GAFusion: Adaptive Fusing LiDAR and Camera with Multiple Guidance for 3D Object Detection
Open accessAbstract
Recent years have witnessed the remarkable progress of 3D multi-modality object detection methods based on the Bird's-Eye-View (BEV) perspective. However, most of them overlook the complementary interaction and guidance be-tween LiDAR and camera. In this work, we propose a novel multi-modality 3D objection detection method, named GA-Fusion, with LiDAR-guided global interaction and adaptive fusion. Specifically, we introduce sparse depth guidance (SDG) and LiDAR occupancy guidance (LOG) to generate 3D features with sufficient depth information. In the following, LiDAR-guided adaptive fusion transformer (LGAFT) is developed to adaptively enhance the interaction of different modal BEV features from a global perspective. Meanwhile, additional downsampling with sparse height compression and multi-scale dual-path transformer (MSDPT) are de-signed to enlarge the receptive fields of different modal features. Finally, a temporal fusion module is introduced to ag-gregate features from previous frames. GAFusion achieves state-of-the-art 3D object detection results with 73.6% mAP and 74.9% NDS on the nuScenes test set.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
RRPN maps radar detections to image coordinate system, generating anchor boxes for quick ROI suggestions.
“RRPN [91] maps radar detections to the image coordinate system, generating anchor boxes for quick ROI suggestions.”
From Multimodal fusion and vision–language models: A survey for robot vision (2026) · cites this paper as [91]
Cites
Swin Transformer: Hierarchical Vision Transformer using Shifted Windows
Liu, Z., Lin, Y., Cao, Y., Hu, H., Wei, Y., Zhang, Z., Lin, S., & Guo, B. (2021) · 2021 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv48922.2021.00986 · Cited by 35,534 · cited 2× · see citing works
“To improve computational efficiency, they both use windowed self-attention [25], and share weights.”
MethodBidirectional recurrent neural networks
Schuster, M., & Paliwal, K. K. (1997) · IEEE Transactions on Signal Processing · doi:10.1109/78.650093 · Cited by 10,172 · cited 1× · see citing works
“3D-VID [46] employs a bidirectional recurrent neural network (Bi-RNN) [32] to model the temporal sequences of multiple point clouds, capturing the motion information and state changes of the targets.”
BackgroundPointNet: Deep Learning on Point Sets for 3D Classification and Segmentation
Charles, R., Su, H., Mo, K., & Guibas, L. J. (2017) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr.2017.16 · Cited by 9,767 · cited 2× · see citing works
“PointNet [31] is the first framework that processes point clouds in an end-to-end manner, by taking unordered point cloud sets as direct inputs and preserving the spatial structure of point clouds.”
BackgroundMethodnuScenes: A Multimodal Dataset for Autonomous Driving
Caesar, H., Bankiti, V., Lang, A. H., Vora, S., Liong, V. E., Xu, Q., Krishnan, A., Pan, Y., Baldan, G., & Beijbom, O. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.01164 · Cited by 7,152 · cited 1× · see citing works
“Similar to previous works [1, 22, 26 nuScenes [2] dataset is a large-scale autonomous driving benchmark, which includes 1000 scenes with images from 6 cameras with surrounding views, points from 5 Radars and 1 LiDAR.”
ResultVoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection
Zhou, Y., & Tuzel, O. (2018) · 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition · doi:10.1109/cvpr.2018.00472 · Cited by 4,813 · cited 3× · see citing works
“VoxelNet [47] dis-cretizes point clouds into voxels, and uses dense convolution to obtain BEV features.”
MethodPointPillars: Fast Encoders for Object Detection From Point Clouds
Lang, A. H., Vora, S., Caesar, H., Zhou, L., Yang, J., & Beijbom, O. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.01298 · Cited by 4,673 · cited 1× · see citing works
“Existing methods [15, 18, 31, 33, 34, 47] either directly predict on point clouds, or convert point clouds into voxels or pillars.”
