Manydepth2: Motion-Aware Self-Supervised Monocular Depth Estimation in Dynamic Scenes

TL;DR

Manydepth2 introduces a motion-aware self-supervised monocular depth estimation method that effectively handles dynamic scenes by integrating optical flow, coarse depth, and attention mechanisms, achieving improved accuracy with computational efficiency.

Contribution

It presents a novel approach combining optical flow, pseudo-static reference frames, and attention-based networks for dynamic scene depth estimation.

Findings

Achieves ~5% reduction in RMSE on KITTI-2015 dataset.

Effectively estimates depth in dynamic scenes.

Maintains computational efficiency comparable to existing methods.

Abstract

Despite advancements in self-supervised monocular depth estimation, challenges persist in dynamic scenarios due to the dependence on assumptions about a static world. In this paper, we present Manydepth2, to achieve precise depth estimation for both dynamic objects and static backgrounds, all while maintaining computational efficiency. To tackle the challenges posed by dynamic content, we incorporate optical flow and coarse monocular depth to create a pseudo-static reference frame. This frame is then utilized to build a motion-aware cost volume in collaboration with the vanilla target frame. Furthermore, to improve the accuracy and robustness of the network architecture, we propose an attention-based depth network that effectively integrates information from feature maps at different resolutions by incorporating both channel and non-local attention mechanisms. Compared to methods with similar computational costs, Manydepth2 achieves a significant reduction of approximately five percent in root-mean-square error for self-supervised monocular depth estimation on the KITTI-2015 dataset. The code could be found at https://github.com/kaichen-z/Manydepth2.