BEVDilation: LiDAR-Centric Multi-Modal Fusion for 3D Object Detection

TL;DR

BEVDilation introduces a LiDAR-centric multi-modal fusion framework for 3D object detection that effectively alleviates spatial misalignment and enhances robustness by prioritizing LiDAR data and using image guidance.

Contribution

It proposes a novel LiDAR-centric fusion method with implicit image guidance, including Sparse Voxel Dilation and Semantic-Guided BEV Dilation blocks, improving detection accuracy and robustness.

Findings

Outperforms state-of-the-art on nuScenes benchmark.

Enhances robustness to depth noise.

Maintains competitive computational efficiency.

Abstract

Integrating LiDAR and camera information in the bird's eye view (BEV) representation has demonstrated its effectiveness in 3D object detection. However, because of the fundamental disparity in geometric accuracy between these sensors, indiscriminate fusion in previous methods often leads to degraded performance. In this paper, we propose BEVDilation, a novel LiDAR-centric framework that prioritizes LiDAR information in the fusion. By formulating image BEV features as implicit guidance rather than naive concatenation, our strategy effectively alleviates the spatial misalignment caused by image depth estimation errors. Furthermore, the image guidance can effectively help the LiDAR-centric paradigm to address the sparsity and semantic limitations of point clouds. Specifically, we propose a Sparse Voxel Dilation Block that mitigates the inherent point sparsity by densifying foreground voxels through image priors. Moreover, we introduce a Semantic-Guided BEV Dilation Block to enhance the LiDAR feature diffusion processing with image semantic guidance and long-range context capture. On the challenging nuScenes benchmark, BEVDilation achieves better performance than state-of-the-art methods while maintaining competitive computational efficiency. Importantly, our LiDAR-centric strategy demonstrates greater robustness to depth noise compared to naive fusion. The source code is available at https://github.com/gwenzhang/BEVDilation.