BEVFusion: Multi-Task Multi-Sensor Fusion with
Unified Bird's-Eye View Representation

Zhijian Liu*, Haotian Tang*, Alexander Amini, Xinyu Yang, Huizi Mao, Daniela L. Rus, Song Han
Massachusetts Institute of Technology (MIT)
(* indicates equal contributions)

If you are interested in getting updates, please sign up here to get notified!

Preprint Paper



  • [06/03/22] BEVFusion ranks first on nuScenes among all solutions.
  • [06/03/22] We release the first version of BEVFusion (with pre-trained checkpoints and evaluations) on GitHub.
  • [05/26/22] BEVFusion is released on arXiv.
  • [05/02/22] BEVFusion ranks first on nuScenes among all solutions that do not use test-time augmentation and model ensemble.


Multi-sensor fusion is essential for an accurate and reliable autonomous driving system. Recent approaches are based on point-level fusion: augmenting the LiDAR point cloud with camera features. However, the camera-to-LiDAR projection throws away the semantic density of camera features, hindering the effectiveness of such methods, especially for semantic-oriented tasks (such as 3D scene segmentation). In this paper, we break this deeply-rooted convention with BEVFusion, an efficient and generic multi-task multi-sensor fusion framework. It unifies multi-modal features in the shared bird's-eye view (BEV) representation space, which nicely preserves both geometric and semantic information. To achieve this, we diagnose and lift key efficiency bottlenecks in the view transformation with optimized BEV pooling, reducing latency by more than 40x. BEVFusion is fundamentally task-agnostic and seamlessly supports different 3D perception tasks with almost no architectural changes. It establishes the new state of the art on nuScenes, achieving 1.3% higher mAP and NDS on 3D object detection and 13.6% higher mIoU on BEV map segmentation, with 1.9x lower computation cost.


Related Projects








  title={BEVFusion: Multi-Task Multi-Sensor Fusion with Unified Bird's-Eye View Representation},
  author={Liu, Zhijian and Tang, Haotian and Amini, Alexander and Yang, Xingyu and Mao, Huizi and Rus, Daniela and Han, Song},

Acknowledgments: We would like to thank Xuanyao Chen and Brady Zhou for their guidance on detection and segmentation evaluation, and Yingfei Liu and Tiancai Wang for their helpful discussions. This work was supported by National Science Foundation, Hyundai Motor, Qualcomm, NVIDIA and Apple. Zhijian Liu was partially supported by the Qualcomm Innovation Fellowship.