{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,30]],"date-time":"2026-07-30T00:19:07Z","timestamp":1785370747511,"version":"3.55.0"},"reference-count":42,"publisher":"Public Library of Science (PLoS)","issue":"12","license":[{"start":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T00:00:00Z","timestamp":1764892800000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"funder":[{"name":"Liaoning Science Foundation of China","award":["2023-MS-322"],"award-info":[{"award-number":["2023-MS-322"]}]}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>Standard detectors such as YOLOv8 face significant challenges when applied to aerial drone imagery, including extreme scale variations, minute targets, and complex backgrounds. Their generic feature fusion architecture is prone to generating false positives and missing small objects. To address these limitations, we propose an improved MFDA-YOLO model based on YOLOv8. The model introduces an Attention-based Intra-scale Feature Interaction (AIFI) module in the backbone network to enhance high-level feature interactions, improve the adaptation to multi-scale targets, and strengthen feature representation. In the neck network, we design the Drone Image Detection Pyramid (DIDP) network, which integrates a space-to-depth convolution module to efficiently propagate multi-scale features from shallow to deep layers. By introducing an omni-kernel module in the cross-stage partial network for image recovery, DIDP can enhance global contextual awareness and eliminate the computational burden to extend the traditional P2 detection layer. Aiming at the problem of insufficient synergy between localization and classification tasks in the detection head, we design the Dynamic Alignment Detection Head (DADH). DADH can achieve cross-task representation optimization through multi-scale feature interaction learning and a dynamic feature selection mechanism, which significantly reduces model complexity and maintains detection accuracy. In addition, we employ the WIoUv3 loss function to dynamically adjust the focusing coefficients and enhance the model\u2019s ability to distinguish small targets. Extensive experimental results demonstrate that MFDA-YOLO outperforms existing state-of-the-art methods such as YOLOv11 and YOLOv13 across the VisDrone2019, HIT-UAV, and NWPU VHR-10 datasets. Particularly on the VisDrone2019 dataset, MFDA-YOLO surpasses the baseline YOLOv8n model, achieving a 4.4 percentage point improvement in mAP0.5 and a 2.7 percentage point increase in mAP0.5:0.95. Furthermore, it reduces parameters by 17.2%, effectively lowering both false negative and false positive rates.<\/jats:p>","DOI":"10.1371\/journal.pone.0337810","type":"journal-article","created":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T18:55:18Z","timestamp":1764960918000},"page":"e0337810","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":2,"title":["MFDA-YOLO: A multiscale feature fusion and dynamic alignment network for UAV small objects detection"],"prefix":"10.1371","volume":"20","author":[{"ORCID":"https:\/\/orcid.org\/0009-0004-2321-3583","authenticated-orcid":true,"given":"Dan","family":"Tian","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xiao","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Dongxin","family":"Liu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Ying","family":"Hao","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2025,12,5]]},"reference":[{"issue":"21","key":"pone.0337810.ref001","doi-asserted-by":"crossref","first-page":"3630","DOI":"10.3390\/rs12213630","article-title":"EFN: field-based object detection for aerial images","volume":"12","author":"J Liu","year":"2020","journal-title":"Remote Sensing"},{"issue":"10","key":"pone.0337810.ref002","doi-asserted-by":"crossref","first-page":"1865","DOI":"10.1109\/JPROC.2017.2675998","article-title":"Remote sensing image scene classification: benchmark and state of the art","volume":"105","author":"G Cheng","year":"2017","journal-title":"Proc IEEE"},{"key":"pone.0337810.ref003","doi-asserted-by":"crossref","first-page":"29294","DOI":"10.1109\/ACCESS.2024.3368848","article-title":"LAYN: lightweight multi-scale attention YOLOv8 network for small object detection","volume":"12","author":"S Ma","year":"2024","journal-title":"IEEE Access"},{"key":"pone.0337810.ref004","doi-asserted-by":"crossref","first-page":"111209","DOI":"10.1016\/j.patcog.2024.111209","article-title":"BiFPN-YOLO: one-stage object detection integrating bi-directional feature pyramid networks","volume":"160","author":"J Doherty","year":"2025","journal-title":"Pattern Recog"},{"key":"pone.0337810.ref005","first-page":"779","author":"J Redmon","year":"2016"},{"issue":"3","key":"pone.0337810.ref006","doi-asserted-by":"crossref","first-page":"642","DOI":"10.1007\/s11263-019-01204-1","article-title":"CornerNet: detecting objects as paired keypoints","volume":"128","author":"H Law","year":"2019","journal-title":"Int J Comput Vis"},{"key":"pone.0337810.ref007","doi-asserted-by":"crossref","DOI":"10.1109\/ICCV.2019.00972","article-title":"FCOS: fully convolutional one-stage object detection.","author":"Z Tian","year":"2019"},{"key":"pone.0337810.ref008","first-page":"10778","article-title":"EfficientDet: scalable and efficient object detection.","author":"M Tan","year":"2020"},{"issue":"2","key":"pone.0337810.ref009","first-page":"2547","article-title":"YOLO-MFD: remote sensing image object detection with multi-scale fusion dynamic head","volume":"79","author":"Z Zhang","year":"2024","journal-title":"Compt Mater