SB-BEVFusion introduces a framework-agnostic module that improves 3D object detection robustness when camera or LiDAR inputs are missing or corrupted, outperforming prior unified BEV approaches on the MultiCorrupt dataset.
Modality invariant multimodal learn- ing to handle missing modalities: A single-branch approach
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
SB-BEVFusion: Enhancing the Robustness against Sensor Malfunction and Corruptions
SB-BEVFusion introduces a framework-agnostic module that improves 3D object detection robustness when camera or LiDAR inputs are missing or corrupted, outperforming prior unified BEV approaches on the MultiCorrupt dataset.