Learning Structurally Consistent Representations for Multi-View Radar Semantic Segmentation

2026-06-30

AI Summary

This work addresses the challenge of dense semantic segmentation in adverse conditions, where radar data suffer from sparsity, high noise levels, and weak semantic content. To tackle this, the authors propose a unified higher-order structural alignment framework that, for the first time, integrates learnable hypergraphs with unbalanced optimal transport (UOT) into multi-view radar segmentation. This approach explicitly models higher-order dependencies among multi-echo signals across range–azimuth (RA), range–Doppler (RD), and azimuth–Doppler (AD) views and achieves cross-view feature alignment without requiring point-wise correspondences. Coupled with adaptive attention-based fusion and structural consistency regularization, the method significantly enhances robustness. It achieves state-of-the-art performance on the CARRADA and RADIal benchmarks with mIoU scores of 63.8% and 83.4%, respectively—improving upon existing best methods by 1.7 and 2.3 mIoU—and thereby validates the efficacy of higher-order relational modeling.

Abstract

Radar sensors provide reliable perception under adverse weather and lighting conditions, but their sparse, noisy, and weakly semantic measurements make dense semantic segmentation challenging. Most existing radar segmentation methods rely on grid-based encodings and pairwise interactions, which struggle to capture the higher-order relational structure formed by multiple radar returns from the same physical object. We introduce a unified higher-order structural alignment framework for multi-view radar segmentation. The proposed method refines radar feature representations using learnable hypergraphs to capture higher-order dependencies among spatially related responses. To ensure consistency across heterogeneous radar projections, we further align view-specific features using Unbalanced Optimal Transport (UOT), enabling correspondence-free alignment under varying measurement densities and partial observations. An adaptive attention mechanism then fuses complementary radar views while emphasising structurally informative responses under sparsity and noise. The resulting architecture learns structurally consistent representations across Range Angle (RA), Range Doppler (RD), and Angle Doppler (AD) views and is trained using supervised segmentation together with cross-view consistency regularisation. Experiments on the CARRADA and RADIal benchmarks demonstrate consistent improvements over strong radar-specific baselines, achieving 63.8% mIoU on CARRADA and 83.4% mIoU on RADIal, improving the previous best methods by +1.7 and +2.3 mIoU, respectively. These results highlight the importance of higher-order relational modelling for robust radar perception.

Problem

• radar semantic segmentation
• multi-view
• higher-order relations
• structural consistency
• sparse measurements

Innovation

• higher-order structural alignment
• learnable hypergraphs
• Unbalanced Optimal Transport
• multi-view radar segmentation
• adaptive attention mechanism