Information Fusion最新文献

A Physics-Guided and Geometry-Constrained Fusion Framework for Underwater Light-Beacon Salient Object Detection in Optical Docking 光学对接中水下光信标显著目标检测的物理引导和几何约束融合框架

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-02 DOI: 10.1016/j.inffus.2026.104741

Miao Yuhong, Wu Guojun, Chen Qingyan, Cheng Chensheng, Wu Yafeng

{"title":"A Physics-Guided and Geometry-Constrained Fusion Framework for Underwater Light-Beacon Salient Object Detection in Optical Docking","authors":"Miao Yuhong, Wu Guojun, Chen Qingyan, Cheng Chensheng, Wu Yafeng","doi":"10.1016/j.inffus.2026.104741","DOIUrl":"https://doi.org/10.1016/j.inffus.2026.104741","url":null,"abstract":"Accurate detection of active light beacons is a prerequisite for reliable monocular pose estimation in autonomous underwater vehicle docking. In underwater optical environments, beacon perception is degraded by backscatter-induced contrast loss and forward-scattering-induced spatial distortion, leading to high uncertainty in both photometric and structural cues. This work approaches underwater light-beacon detection as a task-driven multi-source information fusion problem. We propose a Physics-Guided and Geometry-Constrained Fusion framework for Underwater Light-Beacon detection (PGF-ULB) that integrates heterogeneous but complementary information sources derived from physical imaging models and intrinsic target geometry. A physics-guided saliency map is generated by reformulating the classical underwater imaging model into an inverse guidance process, enabling effective suppression of backscatter and attenuation of forward-scattering halos. In parallel, a geometry-constrained saliency map is constructed by exploiting the radial symmetry of active light beacons through gradient field consistency analysis, providing a structure-based cue that is robust to intensity degradation. These two complementary saliency maps are integrated via a two-level adaptive fusion framework based on a Conditional Random Field (CRF): image-level Pearson weighting balances contributions as adaptive unary potentials, and pixel-level pairwise potentials enforce spatial coherence, achieving robust and degradation-adaptive saliency estimation. Extensive experiments on a real-world underwater docking dataset demonstrate that the proposed framework consistently outperforms both state-of-the-art saliency detection methods and conventional beacon detection pipelines in structural accuracy, threshold robustness, and centroid localization precision. Furthermore, integration into a downstream monocular pose estimation pipeline improves pose availability, validating the practical effectiveness of the proposed fusion strategy for autonomous underwater docking.","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"38 1","pages":""},"PeriodicalIF":18.6,"publicationDate":"2026-09-02","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148884303","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

DAE-Net: A Dual-View Asymmetric Extraction Network with Pathology-Aware Multi-level Fusion for Gastrointestinal Tumour Classification DAE-Net：一个具有病理感知的多级融合的双视图非对称提取网络用于胃肠道肿瘤分类

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-01 DOI: 10.1016/j.inffus.2026.104746

Siyao Liu,Jiayi Xu,Jiaxi Mao,Yazhou Wang,Hengrui Liu,Wenyuan Guo,Yong Zhang,Yueming Zhang

引用次数: 0

FuSS-Net: Multi-level Fusion of Frequency Priors and Semantic State Space Dynamics for Synergistic Lesion Segmentation and Classification fss - net：基于频率先验和语义状态空间动态的多级融合协同病灶分割与分类

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-01 DOI: 10.1016/j.inffus.2026.104742

Tianxiang Li,Yang Liu,Witold Pedrycz,Weiping Ding,Yuchun Sun,Sukun Tian

引用次数: 0

Adaptive Cross-View EEG Fusion of External Scalp and Internal Neural Representations for Affective Brain-Computer Interfaces 基于情感脑机接口的外部头皮自适应交叉视点脑电融合与内部神经表征

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-01 DOI: 10.1016/j.inffus.2026.104754

Tao Wang,Peng Wang,Suiyu Liu,Xiaoyan Wang,Li Shen,Shuang Liu,Dong Ming

引用次数: 0

Reliability-Guided Hypergraph Feature Fusion for Underwater Object Detection 基于可靠性引导的超图特征融合水下目标检测

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-01 DOI: 10.1016/j.inffus.2026.104761

Jingchun Zhou,Yuang Li,Dehuan Zhang,Chenzhao Bai,Guangyong Chen,Qiuping Jiang,Ghulam Muhammad

