{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,7]],"date-time":"2026-08-07T14:31:17Z","timestamp":1786113077742,"version":"build-2736575974"},"reference-count":53,"publisher":"Public Library of Science (PLoS)","issue":"8","license":[{"start":{"date-parts":[[2026,8,5]],"date-time":"2026-08-05T00:00:00Z","timestamp":1785888000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":["www.plosone.org"],"crossmark-restriction":false},"short-container-title":["PLoS One"],"abstract":"<jats:p>\n                    Object detection in complex traffic environments remains challenging due to illumination-induced feature degradation and weak responses from distant small objects. To address these limitations, we propose a two-stage multimodal detection framework that enhances perception under degraded conditions. The first stage employs a Dual-Path Attention Fusion Module (DPAFM) to adaptively integrate RGB and thermal features via learnable gated attention, mitigating modality-specific degradation. The second stage introduces a Hierarchical Attention Graph Neural Network (HA-GNN), which models detection candidates as graph nodes and performs hierarchical relational reasoning over their appearance and spatial relationships to compensate for weakened local cues. Extensive experiments on the KAIST and R-LiViT datasets demonstrate that our framework significantly improves robustness across day and night scenarios, reducing the day-night performance gap from 72.4% to 7.4% and substantially enhancing the detection of small objects. Moreover, the method achieves a real-time inference speed of 27.4 FPS, offering a favorable trade-off between accuracy and efficiency for autonomous driving perception in challenging environments.The code is available at:\n                    <jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" ext-link-type=\"uri\" xlink:href=\"https:\/\/github.com\/yangjiepry\/YOLO-HA-GNN.git\" xlink:type=\"simple\">https:\/\/github.com\/yangjiepry\/YOLO-HA-GNN.git<\/jats:ext-link>\n                    .\n                  <\/jats:p>","DOI":"10.1371\/journal.pone.0354466","type":"journal-article","created":{"date-parts":[[2026,8,5]],"date-time":"2026-08-05T17:43:03Z","timestamp":1785951783000},"page":"e0354466","update-policy":"https:\/\/doi.org\/10.1371\/journal.pone.corrections_policy","source":"Crossref","is-referenced-by-count":0,"title":["Learning to recover weak signals: A two-stage graph-based framework for all-weather RGB-T object detection"],"prefix":"10.1371","volume":"21","author":[{"ORCID":"https:\/\/orcid.org\/0009-0006-3337-9063","authenticated-orcid":true,"given":"Ruoyan","family":"Pei","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Pengge","family":"Ma","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jinwang","family":"Qian","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hai","family":"Yang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Junling","family":"Sun","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"340","published-online":{"date-parts":[[2026,8,5]]},"reference":[{"key":"pone.0354466.ref001","doi-asserted-by":"crossref","DOI":"10.1109\/CVPR.2012.6248074","volume-title":"Are we ready for autonomous driving? The KITTI vision benchmark suite","author":"A Geiger","year":"2012"},{"key":"pone.0354466.ref002","author":"J Redmon","year":"2018"},{"key":"pone.0354466.ref003","doi-asserted-by":"crossref","first-page":"106977","DOI":"10.1016\/j.patcog.2019.106977","article-title":"RGB-T object tracking: Benchmark and baseline","volume":"96","author":"C Li","year":"2019","journal-title":"Pattern