MethodMulti-view 3D Object Detection Network for Autonomous Driving
Chen, X., Ma, H., Wan, J., Li, B., & Xia, T. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.691 · Cited by 3,450 · cited 3× · see citing works
“MV3D [4] associates 3D proposals with 2D RoI features and converts 3D representation into 2D pseudo images, enabling the network to leverage 2D convolutions for geometric refinement.”
BackgroundMethodSECOND: Sparsely Embedded Convolutional Detection
Yan, Y., Mao, Y., & Li, B. (2018) · Sensors · doi:10.3390/s18103337 · Cited by 3,352 · cited 1× · see citing works
“For the LiDAR stream , a common method [1, 22, 26] is to use 3D sparse convolution [41] to extract single-scale features from the voxelized point clouds, which has a weak feature representation with limited receptive fields.”
MethodPV-RCNN: Point-Voxel Feature Set Abstraction for 3D Object Detection
Shi, S., Guo, C., Jiang, L., Wang, Z., Shi, J., Wang, X., & Li, H. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.01054 · Cited by 2,341 · cited 1× · see citing works
“Existing methods [15, 18, 31, 33, 34, 47] either directly predict on point clouds, or convert point clouds into voxels or pillars.”
MethodCenter-based 3D Object Detection and Tracking
Yin, T., Zhou, X., & Krähenbühl, P. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.01161 · Cited by 2,197 · see citing works
Joint 3D Proposal Generation and Object Detection from View Aggregation
Ku, J., Mozifian, M., Lee, J., Harakeh, A., & Waslander, S. L. (2018) · 2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS) · doi:10.1109/iros.2018.8594049 · Cited by 1,697 · cited 2× · see citing works
“Early methods [4, 14, 35, 45] achieve LiDAR-camera fusion by projecting 3D LiDAR point clouds (or region proposals) onto 2D images.”
BackgroundMethodLi, Z., Wang, W., Li, H., Xie, E., Sima, C., Lu, T., Yu, Q., & Dai, J. (2022) · arXiv · doi:10.48550/arxiv.2203.17270 · Cited by 1,419 · cited 1× · see citing works
“BEVformer [19] is a transformer-based 3D object detection model that uses BEV to represent the scene and exploits multiple frames of images for spatial-temporal information fusion.”
BackgroundBevfusion: Multi-task multi-sensor fusion with unified bird’s-eye view representation
Liu, Z., Tang, H., Amini, A., Yang, X., Mao, H., Rus, D. L., & Han, S. (2023) · 2023 IEEE International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra48891.2023.10160968 · Cited by 1,307 · cited 10× · see citing works
“Sparse features F c and their positions P c are obtained as follows: For the camera stream , Following the previous works [22, 26], we input multi-view images into backbone to obtain 2D image features F c ∈ R N c × C × H × W with sufficient semantic information, where N c , C , H , W denote the…”
BackgroundMethodResultInfluentialLift, Splat, Shoot: Encoding Images from Arbitrary Camera Rigs by Implicitly Unprojecting to 3D
Philion, J., & Fidler, S. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58568-6_12 · Cited by 1,271 · cited 1× · see citing works
“Recently, BEVFusion [22, 26] are proposed to fuse LiDAR features and camera features in BEV space and apply a lift-splat-shoot (LSS) [30] operation to project image features, resulting in semantic-rich features.”
MethodPointPainting: Sequential Fusion for 3D Object Detection
Vora, S., Lang, A. H., Helou, B., & Beijbom, O. (2020) · 2020 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr42600.2020.00466 · Cited by 1,236 · cited 1× · see citing works
“PointPainting [36] enriches point clouds with semantic labels from images.”
BackgroundBEVDepth: Acquisition of Reliable Depth for Multi-view 3D Object Detection
Li, Y., Ge, Z., Yu, G., et al. (2022) · AAAI Conference on Artificial Intelligence · doi:10.48550/arxiv.2206.10092 · Cited by 1,112 · cited 1×
“BEVDet [11] and BEVDepth [16] predict the depth distribution to lift the image features to a 3D frustum meshgrid.”