引用次数: 0

LoRAFusion: Infrared and Visible Image Fusion via Low-Rank Adaptation for Multi-Requirement Tasks LoRAFusion：基于低秩自适应的红外和可见光图像融合

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-01 DOI: 10.1016/j.inffus.2026.104759

Liangjie Song,Hanlin Xu,Gang Liu,Gang Xiao,Durga Prasad Bavirisetti

引用次数: 0

STReF-NDVI: A spatio-temporal SAR-optical fusion framework for NDVI time series reconstruction under persistent cloud cover STReF-NDVI：用于持续云层覆盖下NDVI时间序列重建的时空sar -光学融合框架

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-01 DOI: 10.1016/j.inffus.2026.104760

Pengfei Jia,Qiqi Zhu,Zhi Zheng,Helmi Zulhaidi Mohd Shafri,Abdul Rashid Mohamed Shariff,Qingfeng Guan

引用次数: 0

Fusion of audio-visual data through illumination aware and parallel cross-guided for vehicle detection 通过光照感知和平行交叉引导融合视听数据实现车辆检测

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-01 DOI: 10.1016/j.inffus.2026.104763

Mingyang Hao,Fangli Ning,Zhongshan Wang,Feng Wang

引用次数: 0

CL3E-GNN: Three-Phase Attention with Feature Fusion based on Curriculum Implementation Strategy for Class Imbalanced Node Classification CL3E-GNN：基于类不平衡节点分类课程实施策略的三阶段关注特征融合

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-09-01 DOI: 10.1016/j.inffus.2026.104745

Abdul Joseph Fofanah,Lian Wen,David Chen,Shaoyang Zhang,Alpha Alimamy Kamara

引用次数: 0

DPG-CD: Depth-Prior-Guided Cross-Modal Joint 2D–3D Change Detection DPG-CD：深度先验引导的跨模态关节2D-3D变化检测

IF 18.6 1区计算机科学

Information Fusion Pub Date : 2026-08-31 DOI: 10.1016/j.inffus.2026.104743

Luqi Zhang, Zhen Dong, Bisheng Yang

{"title":"DPG-CD: Depth-Prior-Guided Cross-Modal Joint 2D–3D Change Detection","authors":"Luqi Zhang, Zhen Dong, Bisheng Yang","doi":"10.1016/j.inffus.2026.104743","DOIUrl":"https://doi.org/10.1016/j.inffus.2026.104743","url":null,"abstract":"Urban spatial evolution is manifested not only through horizontal expansion but also through vertical structural changes. Consequently, jointly capturing 2D semantic changes and 3D height changes is essential for urban morphology analysis and emergency management. In practical scenarios, collecting 3D observations is often constrained by high acquisition costs and the inability to support frequent updates. The multi-temporal cross-modal input consisting of pre-event Digital Surface Model (DSM) and post-event imagery provides a practical solution for 3D change detection in high-frequency urban monitoring, disaster assessment, and emergency response scenarios. However, this setting remains challenging as imagery and DSM data exhibit significant spectral-geometric representation discrepancy. Moreover, modality differences may be confused with actual changes, and robust change detection requires effective fusion of semantic and geometric features from multi-temporal data. In this paper, we propose DPG-CD, a depth-prior-guided multi-temporal cross-modal fusion framework for joint 2D semantic and 3D height change detection. Specifically, an estimated depth prior is introduced into the image modality to provide complementary geometric and structural information for the optical modality. A gated fusion mechanism then selectively injects geometric cues from depth prior while preserving discriminative spectral representations. Subsequently, a multi-stage cross-temporal cross-modal feature fusion architecture is employed to extract change-aware features. Finally, a multi-task decoder jointly predicts 2D semantic changes and 3D height changes, complemented by an auxiliary DSM prediction task to improve structural consistency and height estimation accuracy. Experiments on two public datasets, Hi-BCD and 3DCD, and a new dataset, NYC-MMCD, demonstrate that DPG-CD outperforms state-of-the-art methods on both 2D and 3D change detection tasks. The code and the NYC-MMCD dataset will be made publicly available online at: <ce:inter-ref xlink:href=\"https://github.com/zhangluqi0209/DPG-CD\" xlink:type=\"simple\">https://github.com/zhangluqi0209/DPG-CD</ce:inter-ref>.","PeriodicalId":50367,"journal":{"name":"Information Fusion","volume":"135 1","pages":""},"PeriodicalIF":18.6,"publicationDate":"2026-08-31","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148884304","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"计算机科学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0