Terrain-Enhanced Resolution-aware Refinement Attention for Off-Road Segmentation

TL;DR

This paper presents a resolution-aware token decoder for off-road segmentation that balances global and local features efficiently, improving boundary accuracy and robustness with minimal computational overhead.

Contribution

The proposed method introduces a novel resolution-aware token decoder with integrated boundary regularization, enhancing off-road segmentation accuracy and stability over existing approaches.

Findings

Achieves competitive segmentation performance on off-road datasets.

Improves boundary accuracy and local coherence in segmentation results.

Maintains low computational cost with high-resolution detail injection.

Abstract

Off-road semantic segmentation suffers from thick, inconsistent boundaries, sparse supervision for rare classes, and pervasive label noise. Designs that fuse only at low resolution blur edges and propagate local errors, whereas maintaining high-resolution pathways or repeating high-resolution fusions is costly and fragile to noise. We introduce a resolutionaware token decoder that balances global semantics, local consistency, and boundary fidelity under imperfect supervision. Most computation occurs at a low-resolution bottleneck; a gated cross-attention injects fine-scale detail, and only a sparse, uncertainty-selected set of pixels is refined. The components are co-designed and tightly integrated: global self-attention with lightweight dilated depthwise refinement restores local coherence; a gated cross-attention integrates fine-scale features from a standard high-resolution encoder stream without amplifying noise; and a class-aware point refinement corrects residual ambiguities with negligible overhead. During training, we add a boundary-band consistency regularizer that encourages coherent predictions in a thin neighborhood around annotated edges, with no inference-time cost. Overall, the results indicate competitive performance and improved stability across transitions.