BackgroundVoxel R-CNN: Towards High Performance Voxel-based 3D Object Detection
Deng, J., Shi, S., Li, P., Zhou, W., Zhang, Y., & Li, H. (2021) · Proceedings of the AAAI Conference on Artificial Intelligence · doi:10.1609/aaai.v35i2.16207 · Cited by 1,048 · cited 1× · see citing works
“They are used as the input of view transformation, and finally voxel pooling [9] is employed to generate the image 3D feature volume, which is denoted as F ′ c , F ′ c ∈ R C × Z × H × W .”
MethodTransFusion: Robust LiDAR-Camera Fusion for 3D Object Detection with Transformers
Bai, X., Hu, Z., Zhu, X., Huang, Q., Chen, Y., Fu, H., & Tai, C.-L. (2022) · 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr52688.2022.00116 · Cited by 990 · cited 5× · see citing works
“To prove the validity and generalization of this module, we separately insert the developed module into TransFusion [1] and BEVFusion [26], as shown in Table 3.”
MethodResultInfluentialPETR: Position Embedding Transformation for Multi-View 3D Object Detection
Liu, Y.-F., Wang, T.-C., Zhang, X., & Sun, J. (2022) · European Conference on Computer Vision · doi:10.48550/arxiv.2203.05625 · Cited by 929 · cited 1×
“DETR3D [38] and PETR [23] introduce transformer into the framework, wherein the former aggregates 2D features into 3D Query, and the latter embeds coordinate information into 2D features.”
BackgroundShi, S., Jiang, L., Deng, J., Wang, Z., Guo, C., Shi, J., Wang, X., & Li, H. (2022) · International Journal of Computer Vision · doi:10.1007/s11263-022-01710-9 · Cited by 571 · cited 1× · see citing works
“Existing methods [15, 18, 31, 33, 34, 47] either directly predict on point clouds, or convert point clouds into voxels or pillars.”
MethodMvx-net: Multimodal voxelnet for 3d object detection
Sindagi, V. A., Zhou, Y., & Tuzel, O. (2019) · 2019 International Conference on Robotics and Automation (ICRA) · doi:10.1109/icra.2019.8794195 · Cited by 554 · cited 2× · see citing works
“Early methods [4, 14, 35, 45] achieve LiDAR-camera fusion by projecting 3D LiDAR point clouds (or region proposals) onto 2D images.”
BackgroundMethodBEVDet4D: Exploit Temporal Cues in Multi-camera 3D Object DetectionPreprint
Huang, J., & Huang, G. (2022) · arXiv.org · doi:10.48550/arxiv.2203.17054 · Cited by 554 · cited 3× · see citing works
“We follow the fusion scheme of BEVDet4D [10] and store the BEV features of historical frames in a memory buffer, and fuse the BEV features of the previous frame at each time.”
BackgroundMethodInfluentialYoo, J. H., Kim, Y., Kim, J., & Choi, J. W. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58583-9_43 · Cited by 520 · cited 2× · see citing works
“Early methods [4, 14, 35, 45] achieve LiDAR-camera fusion by projecting 3D LiDAR point clouds (or region proposals) onto 2D images.”
BackgroundMethodPETRv2: A Unified Framework for 3D Perception from Multi-Camera Images
Liu, Y., Yan, J., Jia, F., Li, S., Gao, A., Wang, T., & Zhang, X. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00302 · Cited by 400 · see citing works
Tri-Perspective View for Vision-Based 3D Semantic Occupancy Prediction
Huang, Y., Zheng, W., Zhang, Y., Zhou, J., & Lu, J. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00890 · Cited by 317 · cited 2× · see citing works
“Inspired by the occupancy task [12], LOG guides a 3D feature volume generated by view transformation with occupancy features, and focuses on the targets in the 3D feature volume, thus providing more valuable information for fusion.”
