Exploiting Playbacks in Unsupervised Domain Adaptation for 3D Object Detection
This addresses the critical problem of domain adaptation for 3D object detection in autonomous vehicles, which is essential for safe operation in new environments, though it appears to be an incremental improvement on existing pseudo-labeling techniques.
The paper tackles the problem of 3D object detectors overfitting to domain idiosyncrasies in autonomous driving by proposing a novel learning approach that fine-tunes detectors on pseudo-labels generated from replays of previously recorded driving sequences. The method reduces the domain gap to new driving environments, yielding drastic improvements in accuracy and detection reliability across five autonomous driving datasets.
Self-driving cars must detect other vehicles and pedestrians in 3D to plan safe routes and avoid collisions. State-of-the-art 3D object detectors, based on deep learning, have shown promising accuracy but are prone to over-fit to domain idiosyncrasies, making them fail in new environments -- a serious problem if autonomous vehicles are meant to operate freely. In this paper, we propose a novel learning approach that drastically reduces this gap by fine-tuning the detector on pseudo-labels in the target domain, which our method generates while the vehicle is parked, based on replays of previously recorded driving sequences. In these replays, objects are tracked over time, and detections are interpolated and extrapolated -- crucially, leveraging future information to catch hard cases. We show, on five autonomous driving datasets, that fine-tuning the object detector on these pseudo-labels substantially reduces the domain gap to new driving environments, yielding drastic improvements in accuracy and detection reliability.