TopoBDA: Towards Bezier Deformable Attention for Road Topology Understanding

TL;DR

TopoBDA introduces a Bezier deformable attention mechanism within a transformer framework to improve road topology understanding from multi-camera imagery, achieving state-of-the-art results in lane detection and topology reasoning.

Contribution

The paper proposes a novel Bezier Deformable Attention module integrated into a transformer for enhanced road topology comprehension from multi-camera data.

Findings

Outperforms existing methods on OpenLane-V2 for centerline detection.

Achieves state-of-the-art results on OpenLane-V1 in 3D lane detection.

Multimodal data integration further improves topology understanding.

Abstract

Understanding road topology is crucial for autonomous driving. This paper introduces TopoBDA (Topology with Bezier Deformable Attention), a novel approach that enhances road topology comprehension by leveraging Bezier Deformable Attention (BDA). TopoBDA processes multi-camera 360-degree imagery to generate Bird's Eye View (BEV) features, which are refined through a transformer decoder employing BDA. BDA utilizes Bezier control points to drive the deformable attention mechanism, improving the detection and representation of elongated and thin polyline structures, such as lane centerlines. Additionally, TopoBDA integrates two auxiliary components: an instance mask formulation loss and a one-to-many set prediction loss strategy, to further refine centerline detection and enhance road topology understanding. Experimental evaluations on the OpenLane-V2 dataset demonstrate that TopoBDA outperforms existing methods, achieving state-of-the-art results in centerline detection and topology reasoning. TopoBDA also achieves the best results on the OpenLane-V1 dataset in 3D lane detection. Further experiments on integrating multi-modal data -- such as LiDAR, radar, and SDMap -- show that multimodal inputs can further enhance performance in road topology understanding.