Summary of Enhancing Vectorized Map Perception with Historical Rasterized Maps, by Xiaoyu Zhang et al.
Enhancing Vectorized Map Perception with Historical Rasterized Maps
by Xiaoyu Zhang, Guangwei Liu, Zihao Liu, Ningyi Xu, Yunhui Liu, Ji Zhao
First submitted to arxiv on: 1 Sep 2024
Categories
- Main: Computer Vision and Pattern Recognition (cs.CV)
- Secondary: Artificial Intelligence (cs.AI)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The proposed HRMapNet model leverages a low-cost Historical Rasterized Map to enhance online vectorized map perception in bird’s-eye-view (BEV) space, aiming to replace traditional high-cost offline high-definition (HD) maps. The authors propose two novel modules: feature aggregation and query initialization, which exploit the historical map’s complementary information. HRMapNet is integrated with state-of-the-art methods on nuScenes and Argoverse 2 datasets, significantly improving their performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper proposes a new way to improve online map perception in autonomous driving. It uses a special kind of map that combines past predictions to help the system better understand its surroundings. This helps when there are obstacles or bad weather making it hard for the system to rely only on what it sees around it. The authors test their idea with two popular methods and show that it works really well. |
