Search papers, labs, and topics across Lattice.
This paper addresses the performance degradation of radar-camera BEV perception across different datasets by modeling source-domain variations to enhance the robustness of 3D detectors. The authors introduce a novel framework that analyzes visual scene variations in the frequency domain, synthesizing diverse source-domain views to regularize the detector's training process. Experimental results show significant improvements in detection performance across multiple BEV fusion backbones, even when limited target-domain data is available, highlighting the effectiveness of the proposed method in maintaining stability under scene variations.
By leveraging frequency domain analysis, this approach significantly enhances the robustness of 3D perception systems against diverse driving conditions without needing target-domain samples.
Radar-camera BEV perception often suffers from degraded performance when evaluated across datasets, as changes in driving scenes, sensor configurations, and environmental conditions can alter both the input observations and the internal fused representations. This work studies this issue from the perspective of source-domain variation modeling, aiming to improve the robustness of BEV-based 3D detectors without relying on target-domain samples. We introduce a framework that characterizes visual scene variations in the frequency domain and uses them to synthesize diverse source-domain views. By comparing the resulting fused BEV representations, the framework further captures how image-level variations influence multi-modal BEV features. These variation patterns are then used to regularize the detector, encouraging the learned fusion space to remain stable under latent scene changes. The proposed method is applied only during training and leaves the inference pipeline unchanged. Experiments on cross-dataset radar-camera 3D detection between View-of-Delft and TJ4DRadSet demonstrate consistent improvements over multiple BEV fusion backbones, and the gains remain effective when a small amount of target-domain data is available.