BackgroundMethodBEVFusion: A Simple and Robust LiDAR-Camera Fusion FrameworkPreprint
Liang, T., Xie, H., Yu, K., Xia, Z., Lin, Z., Wang, Y., Tang, T., Wang, B., & Tang, Z. (2022) · arXiv (Cornell University) · doi:10.48550/arxiv.2205.13790 · Cited by 310 · cited 8× · see citing works
“Sparse features F c and their positions P c are obtained as follows: For the camera stream , Following the previous works [22, 26], we input multi-view images into backbone to obtain 2D image features F c ∈ R N c × C × H × W with sufficient semantic information, where N c , C , H , W denote the…”
BackgroundMethodResultInfluentialBEVDet: High-performance Multi-camera 3D Object Detection in Bird-Eye-ViewPreprint
Huang, J., Huang, G., Zhu, Z. D., Ye, Y., & Du, D. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2112.11790 · Cited by 292 · cited 1× · see citing works
“BEVDet [11] and BEVDepth [16] predict the depth distribution to lift the image features to a 3D frustum meshgrid.”
BackgroundVoxFormer: Sparse Voxel Transformer for Camera-Based 3D Semantic Scene Completion
Li, Y., Yu, Z., Choy, C., Xiao, C., Álvarez, J. M., Fidler, S., Feng, C., & Anandkumar, A. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.00877 · Cited by 252 · cited 1× · see citing works
“VoxFromer [17] adopts a two-stage design and generates a set of sparse visible and occupied voxel queries from depth estimation.”
MethodLiDAR R-CNN: An Efficient and Universal 3D Object Detector
Li, Z., Wang, F., & Wang, N. (2021) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr46437.2021.00746 · Cited by 246 · cited 1× · see citing works
“Existing methods [15, 18, 31, 33, 34, 47] either directly predict on point clouds, or convert point clouds into voxels or pillars.”
MethodFB-OCC: 3D Occupancy Prediction based on Forward-Backward View Transformation
Li, Z., Yu, Z.-D., Austin, D., et al. (2023) · arXiv.org · doi:10.48550/arxiv.2307.01492 · Cited by 220 · cited 1×
“Based on FB-BEV [21], FB-OCC [20] emphasizes the importance of model scale and pre-training.”
BackgroundOccFormer: Dual-path Transformer for Vision-based 3D Semantic Occupancy Prediction
Zhang, Y., Zhu, Z. D., & Du, D. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00865 · Cited by 216 · cited 2× · see citing works
“VoxFromer [17] adopts a two-stage design and generates a set of sparse visible and occupied voxel queries from depth estimation.”
MethodOpenOccupancy: A Large Scale Benchmark for Surrounding Semantic Occupancy Perception
Wang, X., Zhu, Z., Xu, W., Zhang, Y., Wei, Y., Chi, X., Ye, Y., Du, D., Lu, J., & Wang, X. (2023) · 2023 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01636 · Cited by 192 · cited 1× · see citing works
“OpenOccu-pancy [37] is the first omnidirectional semantic occupancy perception benchmark.”
BackgroundDETR3D: 3D Object Detection from Multi-view Images via 3D-to-2D Queries
Wang, Y., Guizilini, V., Zhang, T., Wang, Y., Zhao, H., & Solomon, J. (2021) · arXiv · doi:10.48550/arxiv.2110.06922 · Cited by 182 · cited 1× · see citing works
“DETR3D [38] and PETR [23] introduce transformer into the framework, wherein the former aggregates 2D features into 3D Query, and the latter embeds coordinate information into 2D features.”
BackgroundVision-Centric BEV Perception: A Survey
Ma, Y., Wang, T., Bai, X., Yang, H., Hou, Y., Wang, Y., Qiao, Y., Yang, R., & Zhu, X. (2024) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2024.3449912 · Cited by 180 · cited 1× · see citing works
“Recent works [1, 3, 13, 22, 26, 39, 40, 42] adopt different query generation strategies or create a unified Bird’s-Eye-View (BEV) [28] intermediate feature to fuse multi-modality features.”
MethodJiao, Y., Jie, Z., Chen, S., Chen, J., Ma, L., & Jiang, Y. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.02073 · Cited by 162 · cited 2× · see citing works
“The re-sults show that, on the test set, GAFusion surpasses all the existing methods with 73.6 % mAP and 74.9 % NDS, such as MSMDFusion [13] and CMT [40].”
