"The Pedestrian next to the Lamppost" Adaptive Object Graphs for Better Instantaneous Mapping

TL;DR

This paper introduces an adaptive object graph approach using graph neural networks to improve bird's-eye-view mapping from monocular images, significantly reducing localization errors especially for distant objects.

Contribution

It proposes a novel graph neural network that models spatial relationships between objects to enhance BEV map accuracy from single images.

Findings

Achieves state-of-the-art BEV estimation accuracy across three datasets.

50% relative improvement in object localization on nuScenes.

Reduces error increase with distance in BEV mapping.

Abstract

Estimating a semantically segmented bird's-eye-view (BEV) map from a single image has become a popular technique for autonomous control and navigation. However, they show an increase in localization error with distance from the camera. While such an increase in error is entirely expected - localization is harder at distance - much of the drop in performance can be attributed to the cues used by current texture-based models, in particular, they make heavy use of object-ground intersections (such as shadows), which become increasingly sparse and uncertain for distant objects. In this work, we address these shortcomings in BEV-mapping by learning the spatial relationship between objects in a scene. We propose a graph neural network which predicts BEV objects from a monocular image by spatially reasoning about an object within the context of other objects. Our approach sets a new state-of-the-art in BEV estimation from monocular images across three large-scale datasets, including a 50% relative improvement for objects on nuScenes.