Focus on BEV: Self-calibrated Cycle View Transformation for Monocular Birds-Eye-View Segmentation

TL;DR

This paper introduces FocusBEV, a novel framework for monocular Birds-Eye-View segmentation that enhances view transformation accuracy and temporal consistency, achieving state-of-the-art results on popular benchmarks.

Contribution

The paper presents a self-calibrated view transformation, ego-motion-based temporal fusion, and an occupancy-agnostic loss, advancing BEV segmentation methods.

Findings

Achieves 29.2% mIoU on nuScenes

Achieves 35.2% mIoU on Argoverse

Outperforms previous state-of-the-art methods

Abstract

Birds-Eye-View (BEV) segmentation aims to establish a spatial mapping from the perspective view to the top view and estimate the semantic maps from monocular images. Recent studies have encountered difficulties in view transformation due to the disruption of BEV-agnostic features in image space. To tackle this issue, we propose a novel FocusBEV framework consisting of $(i)$ a self-calibrated cross view transformation module to suppress the BEV-agnostic image areas and focus on the BEV-relevant areas in the view transformation stage, $(ii)$ a plug-and-play ego-motion-based temporal fusion module to exploit the spatiotemporal structure consistency in BEV space with a memory bank, and $(iii)$ an occupancy-agnostic IoU loss to mitigate both semantic and positional uncertainties. Experimental evidence demonstrates that our approach achieves new state-of-the-art on two popular benchmarks,\ie, 29.2\% mIoU on nuScenes and 35.2\% mIoU on Argoverse.