MethodInfluentialSparseFusion: Fusing Multi-Modal Sparse Representations for Multi-Sensor 3D Object Detection
Xie, Y., Xu, C., Rakotosaona, M., Rim, P., Tombari, F., Keutzer, K., Tomizuka, M., & Zhan, W. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.01613 · Cited by 147 · cited 1× · see citing works
“Recent works [1, 3, 13, 22, 26, 39, 40, 42] adopt different query generation strategies or create a unified Bird’s-Eye-View (BEV) [28] intermediate feature to fuse multi-modality features.”
MethodFB-BEV: BEV Representation from Forward-Backward View Transformations
Li, Z., Yu, Z., Wang, W., Anandkumar, A., Lü, T., & Álvarez, J. M. (2023) · IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv51070.2023.00637 · Cited by 141 · cited 1× · see citing works
“Based on FB-BEV [21], FB-OCC [20] emphasizes the importance of model scale and pre-training.”
BackgroundLargeKernel3D: Scaling up Kernels in 3D Sparse CNNs
Chen, Y., Liu, J., Zhang, X., Qi, X., & Jia, J. (2023) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52729.2023.01296 · Cited by 139 · see citing works
Multimodal Virtual Point 3D DetectionPreprint
Yin, T., Zhou, X., & Krähenbühl, P. (2021) · arXiv (Cornell University) · doi:10.48550/arxiv.2111.06881 · Cited by 130 · see citing works
DeepInteraction: 3D Object Detection via Modality Interaction
Yang, Z., Chen, J., Miao, Z., Li, W., Zhu, X., & Zhang, L. (2022) · arXiv · doi:10.48550/arxiv.2208.11112 · Cited by 108 · cited 1× · see citing works
“Recent works [1, 3, 13, 22, 26, 39, 40, 42] adopt different query generation strategies or create a unified Bird’s-Eye-View (BEV) [28] intermediate feature to fuse multi-modality features.”
MethodAutoAlignV2: Deformable Feature Aggregation for Dynamic Multi-Modal 3D Object Detection
Chen, Z., Li, Z.-Y., Zhang, S., et al. (2022) · arXiv.org · doi:10.48550/arxiv.2207.10316 · Cited by 98
Cai, H., Zhang, Z., Zhou, Z., et al. (2023) · arXiv.org · doi:10.48550/arxiv.2303.17099 · Cited by 77 · cited 1×
“Recent works [1, 3, 13, 22, 26, 39, 40, 42] adopt different query generation strategies or create a unified Bird’s-Eye-View (BEV) [28] intermediate feature to fuse multi-modality features.”
MethodOccDepth: A Depth-Aware Method for 3D Semantic Scene CompletionPreprint
Miao, R., Liu, W., Chen, M., Gong, Z., Xu, W., Chen, H., & Zhou, S. (2023) · arXiv (Cornell University) · doi:10.48550/arxiv.2302.13540 · Cited by 38 · cited 1× · see citing works
“OccDepth [29] leverages the implicit depth information from depth images (or RGBD images) to help recover the 3D geometry.”
BackgroundMuti-Frame Point Cloud Feature Fusion Based on Attention Mechanisms for 3D Object Detection
Zhai, Z., Wang, Q., Pan, Z., Gao, Z., & Hu, W. (2022) · Sensors · doi:10.3390/s22197473 · Cited by 11 · cited 1× · see citing works
“3D-VID [46] employs a bidirectional recurrent neural network (Bi-RNN) [32] to model the temporal sequences of multiple point clouds, capturing the motion information and state changes of the targets.”
Background
Flagged first, then by citations
Cited by
Sorted by citations
Multimodal fusion and vision–language models: A survey for robot vision
Han, X., Chen, S., Fu, Z., Feng, Z., Fan, L., An, D., Wang, C., Guo, L., Meng, W., Zhang, X., et al. (2026) · Information Fusion · doi:10.1016/j.inffus.2025.103652 · Cited by 41 · see citing works
“Future research will likely focus on improving scalability, real-time capabilities, and developing unified frameworks that integrate localization with high-level perception and decision-making tasks [145].”
