Weakly But Deeply Supervised Occlusion-Reasoned Parametric Road Layouts

TL;DR

This paper introduces an end-to-end neural network that infers occlusion-aware road layouts in both perspective and bird's-eye views using minimal human annotations, achieving state-of-the-art results.

Contribution

It presents a novel deep learning approach that requires only parametric annotations for supervision, reducing labeling effort compared to dense semantic labels.

Findings

Achieves state-of-the-art performance on KITTI and NuScenes datasets.

Uses minimal human supervision with parametric annotations.

Incorporates inductive biases for occlusion reasoning and geometric transformations.

Abstract

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.