Recognition"},{"key":"pone.0354466.ref004","doi-asserted-by":"crossref","DOI":"10.1109\/ICSC.2020.00085","volume-title":"Intelligent Active and Passive Learning for Integrated Semantic Computing for Vision Data Annotation","author":"IEW Rachmawan","year":"2020"},{"issue":"000","key":"pone.0354466.ref005","first-page":"1","article-title":"Multispectral Semantic Segmentation for UAVs: A Benchmark Dataset and Baseline. Geoscience and Remote Sensing","volume":"62","author":"Q Li","year":"2024","journal-title":"IEEE Transactions on"},{"key":"pone.0354466.ref006","doi-asserted-by":"crossref","unstructured":"Mirlach J, Wan L, Wiedholz A, Keen HE, Eich A, editors. R-LiViT: A LiDAR-Visual-Thermal Dataset Enabling Vulnerable Road User Focused Roadside Perception2025.","DOI":"10.1109\/ICCV51701.2025.02635"},{"key":"pone.0354466.ref007","unstructured":"Narayanan V, Sai B, Ahuja R, Likhar P, Kumar VR, Yogamani S. MambaFusion: Adaptive State-Space Fusion for Multimodal 3D Object Detection. 2026."},{"key":"pone.0354466.ref008","unstructured":"Huang Y, Lin J, Zhou C, Yang H, Huang L. Modality Competition: What Makes Joint Training of Multi-modal Network Fail in Deep Learning? (Provably). 2022."},{"issue":"23","key":"pone.0354466.ref009","doi-asserted-by":"crossref","first-page":"7855","DOI":"10.3390\/s24237855","article-title":"Tackling Heterogeneous Light Detection and Ranging-Camera Alignment Challenges in Dynamic Environments: A Review for Object Detection","volume":"24","author":"Y Wang","year":"2024","journal-title":"Sensors (Basel)"},{"issue":"1","key":"pone.0354466.ref010","article-title":"Modality balancing network for pedestrian detection based on cross-modal compensation fusion and multimodal feature alignment","volume":"42","author":"Z Li","year":"2025","journal-title":"Vis Comput"},{"key":"pone.0354466.ref011","first-page":"2735","volume-title":"ResNeSt: Split-Attention Networks. 2022 IEEE\/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW","author":"H Zhang","year":"2022"},{"key":"pone.0354466.ref012","doi-asserted-by":"crossref","first-page":"129718","DOI":"10.1016\/j.neucom.2025.129718","article-title":"MambaSOD: Dual Mamba-driven cross-modal fusion network for RGB-D Salient Object Detection","volume":"631","author":"Y Zhan","year":"2025","journal-title":"Neurocomputing"},{"key":"pone.0354466.ref013","doi-asserted-by":"crossref","unstructured":"Shen J, Zhan H, Dong S, Zuo X, Yang W, Ling H. Multispectral State-Space Feature Fusion: Bridging Shared and Cross-Parametric Interactions for Object Detection. 2025.","DOI":"10.1016\/j.inffus.2025.103895"},{"key":"pone.0354466.ref014","doi-asserted-by":"crossref","first-page":"103260","DOI":"10.1016\/j.jvcir.2021.103260","article-title":"An Interconnected Feature Pyramid Networks for object detection","volume":"79","author":"Q Wang","year":"2021","journal-title":"Journal of Visual Communication and Image Representation"},{"issue":"6","key":"pone.0354466.ref015","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":"S Ren","year":"2017","journal-title":"IEEE Trans Pattern Anal Mach Intell"},{"issue":"2","key":"pone.0354466.ref016","doi-asserted-by":"crossref","first-page":"2751","DOI":"10.32604\/cmc.2025.060860","article-title":"CG-FCLNet: Category-Guided Feature Collaborative Learning Network for Semantic Segmentation of Remote Sensing Images","volume":"83","author":"M Yao","year":"2025","journal-title":"CMC"},{"issue":"1","key":"pone.0354466.ref017","doi-asserted-by":"crossref","first-page":"13533","DOI":"10.1038\/s41598-025-94655-1","article-title":"Drivable area recognition on unstructured roads for autonomous vehicles using an optimized bilateral neural network","volume":"15","author":"X Chen","year":"2025","journal-title":"Sci Rep"},{"key":"pone.0354466.ref018","unstructured":"Reports on Networks Findings from Shandong University Provide New Insights (Wavenet: Wavelet Network With Knowledge Distillation for Rgb-t Salient Object Detection). Network Daily News. 2023."