Integrating Multi-Modal Sensors: A Review of Fusion Techniques for Intelligent Vehicles
Wei, C., Qin, Z., Zhang, Z., Wu, G., & Barth, M. (2025) · IEEE Intelligent Vehicles Symposium · doi:10.1109/iv64158.2025.11097396 · Cited by 35 · see citing works
“…2022 ✓ ✓ ✓ Object Detection BEVfusion [40] 2023 ✓ ✓ ✓ Object Detection RCBEV [41] 2023 ✓ ✓ ✓ Object Detection CRKD [42] 2024 ✓ ✓ ✓ Object Detection GATR [43] 2024 ✓ ✓ ✓ Object Detection GAFusion [44] 2024 ✓ ✓ ✓ Object Detection HyDRa [45] 2024 ✓ ✓ ✓ 3D Semantic Occupancy Prediction CO-Occ [46] 2024”
Collaborating Constrained and Unconstrained Encodings for Cross-Modal Salient Object Detection
Yao, Z., Gao, W., Li, G., & Zhao, T. (2025) · IEEE Transactions on Emerging Topics in Computational Intelligence · doi:10.1109/tetci.2025.3577495 · Cited by 8 · see citing works
“[44]designedaLiDAR-guided adaptivefusiontransformertofurtherenhancetheintegrationof features from both modalities.”
Zhu, Y., Jia, X., Yang, X., & Yan, J. (2025) · Proceedings - IEEE International Conference on Robotics and Automation/Proceedings · doi:10.1109/icra55743.2025.11128084 · Cited by 7 · see citing works
“However, they heavily rely on the feature quality of single branch and thus require large backbones (e.g., VoVNet, Swin-Tiny) which is less efficient.”
MoRAL: Motion-Aware Multi-Frame 4D Radar and LiDAR Fusion for Robust 3D Object Detection
Peng, X., Wang, Y., Tang, M., et al. (2025) · IEEE International Conference on Intelligent Transportation Systems (ITSC) · doi:10.1109/itsc60802.2025.11423633 · Cited by 4 · see citing works
Ev-3DOD: Pushing the Temporal Boundaries of 3D Object Detection with Event Cameras
Cho, H., Kang, J., Kim, Y.-H., & Yoon, K. (2025) · IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings · doi:10.1109/cvpr52734.2025.02533 · Cited by 7 · see citing works
“Multi-modal sensor fusion [5, 34, 36, 47, 48, 56, 57, 59, 90, 101, 102, 113], where different sensors complement each other, typically utilizes a combination of cameras and LiDAR.”
RCFusionNet: A Radar-Guided Two-Stage Fusion Approach to Effective Multimodal Perception
Xu, W., Meng, X., Hu, Y. H., & Deng, H. (2025) · IEEE Sensors Journal · doi:10.1109/jsen.2025.3592236 · Cited by 3 · see citing works
“GA-Fusion [37] introduces a LiDAR-guided fusion strategy that utilizes LiDAR occupancy to generate camera BEV features and employs a transformer to adaptively integrate multimodal features.”
Robust 3D Object Detection in Complex Traffic via Unified Feature Alignment in Bird’s Eye View
Liu, A., Zhang, Y., Shi, H., & Chen, J. (2025) · World Electric Vehicle Journal · doi:10.3390/wevj16100567 · Cited by 3 · see citing works
“In contrast, unified BEV fusion methods like GAFusion [5], and BEVFusion [9], project image features into the BEV space, where they are deeply fused with LiDAR representations to form a unified spatial understanding.”
InfluentialConfidence-Guided LiDAR Depth Completion for Robust 3D Object Detection
Hwang, S., & Lee, K. (2025) · IEEE Access · doi:10.1109/access.2025.3609143 · Cited by 2 · see citing works
“Recent deep fusion models, such as GAFusion [58], BEVFusion [59], and FUTR3D [60], employ transformer architectures and unified BEV representations to achieve robust performance across diverse scenarios.”
