MachMap: End-to-End Vectorized Solution for Compact HD-Map Construction

TL;DR

MachMap is an end-to-end vectorized framework for high-definition map construction in autonomous driving, achieving state-of-the-art accuracy by innovative vectorization and map-compaction techniques.

Contribution

The paper introduces MachMap, a novel end-to-end architecture with a map-compaction scheme that significantly reduces vector points and improves HD-map construction accuracy.

Findings

Achieves 83.5 mAP on Argoverse2 benchmark.

Reduces vectorized points by 93% without performance loss.

Outperforms all other online HD-map construction methods.

Abstract

This report introduces the 1st place winning solution for the Autonomous Driving Challenge 2023 - Online HD-map Construction. By delving into the vectorization pipeline, we elaborate an effective architecture, termed as MachMap, which formulates the task of HD-map construction as the point detection paradigm in the bird-eye-view space with an end-to-end manner. Firstly, we introduce a novel map-compaction scheme into our framework, leading to reducing the number of vectorized points by 93% without any expression performance degradation. Build upon the above process, we then follow the general query-based paradigm and propose a strong baseline with integrating a powerful CNN-based backbone like InternImage, a temporal-based instance decoder and a well-designed point-mask coupling head. Additionally, an extra optional ensemble stage is utilized to refine model predictions for better performance. Our MachMap-tiny with IN-1K initialization achieves a mAP of 79.1 on the Argoverse2 benchmark and the further improved MachMap-huge reaches the best mAP of 83.5, outperforming all the other online HD-map construction approaches on the final leaderboard with a distinct performance margin (> 9.8 mAP at least).