},{"issue":"1","key":"pone.0354466.ref019","doi-asserted-by":"crossref","first-page":"20","DOI":"10.1016\/0166-2236(92)90344-8","article-title":"Separate visual pathways for perception and action","volume":"15","author":"MA Goodale","year":"1992","journal-title":"Trends Neurosci"},{"issue":"4","key":"pone.0354466.ref020","doi-asserted-by":"crossref","first-page":"428","DOI":"10.1016\/S0959-4388(03)00105-3","article-title":"Visual attention: the where, what, how and why of saliency","volume":"13","author":"S Treue","year":"2003","journal-title":"Curr Opin Neurobiol"},{"key":"pone.0354466.ref021","first-page":"71","article-title":"Laws of Organization in Perceptual Forms","volume":"4","author":"M Wertheimer","year":"1938","journal-title":"Psychologische Forschung"},{"issue":"000","key":"pone.0354466.ref022","volume":"111","author":"C V\u00e1zquez","year":"2020","journal-title":"High spatial resolution optical fiber thermometers for applications in harsh environments"},{"key":"pone.0354466.ref023","doi-asserted-by":"crossref","unstructured":"Joshi R, Mishra S. Learning Graph Representations. 2021.","DOI":"10.1007\/978-981-16-3398-0_10"},{"key":"pone.0354466.ref024","doi-asserted-by":"crossref","first-page":"378","DOI":"10.1007\/978-981-99-8555-5_30","article-title":"Object Centric Body Part Attention Network forHuman-Object Interaction Detection","author":"Z Liu","year":"2024","journal-title":"Lecture Notes in Computer Science"},{"key":"pone.0354466.ref025","volume-title":"Multi-Modal Graph Neural Network for Joint Reasoning on Vision and Scene Text","author":"D Gao","year":"2020"},{"issue":"1","key":"pone.0354466.ref026","doi-asserted-by":"crossref","first-page":"37794","DOI":"10.1038\/s41598-025-21550-0","article-title":"A lightweight infrared target detection network suitable for land and water surfaces","volume":"15","author":"F Yang","year":"2025","journal-title":"Sci Rep"},{"key":"pone.0354466.ref027","doi-asserted-by":"crossref","unstructured":"Li J, Xie C, Ao J, Leng D, Yin Y, editors. LMM-Det: Make Large Multimodal Models Excel in Object. Detection. 2025.","DOI":"10.1109\/ICCV51701.2025.00036"},{"key":"pone.0354466.ref028","doi-asserted-by":"crossref","first-page":"105203","DOI":"10.1016\/j.trc.2025.105203","article-title":"Context-aware multi-task learning for pedestrian intent and trajectory prediction","volume":"178","author":"F Munir","year":"2025","journal-title":"Transportation Research Part C: Emerging Technologies"},{"key":"pone.0354466.ref029","doi-asserted-by":"crossref","unstructured":"Zhou M, Zhao X, Luo F, Luo J, Pu H, Xiang T. Robust RGB-T Tracking via Adaptive Modality Weight Correlation Filters and Cross-modality Learning. ACM Transactions on Multimedia Computing, Communications, and Applications (TOMCCAP). 2024;20(4):20.","DOI":"10.1145\/3630100"},{"key":"pone.0354466.ref030","doi-asserted-by":"crossref","unstructured":"Reinders C, Berdan R, Besbinar B, Otsuka J, Iso D. RAW-Diffusion: RGB-Guided Diffusion Models for High-Fidelity RAW Image Generation. 2024.","DOI":"10.1109\/WACV61041.2025.00817"},{"key":"pone.0354466.ref031","unstructured":"Wang G, Li J, Sun Y, Chen X, Liu C, Wu Y, et al. Hierarchical Reasoning Model. 