Liu, Rong (Beijing University of Posts and Telecommunications), Deng, Jiayin (Beijing University of Posts and Telecommunications), ZHU, BONING (BUPT), Hu, Zhiqun (Beijing University of Posts and Telecommunications), Lu, Zhaoming (Beijing University of Posts and Telecommunications)

DiffRCF: Diffusion Model for Robust 3D Object Detection with Radar-Camera Fusion

Scheduled for presentation during the Regular Session "S28c-Multi-Sensor Fusion and Perception for Robust Autonomous Driving" (TH-LA-T28), Thursday, November 20, 2025, 16:00−16:20, Stradbroke

2025 IEEE 28th International Conference on Intelligent Transportation Systems (ITSC), November 18-21, 2025, Gold Coast, Australia

This information is tentative and subject to change. Compiled on October 18, 2025

Abstract

Three-dimensional object detection is a critical task in autonomous driving.Although recent radar-camera fusion methods have achieved promising results in 3D detection, including under challenging conditions such as low illumination or adverse weather, they overlook the problem of sensor data loss and fail to fully exploit the correlations between different modal features. In this paper, we propose DiffRCF, a novel radar-camera fusion framework. Specifically, we leverage sparse yet accurate radar points to enhance perspective image features and transform them into bird's-eye-view (BEV) representations. DiffRCF integrates a modality-aware weighting mechanism to adaptively assess the importance of each modality under varying conditions and a conditional diffusion model to reconstruct missing information. Additionally, we employ deformable cross-attention and spatial attention mechanisms to better align and fuse multi-modal features. Experiments on the nuScenes dataset demonstrate that DiffRCF achieves state-of-the-art performance among single-frame radar-camera fusion methods and exhibits strong robustness against poor lighting and sensor degradation.