Abstract
In the field of autonomous driving, self-training is widely applied to mitigate distribution shifts in LiDAR-based 3D object detectors. This eliminates the need for expensive, high-quality labels whenever the environment changes (e.g., geographic location, sensor setup, weather condition). State-of-the-art self-training approaches, however, mostly ignore the temporal nature of autonomous driving data. To address this issue, we propose a flow-aware self-training method that enables unsupervised domain adaptation for 3D object detectors on continuous LiDAR point clouds. In order to get reliable pseudo-labels, we leverage scene flow to propagate detections through time. In particular, we introduce a flow-based multi-target tracker, that exploits flow consistency to filter and refine resulting tracks. The emerged precise pseudo-labels then serve as a basis for model re-training. Starting with a pre-trained KITTI model, we conduct experiments on the challenging Waymo Open Dataset to demonstrate the effectiveness of our approach. Without any prior target domain knowledge, our results show a significant improvement over the state-of-the-art.
Original language | English |
---|---|
Title of host publication | British Machine Vision Conference (BMVC) 2021 |
Publisher | The British Machine Vision Association |
Number of pages | 14 |
Publication status | Published - 23 Nov 2021 |
Event | 32nd British Machine Vision Conference: BMVC 2021 - Virtuell, United Kingdom Duration: 22 Nov 2021 → 25 Nov 2021 |
Conference
Conference | 32nd British Machine Vision Conference |
---|---|
Abbreviated title | BMVC 2021 |
Country/Territory | United Kingdom |
City | Virtuell |
Period | 22/11/21 → 25/11/21 |