Skip to yearly menu bar Skip to main content


Poster

ReSimAD: Zero-Shot 3D Domain Transfer for Autonomous Driving with Source Reconstruction and Target Simulation

Bo Zhang · Xinyu Cai · Jiakang Yuan · Donglin Yang · Jianfei Guo · Xiangchao Yan · Renqiu Xia · Botian Shi · Min Dou · Tao Chen · Si Liu · Junchi Yan · Yu Qiao

Halle B
[ ]
Wed 8 May 1:45 a.m. PDT — 3:45 a.m. PDT

Abstract:

Domain shifts such as sensor type changes and geographical situation variations are prevalent in Autonomous Driving (AD), which poses a challenge since AD model relying on the previous domain knowledge can be hardly directly deployed to a new domain without additional costs. In this paper, we provide a new perspective and approach of alleviating the domain shifts, by proposing a Reconstruction-Simulation-Perception (ReSimAD) scheme. Specifically, the implicit reconstruction process is based on the knowledge from the previous old domain, aiming to convert the domain-related knowledge into domain-invariant representations, e.g., 3D scene-level meshes. Besides, the point clouds simulation process of multiple new domains is conditioned on the above reconstructed 3D meshes, where the target-domain-like simulation samples can be obtained, thus reducing the cost of collecting and annotating new-domain data for the subsequent perception process. For experiments, we consider different cross-domain situations such as Waymo-to-KITTI, Waymo-to-nuScenes, etc, to verify the zero-shot target-domain perception using ReSimAD. Results demonstrate that our method is beneficial to boost the domain generalization ability, even promising for 3D pre-training. Code and simulated points are available at: https://github.com/PJLab-ADG/3DTrans

Chat is not available.