PivotNet: Vectorized Pivot Learning for End-to-end HD Map Construction

TL;DR

PivotNet introduces a novel vectorized approach for high-definition map construction in autonomous driving, utilizing pivot-based representations and sequence matching to improve accuracy and detail over existing methods.

Contribution

It proposes a unified pivot-based map representation and a set prediction framework with novel modules for topology modeling and dynamic sequence loss.

Findings

PivotNet outperforms state-of-the-art methods by at least 5.9 mAP.

The point-to-line mask module effectively encodes geometric priors.

Extensive experiments validate the superiority of the proposed architecture.

Abstract

Vectorized high-definition map online construction has garnered considerable attention in the field of autonomous driving research. Most existing approaches model changeable map elements using a fixed number of points, or predict local maps in a two-stage autoregressive manner, which may miss essential details and lead to error accumulation. Towards precise map element learning, we propose a simple yet effective architecture named PivotNet, which adopts unified pivot-based map representations and is formulated as a direct set prediction paradigm. Concretely, we first propose a novel point-to-line mask module to encode both the subordinate and geometrical point-line priors in the network. Then, a well-designed pivot dynamic matching module is proposed to model the topology in dynamic point sequences by introducing the concept of sequence matching. Furthermore, to supervise the position and topology of the vectorized point predictions, we propose a dynamic vectorized sequence loss. Extensive experiments and ablations show that PivotNet is remarkably superior to other SOTAs by 5.9 mAP at least. The code will be available soon.