Qu, X., Qin, K., Li, Y.-P., Zhang, S. B., Li, Y., Shen, S., & Gao, Y. (2025) · Journal of Imaging · doi:10.3390/jimaging11090319 · Cited by 3 · see citing works
“Recent advancements attempt to overcome these limitations from different angles: GAFusion [46] introduces LiDAR-guided global and local depth priors for adaptive fusion; SimpleBEV [47] offers a lightweight and efficient BEV fusion architecture; GraphBEV [43] achieves robust BEV feature alignment…”
Noguchi, C., & Yamamoto, T. (2025) · IEEE International Conference on Computer Vision Workshops · doi:10.1109/iccvw69036.2025.00323 · Cited by 1 · see citing works
DepthFusion: Depth-Aware Hybrid Feature Fusion for LiDAR-Camera 3D Object Detection
Ji, M., Yang, J. J., & Zhang, S. (2026) · IEEE Transactions on Multimedia · doi:10.1109/tmm.2026.3668596 · Cited by 1 · see citing works
Dual-View Stereo Vision Sensor Based on Single Camera and Planar Mirror for 3-D Measurement
Xing, R., & Zhou, F. (2026) · IEEE Sensors Journal · doi:10.1109/jsen.2026.3696981 · Cited by 1 · see citing works
XMatchAD: A Cross-Modal Matching Perspective on Reconstruction-Based Anomaly Detection
Cai, M., Zhang, Z., Wu, G., et al. (2026) · IEEE Transactions on Image Processing · doi:10.1109/tip.2026.3718437 · Cited by 1 · see citing works
“In a related approach, Li et al. [42] employ sparse depth supervision and LiDAR occupancy guidance to enhance representations.”
Dense projection fusion for 3D object detection
Chen, Z., Hu, B., Luo, C., Chen, G., & Zhu, H. (2024) · Scientific Reports · doi:10.1038/s41598-024-74679-9 · Cited by 1 · see citing works
“Most of the latest approaches 18,21,28,29 use TransFusion-L 19 as their LiDAR-only baseline and add customized image branch on top of it.”
MethodTargetless Extrinsic Calibration for LiDAR and Camera Systems with Intensity Information
Jiang, W., Zhang, Y., Lü, T., et al. (2024) · International Conference on Mechatronics, Robotics and Automation (ICMRA) · doi:10.1109/icmra62519.2024.10809102
MGAF: LiDAR-Camera 3D Object Detection With Multiple Guidance and Adaptive Fusion
Fan, B., Li, X., Zhou, Y., Xia, C., Fan, H., Xu, F., & Tian, J. (2025) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2025.3612958
Self-citationA multi-graph aggregation network for robust multi-drone trajectory prediction
Liu, X., Zhang, G., Zhong, L., Wang, B., & Zhou, B. (2026) · Journal of King Saud University - Computer and Information Sciences · doi:10.1007/s44443-026-00580-2
MVOD - A Multi-View Fusion Network for 3D Obstacle Detection with Point Clouds and Images
Wang, J., Wang, Z., Mi, H., et al. (2025) · International Conference on Robotics, Automation and Artificial Intelligence (RAAI) · doi:10.1109/raai67517.2025.11423379
TGEM: Text-Guided Fusion of LiDAR and Camera for Robust 2D Object Detection
Jiang, T., Ma, L., & Hua, C. (2026) · International Conference on Big Data, Information and Computer Network (BDICN) · doi:10.1145/3801228.3801346
UniPT: A Unified Representation Pre-training for Multi-dataset 3D Object Detection
Zheng, Z., Wang, C., Huang, Z., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-981-95-7084-3_14
Daohu, Z., Fu, D., Sai, Q., et al. (2026) · Transactions of the Institute of Measurement and Control · doi:10.1177/01423312261445534
Chameleon: 3-D Object Detection via Adaptive Multisensor Decoupling for Autonomous Vehicles
Xing, L., Huang, Y., Deng, K., Gao, J., Wu, H., & Ma, H. (2026) · IEEE Internet of Things Journal · doi:10.1109/jiot.2026.3695856
“GAFusion [21] guides image depth estimation with LiDAR, enhances multiscale image features via transformers, and fuses BEV representations across sensors.”