Contin"},{"issue":"5","key":"pone.0337810.ref010","doi-asserted-by":"crossref","first-page":"1483","DOI":"10.1109\/TPAMI.2019.2956516","article-title":"Cascade R-CNN: high quality object detection and instance segmentation","volume":"43","author":"Z Cai","year":"2021","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"key":"pone.0337810.ref011","first-page":"936","author":"T-Y Lin","year":"2017"},{"key":"pone.0337810.ref012","first-page":"9992","author":"Z Liu","year":"2021"},{"issue":"6","key":"pone.0337810.ref013","article-title":"ASG-YOLOv5: Improved YOLOv5 unmanned aerial vehicle remote sensing aerial images scenario for small object detection based on attention and spatial gating","volume":"19","author":"H Shi","year":"2024","journal-title":"PLoS One"},{"key":"pone.0337810.ref014"},{"issue":"3","key":"pone.0337810.ref015","doi-asserted-by":"crossref","DOI":"10.1371\/journal.pone.0299058","article-title":"Development of a deep learning-based surveillance system for forest fire detection and monitoring using UAV","volume":"19","author":"I Shamta","year":"2024","journal-title":"PLoS One"},{"issue":"11","key":"pone.0337810.ref016","doi-asserted-by":"crossref","first-page":"609","DOI":"10.3390\/drones8110609","article-title":"YOLO-DroneMS: multi-scale object detection network for unmanned aerial vehicle (UAV) images","volume":"8","author":"X Zhao","year":"2024","journal-title":"Drones"},{"issue":"9","key":"pone.0337810.ref017","doi-asserted-by":"crossref","first-page":"453","DOI":"10.3390\/drones8090453","article-title":"HSP-YOLOv8: UAV aerial photography small target detection algorithm","volume":"8","author":"H Zhang","year":"2024","journal-title":"Drones"},{"key":"pone.0337810.ref018","first-page":"16965","article-title":"DETRs beat YOLOs on real-time object detection.","author":"Y Zhao","year":"2024"},{"issue":"6","key":"pone.0337810.ref019","doi-asserted-by":"crossref","first-page":"240","DOI":"10.3390\/drones8060240","article-title":"PHSI-RTDETR: a lightweight infrared small target detection algorithm based on UAV aerial photography","volume":"8","author":"S Wang","year":"2024","journal-title":"Drones"},{"key":"pone.0337810.ref020","author":"R Sunkara","year":"2022","journal-title":"No more strided convolutions or pooling: a new CNN building block for low-resolution images and small objects. In: Joint European Conference on Machine Learning and Knowledge Discovery in Databases"},{"key":"pone.0337810.ref021","author":"C-Y Wang","year":"2020"},{"issue":"2","key":"pone.0337810.ref022","doi-asserted-by":"crossref","first-page":"1426","DOI":"10.1609\/aaai.v38i2.27907","article-title":"Omni-kernel network for image restoration","volume":"38","author":"Y Cui","year":"2024","journal-title":"AAAI"},{"key":"pone.0337810.ref023","author":"C Feng","year":"2021"},{"key":"pone.0337810.ref024","first-page":"9300","author":"X Zhu","year":"2019"},{"key":"pone.0337810.ref025","doi-asserted-by":"crossref","first-page":"3","DOI":"10.1007\/978-3-030-01261-8_1","article-title":"Group normalization","author":"Y Wu","year":"2018","journal-title":"Lecture Notes in Computer Science"},{"key":"pone.0337810.ref026","author":"Z Tong","year":"2023"},{"issue":"3","key":"pone.0337810.ref027","article-title":"Small object detection algorithm incorporating swin transformer for tea buds","volume":"19","author":"M Shi","year":"2024","journal-title":"PLoS One"},{"issue":"3","key":"pone.0337810.ref028","doi-asserted-by":"crossref","first-page":"46","DOI":"10.54097\/fcis.v5i3.13852","article-title":"Object detection of UAV aerial image based on YOLOv8","volume":"5","author":"C Liu","year":"2023","journal-title":"Front Compt Intell Syst"},{"key":"pone.0337810.ref029","author":"D Du","year":"2019"},{"issue":"1","key":"pone.0337810.ref030","doi-asserted-by":"crossref","first-page":"227","DOI":"10.1038\/s41597-023-02066-6","article-title":"HIT-UAV: a high-altitude infrared thermal dataset for unmanned aerial vehicle-based object detection","volume":"10","author":"J Suo","year":"2023","journal-title":"Sci Data"},{"key":"pone.0337810.ref031","doi-asserted-by":"crossref","first-page":"119","DOI":"10.1016\/j.isprsjprs.2014.10.002","article-title":"Multi-class geospatial object detection and geographic image classification based on collection of part detectors","volume":"98","author":"G Cheng","year":"2014","journal-title":"ISPRS J Photogram Remote Sens"},{"key":"pone.0337810.ref032","author":"R Yacouby","year":"2020"},{"key":"pone.0337810.ref033","author":"CY Wang","year":"2024"},{"key":"pone.0337810.ref034","author":"A Wang","year":"2024"},{"key":"pone.0337810.ref035","author":"Z Ge","year":"2021"},{"key":"pone.0337810.ref036","author":"R Khanam","year":"2024"},{"key":"pone.0337810.ref037","author":"Y Tian","year":"2025"},{"key":"pone.0337810.ref038","author":"M Lei","year":"2025"},{"key":"pone.0337810.ref039","first-page":"2999","author":"T-Y Lin","year":"2017"},{"key":"pone.0337810.ref040","author":"C Lyu","year":"2022"},{"key":"pone.0337810.ref041","author":"X Zhu","year":"2020"},{"key":"pone.0337810.ref042","author":"S Zhang","year":"2020"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0337810","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,12,5]],"date-time":"2025-12-05T18:55:27Z","timestamp":1764960927000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0337810"}},"subtitle":[],"editor":[{"given":"Armin","family":"Moghimi","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2025,12,5]]},"references-count":42,"journal-issue":{"issue":"12","published-online":{"date-parts":[[2025,12,5]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0337810","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025,12,5]]}}}