DS-Depth: Dynamic and Static Depth Estimation via a Fusion Cost Volume

TL;DR

DS-Depth introduces a fusion of static and dynamic cost volumes using residual optical flow to improve monocular depth estimation in dynamic scenes, significantly enhancing accuracy over previous methods.

Contribution

The paper proposes a novel fusion approach combining static and dynamic cost volumes with residual optical flow for better depth estimation in dynamic environments.

Findings

Outperforms previous methods on KITTI and Cityscapes datasets.

Effectively handles occlusions and moving objects in depth estimation.

Reduces photometric errors using pyramid distillation and adaptive loss functions.

Abstract

Self-supervised monocular depth estimation methods typically rely on the reprojection error to capture geometric relationships between successive frames in static environments. However, this assumption does not hold in dynamic objects in scenarios, leading to errors during the view synthesis stage, such as feature mismatch and occlusion, which can significantly reduce the accuracy of the generated depth maps. To address this problem, we propose a novel dynamic cost volume that exploits residual optical flow to describe moving objects, improving incorrectly occluded regions in static cost volumes used in previous work. Nevertheless, the dynamic cost volume inevitably generates extra occlusions and noise, thus we alleviate this by designing a fusion module that makes static and dynamic cost volumes compensate for each other. In other words, occlusion from the static volume is refined by the dynamic volume, and incorrect information from the dynamic volume is eliminated by the static volume. Furthermore, we propose a pyramid distillation loss to reduce photometric error inaccuracy at low resolutions and an adaptive photometric error loss to alleviate the flow direction of the large gradient in the occlusion regions. We conducted extensive experiments on the KITTI and Cityscapes datasets, and the results demonstrate that our model outperforms previously published baselines for self-supervised monocular depth estimation.