2025."},{"key":"pone.0354466.ref032","doi-asserted-by":"crossref","first-page":"131570","DOI":"10.1016\/j.eswa.2026.131570","article-title":"A novel implicit cross-attention framework for RGB-T object detection","volume":"314","author":"Z Liu","year":"2026","journal-title":"Expert Systems with Applications"},{"key":"pone.0354466.ref033","doi-asserted-by":"crossref","first-page":"104385","DOI":"10.1016\/j.inffus.2026.104385","article-title":"Im2State: Image-to-state Mamba for hyperspectral and multispectral image fusion","volume":"134","author":"C Li","year":"2026","journal-title":"Information Fusion"},{"issue":"1","key":"pone.0354466.ref034","doi-asserted-by":"crossref","DOI":"10.1049\/ipr2.70111","article-title":"AW\u2010YOLO: A Multi\u2010Object Detection Network for Autonomous Driving Under all Weather Conditions","volume":"19","author":"D Chen","year":"2025","journal-title":"IET Image Processing"},{"issue":"12","key":"pone.0354466.ref035","doi-asserted-by":"crossref","first-page":"8597","DOI":"10.1007\/s11760-024-03493-7","article-title":"Rethinking all-in-one adverse weather removal for object detection","volume":"18","author":"Y Li","year":"2024","journal-title":"SIViP"},{"issue":"1","key":"pone.0354466.ref036","doi-asserted-by":"crossref","DOI":"10.1038\/s43586-026-00470-x","article-title":"Hyperspectral imaging","volume":"6","author":"D Hong","year":"2026","journal-title":"Nat Rev Methods Primers"},{"issue":"8","key":"pone.0354466.ref037","doi-asserted-by":"crossref","first-page":"5227","DOI":"10.1109\/TPAMI.2024.3362475","article-title":"SpectralGPT: Spectral Remote Sensing Foundation Model","volume":"46","author":"D Hong","year":"2024","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"key":"pone.0354466.ref038","doi-asserted-by":"crossref","first-page":"105684","DOI":"10.1016\/j.dsp.2025.105684","article-title":"Lightweight multi-scale and multi-level feature fusion network for RGB-D-T salient object detection","volume":"168","author":"Z Bai","year":"2026","journal-title":"Digital Signal Processing"},{"key":"pone.0354466.ref039","doi-asserted-by":"crossref","first-page":"132307","DOI":"10.1016\/j.eswa.2026.132307","article-title":"MPCANet: Multi-physical prior guided cross-modal attention and fusion network for RGB-T salient object detection","volume":"322","author":"H Gao","year":"2026","journal-title":"Expert Systems with Applications"},{"issue":"3","key":"pone.0354466.ref040","doi-asserted-by":"crossref","first-page":"2697","DOI":"10.1109\/TCSVT.2025.3613770","article-title":"HyPSAM: Hybrid Prompt-Driven Segment Anything Model for RGB-Thermal Salient Object Detection","volume":"36","author":"R Hou","year":"2026","journal-title":"IEEE Trans Circuits Syst Video Technol"},{"issue":"3","key":"pone.0354466.ref041","doi-asserted-by":"crossref","first-page":"1958","DOI":"10.1109\/TPAMI.2024.3511621","article-title":"Divide-and-Conquer: Confluent Triple-Flow Network for RGB-T Salient Object Detection","volume":"47","author":"H Tang","year":"2025","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"issue":"3","key":"pone.0354466.ref042","doi-asserted-by":"crossref","first-page":"2513","DOI":"10.1109\/TCSVT.2024.3489440","article-title":"Intra-Modality Self-Enhancement Mirror Network for RGB-T Salient Object Detection","volume":"35","author":"J Wang","year":"2025","journal-title":"IEEE Trans Circuits Syst Video Technol"},{"key":"pone.0354466.ref043","doi-asserted-by":"crossref","first-page":"112103","DOI":"10.1016\/j.patcog.2025.112103","article-title":"LESOD: Lightweight and efficient network for RGB-D salient