DualCross: Cross-Modality Cross-Domain Adaptation for Monocular BEV Perception
Abstract
Closing the domain gap between training and deployment and incorporating multiple sensor modalities are two challenging yet critical topics for self-driving. Existing work only focuses on single one of the above topics, overlooking the simultaneous domain and modality shift which pervasively exists in real-world scenarios. A model trained with multi-sensor data collected in Europe may need to run in Asia with a subset of input sensors available. In this work, we propose DualCross, a cross-modality cross-domain adaptation framework to facilitate the learning of a more robust monocular bird's-eye-view (BEV) perception model, which transfers the point cloud knowledge from a LiDAR sensor in one domain during the training phase to the camera-only testing scenario in a different domain. This work results in the first open analysis of cross-domain cross-sensor perception and adaptation for monocular 3D tasks in the wild. We benchmark our approach on large-scale datasets under a wide range of domain shifts and show state-of-the-art results against various baselines.
- Publication:
-
arXiv e-prints
- Pub Date:
- May 2023
- DOI:
- 10.48550/arXiv.2305.03724
- arXiv:
- arXiv:2305.03724
- Bibcode:
- 2023arXiv230503724M
- Keywords:
-
- Computer Science - Computer Vision and Pattern Recognition;
- Computer Science - Artificial Intelligence;
- Computer Science - Robotics
- E-Print:
- IROS 2023. Project website: https://yunzeman.github.io/DualCross