BezierFormer: A Unified Architecture for 2D and 3D Lane Detection
Zhiwei Dong, Xi Zhu, Xiya Cao, Ran Ding, Wei Li, Caifa Zhou, Yongliang Wang, Qiangbo Liu
Unverified — Be the first to reproduce this paper.
ReproduceAbstract
Lane detection has made significant progress in recent years, but there is not a unified architecture for its two sub-tasks: 2D lane detection and 3D lane detection. To fill this gap, we introduce B\'ezierFormer, a unified 2D and 3D lane detection architecture based on B\'ezier curve lane representation. B\'ezierFormer formulate queries as B\'ezier control points and incorporate a novel B\'ezier curve attention mechanism. This attention mechanism enables comprehensive and accurate feature extraction for slender lane curves via sampling and fusing multiple reference points on each curve. In addition, we propose a novel Chamfer IoU-based loss which is more suitable for the B\'ezier control points regression. The state-of-the-art performance of B\'ezierFormer on widely-used 2D and 3D lane detection benchmarks verifies its effectiveness and suggests the worthiness of further exploration.