MemFusionMap: Working Memory Fusion for Online Vectorized HD Map Construction

WACV 2025

Jingyu Song1, Xudong Chen2, Liupei Lu2, Jie Li2, Katherine A. Skinner1
1University of Michigan 2NVIDIA
Paper arXiv Video Code (Pre-release)
CRKD

We proposed MemFusionMap, a novel approach for effective online vectorized HD map construction with enhanced temporal reasoning capability.

Abstract

High-definition (HD) maps provide environmental information for autonomous driving systems and are essential for safe planning. While existing methods with single-frame input achieve impressive performance for online vectorized HD map construction, they still struggle with complex scenarios and occlusions. We propose MemFusionMap, a novel temporal fusion model with enhanced temporal reasoning capabilities for online HD map construction. Specifically, we contribute a working memory fusion module that improves the model's memory capacity to reason across a history of frames. We also design a novel temporal overlap heatmap to explicitly inform the model about the temporal overlap information and vehicle trajectory in the Bird's Eye View space. By integrating these two designs, MemFusionMap significantly outperforms existing methods while also maintaining a versatile design for scalability. We conduct extensive evaluation on open-source benchmarks and demonstrate a maximum improvement of 5.4% in mAP over state-of-the-art methods.

Overview

We propose a simple yet effective model to fuse working memory features in BEV space for online vectorized HD map construction. MemFusionMap focuses on improving the network's temporal reasoning capability while also maintaining a versatile design for scalability and compatibility.

We propose a novel design of maintaining a temporal overlap heatmap, providing a strong cue for the model to reason across a history of frames and also implicitly encoding valuable insights of the vehicle's trajectory.

We conduct extensive evaluation on nuScenes and Argoverse2 to demonstrate the effectiveness of MemFusionMap. The proposed method significantly outperforms the state-of-the-art method, achieving a maximum improvement of 5.4% in mAP.

MemFusionMap Overview

Results

Qualitative results on nuScenes at the 100x50 m perception ranges. We show a consecutive sequence of 5 frames.

nuScenes Result 1 nuScenes Result 2

BibTeX

@InProceedings{song2025memfusionmap,
      author    = {Song, Jingyu and Chen, Xudong and Lu, Liupei and Li, Jie and Skinner, Katherine A.},
      title     = {MemFusionMap: Working Memory Fusion for Online Vectorized HD Map Construction},
      booktitle = {Proceedings of the Winter Conference on Applications of Computer Vision (WACV)},
      month     = {February},
      year      = {2025},
      pages     = {9230-9239}
    }