MapGS: Generalizable Pretraining and Data Augmentation for Online Mapping via Novel View Synthesis

TL;DR

MapGS introduces a novel view synthesis-based framework using Gaussian splatting to improve online mapping for autonomous vehicles, enhancing generalization across sensor configurations and reducing data labeling efforts.

Contribution

The paper presents a new framework leveraging Gaussian splatting for scene reconstruction and view synthesis, enabling better sensor configuration generalization and data augmentation in online mapping.

Findings

Achieves 18% performance improvement with dataset augmentation

Faster convergence and efficient training compared to previous methods

Exceeds state-of-the-art with only 25% of original training data

Abstract

Online mapping reduces the reliance of autonomous vehicles on high-definition (HD) maps, significantly enhancing scalability. However, recent advancements often overlook cross-sensor configuration generalization, leading to performance degradation when models are deployed on vehicles with different camera intrinsics and extrinsics. With the rapid evolution of novel view synthesis methods, we investigate the extent to which these techniques can be leveraged to address the sensor configuration generalization challenge. We propose a novel framework leveraging Gaussian splatting to reconstruct scenes and render camera images in target sensor configurations. The target config sensor data, along with labels mapped to the target config, are used to train online mapping models. Our proposed framework on the nuScenes and Argoverse 2 datasets demonstrates a performance improvement of 18% through effective dataset augmentation, achieves faster convergence and efficient training, and exceeds state-of-the-art performance when using only 25% of the original training data. This enables data reuse and reduces the need for laborious data labeling. Project page at https://henryzhangzhy.github.io/mapgs.