Agglomerative Model Meets Multi-Scale Adaptive Fusion for Cross-Modal Unsupervised Domain Adaptation
Wang, Z., Kuang, L., Wang, S., Jiao, S., Chen, Z., & Xiong, F. (2026) · International Conference on Multimedia Retrieval · doi:10.1145/3805622.3810817
Real-Time 3D Object Detection Algorithm Based on the Fusion of Image and Point Cloud
Shan, W., Chen, X., & Xu, Z. (2026) · Chinese Control and Decision Conference (CCDC) · doi:10.1109/ccdc69976.2026.11559944
“GAFusion [13] introduces LiDAR-guided adaptive transformers with multiple guidance signals to enhance global interaction between BEV features from different sensors, demonstrating improved 3D detection performance on benchmarks such as nuScenes.”
Yu, X., Xu, L., Qian, Y., et al. (2026) · IET Intelligent Transport Systems · doi:10.1049/itr2.70273
FineVib: A Fine-grained Way to Learn Vibration Features via mmWave Radar
Tian, W., Xiong, R., Yan, X., et al. (2026) · International Conference on Computer Supported Cooperative Work in Design (CSCWD) · doi:10.1109/cscwd68734.2026.11582230
“Cameras and 3D LIDAR [10], [11] are utilized and a point cloud and a convolutional neural network (CNN) are used respectively for further object recognition.”
Park, H., Jeong, J., & Yoon, K. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37517-9_12 · Cited by 1 · see citing works
Wei, W., Zhang, Y., Peng, G., et al. (2026) · Electronics · doi:10.3390/electronics15143110
Zhanga, G., Feng, Y., Zhoub, L., et al. (2026) · International Conference on Electronic Engineering and Informatics (EEI) · doi:10.1109/eei70303.2026.11640663
Qi, J., & Wang, J. (2026) · Journal of King Saud University - Computer and Information Sciences · doi:10.1007/s44443-026-01134-2
Ma, L., Hao, M., Tian, Y., Bi, Y., & Zhang, L. (2026) · IEEE Transactions on Instrumentation and Measurement · doi:10.1109/tim.2026.3726219
“GAFusion introduces LiDAR-guided depth and occupancy information together with adaptive cross-modal interaction to enhance camera–LiDAR BEV features [38], while IS-Fusion jointly models scene-and instance-level multimodal context for 3-”
A LiDAR-Based Multimodal 3D Object Detection Algorithm for Intelligent Driving in Open-Pit Mines
Wang, S., & Zhang, X. (2026) · Electronics · doi:10.3390/electronics15184123
BitRIC: Efficient Neural Compression of LiDAR Range Images via Hierarchical Bitplanes
You, K., Chen, T., Ding, D., et al. (2026) · Lecture notes in computer science · doi:10.1007/978-3-032-37174-4_32
FGP-BEVFusion for Lightweight Camera–LiDAR 3D Object Detection
Shen, Z., Huang, C., Huo, G., et al. (2026) · International Conference on Computer Science and Blockchain (CCSB) · doi:10.1109/ccsb71176.2026.11716350
- Authors
- Li, X., Fan, B., Tian, J., & Fan, H.
- Affiliations
- Nanjing University of Posts and Telecommunications (CN)
- Shenyang Institute of Automation (CN)
- Venue
- 2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2024-06-16
- Pages
- 21209-21218
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W4402780115
- Cited by
- 42
- References
- 44
- Citations by year
- 2026: 26 · 2025: 14 · 2024: 2
- FWCI
- 8.62
- Type
- conference paper
- Language
- en
- Open access
- green
- arxiv.org
- Locations
- IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings - IEEE Computer Society Conference on Computer Vision and Pattern Recognition/Proceedings (publishedVersion)
- arXiv (Cornell University) (submittedVersion)
- arXiv (Cornell University)
- Funders
- National Natural Science Foundation of China (U2013210)
- Topics
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Neural Network Applications
- Physical Sciences › Engineering › Automotive Engineering › Autonomous Vehicle Technology and Safety
- Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Advanced Vision and Imaging
- Keywords
- 3D object detection, LiDAR-camera fusion, multimodal fusion, bird's-eye view, temporal fusion, nuScenes
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
