Home

Awesome

<div align="center"> <h1>HRMapNet</h1> <h2>Enhancing Vectorized Map Perception with Historical Rasterized Maps</h2>

Xiaoyu Zhang<sup>1*</sup>, Guangwei Liu<sup>2*</sup>, Zihao Liu<sup>3</sup>, Ningyi Xu<sup>3</sup>, Yunhui Liu<sup>1</sup><sup>:envelope:</sup>, Ji Zhao<sup>2#</sup>,

<sup>*</sup>Equal contribution. <sup>:envelope:</sup>Corresponding author. <sup>#</sup>Project lead

<sup>1</sup> The Chinese University of Hong Kong, <sup>2</sup> Huixi Technology, <sup>3</sup> Shanghai Jiao Tong University

ArXiv Preprint (arXiv 2409.00620)

Accepted by ECCV 2024

</div>

Overview

pipeline This project introduces HRMapNet, leveraging a low-cost Historical Rasterized Map to enhance online vectorized map perception. The historical rasterized map can be easily constructed from past predicted vectorized results and provides valuable complementary information. To fully exploit a historical map, we propose two novel modules to enhance BEV features and map element queries. For BEV features, we employ a feature aggregation module to encode features from both onboard images and the historical map. For map element queries, we design a query initialization module to endow queries with priors from the historical map. The two modules contribute to leveraging map information in online perception. Our HRMapNet can be integrated with most online vectorized map perception methods, significantly improving their performance on both the nuScenes and Argoverse 2 datasets.

Example of online perception from an emplt map

YouTube

<div align="center">

https://github.com/user-attachments/assets/42f7fbf5-9cf9-4032-a4f9-bdc91cfcb5fb

</div>

Models

MapTRv2 as Baseline

nuScenes dataset

MethodEpochAP<sub>div</sub>AP<sub>ped</sub>AP<sub>bou</sub>mAPInitial MapConfigDownload
MapTRv2+<br/>HRMapNet2467.465.668.567.2Emptyconfigmodel
MapTRv2+<br/>HRMapNet2472.173.073.973.0Testing Mapconfigmodel
MapTRv2+<br/>HRMapNet2486.281.083.683.6Training Mapconfigmodel
MapTRv2+<br/>HRMapNet11072.772.275.773.5Emptyconfigmodel

Argoverse 2 dataset

MethodEpochAP<sub>div</sub>AP<sub>ped</sub>AP<sub>bou</sub>mAPInitial MapConfigDownload
MapTRv23068.760.064.264.3-configmodel
MapTRv2+<br/>HRMapNet3071.465.168.668.3Emptyconfigmodel

MapQR as Baseline

Here we also provide results based on our MapQR. This is not included in our paper.

nuScenes dataset

MethodEpochAP<sub>div</sub>AP<sub>ped</sub>AP<sub>bou</sub>mAPInitial MapConfigDownload
MapQR+<br/>HRMapNet2470.170.371.170.5Emptyconfigmodel
MapQR+<br/>HRMapNet*2473.172.272.572.6Emptyconfigmodel

*Fix a bug in MapQR.

Getting Started

These settings are similar with MapTRv2

Acknowledgements

MapQR is mainly based on MapTRv2 and NMP.

It is also greatly inspired by the following outstanding contributions to the open-source community: BEVFormer, MapQR, BoundaryFormer.

Citation

If you find this project is useful in your research or applications, please consider giving us a star 🌟 and citing it by the following BibTeX entry.

@inproceedings{zhang2024hrmapnet,
  title={Enhancing Vectorized Map Perception with Historical Rasterized Maps},
  author={Zhang, Xiaoyu and Liu, Guangwei and Liu, Zihao and Xu, Ningyi and Liu, Yunhui and Zhao, Ji},
  booktitle={European Conference on Computer Vision},
  year={2024}
}