Weakly But Deeply Supervised Occlusion-Reasoned Parametric Road Layouts
This addresses the need for efficient road scene understanding in autonomous driving by reducing annotation costs, though it is incremental in using weakly supervised methods.
The paper tackles the problem of generating occlusion-reasoned road layouts from single RGB images with minimal human supervision, achieving state-of-the-art results on KITTI and NuScenes datasets.
We propose an end-to-end network that takes a single perspective RGB image of a complex road scene as input, to produce occlusion-reasoned layouts in perspective space as well as a parametric bird's-eye-view (BEV) space. In contrast to prior works that require dense supervision such as semantic labels in perspective view, our method only requires human annotations for parametric attributes that are cheaper and less ambiguous to obtain. To solve this challenging task, our design is comprised of modules that incorporate inductive biases to learn occlusion-reasoning, geometric transformation and semantic abstraction, where each module may be supervised by appropriately transforming the parametric annotations. We demonstrate how our design choices and proposed deep supervision help achieve meaningful representations and accurate predictions. We validate our approach on two public datasets, KITTI and NuScenes, to achieve state-of-the-art results with considerably less human supervision.