Faster Training, Fewer Labels: Self-Supervised Pretraining for Fine-Grained BEV Segmentation
This work addresses the annotation bottleneck for fine-grained BEV segmentation in autonomous driving, offering a scalable method to reduce labeling effort and training time, though it is incremental as it builds on existing models like BEVFormer and Mask2Former.
The paper tackled the problem of costly and inconsistent BEV ground truth annotations for autonomous driving by introducing a two-phase training strategy that removes full supervision during pretraining and halves training data during fine-tuning, resulting in up to +2.5pp mIoU improvement over the supervised baseline while reducing training time by up to two-thirds.
Dense Bird's Eye View (BEV) semantic maps are central to autonomous driving, yet current multi-camera methods depend on costly, inconsistently annotated BEV ground truth. We address this limitation with a two-phase training strategy for fine-grained road marking segmentation that removes full supervision during pretraining and halves the amount of training data during fine-tuning while still outperforming the comparable supervised baseline model. During the self-supervised pretraining, BEVFormer predictions are differentiably reprojected into the image plane and trained against multi-view semantic pseudo-labels generated by the widely used semantic segmentation model Mask2Former. A temporal loss encourages consistency across frames. The subsequent supervised fine-tuning phase requires only 50% of the dataset and significantly less training time. With our method, the fine-tuning benefits from rich priors learned during pretraining boosting the performance and BEV segmentation quality (up to +2.5pp mIoU over the fully supervised baseline) on nuScenes. It simultaneously halves the usage of annotation data and reduces total training time by up to two thirds. The results demonstrate that differentiable reprojection plus camera perspective pseudo labels yields transferable BEV features and a scalable path toward reduced-label autonomous perception.