object detection","volume":"171","author":"M Zhong","year":"2026","journal-title":"Pattern Recognition"},{"issue":"4","key":"pone.0354466.ref044","doi-asserted-by":"crossref","first-page":"3104","DOI":"10.1109\/TCSVT.2024.3502244","article-title":"Highly Efficient RGB-D Salient Object Detection With Adaptive Fusion and Attention Regulation","volume":"35","author":"H Gao","year":"2025","journal-title":"IEEE Transactions on Circuits and Systems for Video Technology"},{"issue":"8","key":"pone.0354466.ref045","doi-asserted-by":"crossref","first-page":"3067","DOI":"10.1007\/s11263-024-02020-y","article-title":"Cross-Modal Fusion and Progressive Decoding Network for RGB-D Salient Object Detection","volume":"132","author":"X Hu","year":"2024","journal-title":"Int J Comput Vis"},{"key":"pone.0354466.ref046","doi-asserted-by":"crossref","first-page":"112126","DOI":"10.1016\/j.knosys.2024.112126","article-title":"MAGNet: Multi-scale Awareness and Global fusion Network for RGB-D salient object detection","volume":"299","author":"M Zhong","year":"2024","journal-title":"Knowledge-Based Systems"},{"key":"pone.0354466.ref047","doi-asserted-by":"crossref","first-page":"125278","DOI":"10.1016\/j.eswa.2024.125278","article-title":"Perceptual localization and focus refinement network for RGB-D salient object detection","volume":"259","author":"J Han","year":"2025","journal-title":"Expert Systems with Applications"},{"key":"pone.0354466.ref048","author":"Y Zhao","year":"2023"},{"issue":"12","key":"pone.0354466.ref049","doi-asserted-by":"crossref","first-page":"2121","DOI":"10.1109\/JAS.2022.106082","article-title":"SuperFusion: A Versatile Image Registration and Fusion Network with Semantic Awareness","volume":"9","author":"L Tang","year":"2022","journal-title":"IEEE\/CAA J Autom Sinica"},{"key":"pone.0354466.ref050","doi-asserted-by":"crossref","first-page":"131714","DOI":"10.1016\/j.eswa.2026.131714","article-title":"Progressive multimodal synergetic fusion network for salient object detection in urban perception","volume":"315","author":"J Yang","year":"2026","journal-title":"Expert Systems with Applications"},{"key":"pone.0354466.ref051","first-page":"3119","author":"Y Zhang","year":"2022"},{"issue":"6","key":"pone.0354466.ref052","doi-asserted-by":"crossref","first-page":"3180","DOI":"10.1007\/s11263-024-02315-0","article-title":"CSFRNet: integrating clothing status awareness for long-term person re-identification","volume":"133","author":"Y Huang","year":"2024","journal-title":"Int J Comput Vis"},{"issue":"2","key":"pone.0354466.ref053","doi-asserted-by":"crossref","first-page":"59","DOI":"10.3390\/fire8020059","article-title":"TFNet: Transformer-Based Multi-Scale Feature Fusion Forest Fire Image Detection Network","volume":"8","author":"H Liu","year":"2025","journal-title":"Fire"}],"container-title":["PLOS One"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0354466","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,8,5]],"date-time":"2026-08-05T17:43:11Z","timestamp":1785951791000},"score":1,"resource":{"primary":{"URL":"https:\/\/dx.plos.org\/10.1371\/journal.pone.0354466"}},"subtitle":[],"editor":[{"given":"Baicang","family":"Guo","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"editor"}]}],"short-title":[],"issued":{"date-parts":[[2026,8,5]]},"references-count":53,"journal-issue":{"issue":"8","published-online":{"date-parts":[[2026,8,5]]}},"URL":"https:\/\/doi.org\/10.1371\/journal.pone.0354466","relation":{},"ISSN":["1932-6203"],"issn-type":[{"value":"1932-6203","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,8,5]]}}}