","doi-asserted-by":"publisher","first-page":"2405.14458","DOI":"10.48550\/arXiv.2405.14458","article-title":"Yolov10: real-time end-to-end object detection","author":"Wang","year":"2024","journal-title":"Arxiv"},{"key":"ref33","doi-asserted-by":"publisher","first-page":"1945","DOI":"10.3390\/jmse11101945","article-title":"Research on black smoke detection and class evaluation method for ships based on yolov5s-cmbi multi-feature fusion","volume":"11","author":"Wang","year":"2023","journal-title":"J. Mar. Sci. Eng."},{"key":"ref34","doi-asserted-by":"publisher","first-page":"2409.08475","DOI":"10.48550\/arXiv.2409.08475","article-title":"Rt-detrv3: real-time end-to-end object detection with hierarchical dense positive supervision","author":"Wang","year":"2024","journal-title":"Arxiv"},{"key":"ref35","doi-asserted-by":"publisher","first-page":"6283","DOI":"10.1007\/s00521-024-09422-6","article-title":"A review of small object detection based on deep learning","volume":"36","author":"Wei","year":"2024","journal-title":"Neural Comput. Applic."},{"key":"ref36","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR52733.2024.01605","article-title":"Detrs beat yolos on real-time object detection","author":"Zhao","year":"2024"},{"key":"ref37","doi-asserted-by":"publisher","first-page":"1342126","DOI":"10.3389\/fnbot.2024.1342126","article-title":"Improved object detection method for unmanned driving based on transformers","volume":"18","author":"Zhao","year":"2024","journal-title":"Front. Neurorobot."},{"key":"ref38","doi-asserted-by":"publisher","first-page":"1904.07850","DOI":"10.48550\/arXiv.1904.07850","article-title":"Objects as points","author":"Zhou","year":"2019","journal-title":"Arxiv"},{"key":"ref39","doi-asserted-by":"crossref","DOI":"10.1109\/ICCVW54120.2021.00312","article-title":"Tph-yolov5: improved yolov5 based on transformer prediction head for object detection on drone-captured scenarios","author":"Zhu","year":"2021"},{"key":"ref40","doi-asserted-by":"publisher","first-page":"2010.04159","DOI":"10.48550\/arXiv.2010.04159","article-title":"Deformable detr: deformable transformers for end-to-end object detection","author":"Zhu","year":"2021","journal-title":"Arxiv"},{"key":"ref41","doi-asserted-by":"publisher","first-page":"7380","DOI":"10.1109\/tpami.2021.3119563","article-title":"Detection and tracking meet drones challenge","volume":"44","author":"Zhu","year":"2021","journal-title":"IEEE Trans. Pattern Anal. Mach. Intell."}],"container-title":["Frontiers in Neurorobotics"],"original-title":[],"link":[{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnbot.2025.1588565\/full","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,6,10]],"date-time":"2025-06-10T05:23:22Z","timestamp":1749533002000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.frontiersin.org\/articles\/10.3389\/fnbot.2025.1588565\/full"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,6,10]]},"references-count":41,"alternative-id":["10.3389\/fnbot.2025.1588565"],"URL":"https:\/\/doi.org\/10.3389\/fnbot.2025.1588565","relation":{},"ISSN":["1662-5218"],"issn-type":[{"value":"1662-5218","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,6,10]]},"article-number":"1588565"}}