PanoDepth

A Two Stage Approach for Monocular Omnidirectional Depth Estimation

Abstract

Omnidirectional 3D information is essential for a wide range of applications such as Virtual Reality, Autonomous Driving, Robotics, etc. In this paper, we propose a novel, model-agnostic, two-stage pipeline for omnidirectional monocular depth estimation. Our proposed framework PanoDepth takes one 360 image as input, produces one or more synthesized views in the first stage, and feeds the original image and the synthesized images into the subsequent stereo matching stage. In the second stage, we propose a differentiable Spherical Warping Layer to handle omnidirectional stereo geometry efficiently and effectively. By utilizing the explicit stereo-based geometric constraints in the stereo matching stage, PanoDepth can generate dense high-quality depth. We conducted extensive experiments and ablation studies to evaluate PanoDepth with both the full pipeline as well as the individual modules in each stage. Our results show that PanoDepth outperforms the state-of-the-art approaches by a large margin for 360 monocular depth estimation.

General diagram of PanoDepth.
Fig. Illustration of our PanoDepth framework. PanoDepth takes one 360 image as input to generate one or more novel views in the first view synthesis stage. The original and synthesized 360 images are then fed into the second multi-view stereo matching stage to predict final dense depth map.
Result comparison with other methods.
Fig. A qualitative comparison between our method (3rd column) BiFuse (4th column), and OmniDepth (5th column) on 360D. We highlight and zoom in some areas that distinguish the performance of three methods. We can see that our PanoDepth is able to produce sharp edges, predict depth range accurately, and recover surface detail.
PanoDepth Result Example in 3D.
PanoDepth Result Example in 3D.
PanoDepth Result Example in 3D.
PanoDepth Result Example in 3D.
PanoDepth Result Example in 3D.
PanoDepth Result Example in 3D.
Fig. 3D point cloud reconstructed from estimated depth using PanoDepth.

Cite

@inproceedings{li2021panodepth,
    author={Li, Yuyan and Yan, Zhixin and Duan, Ye and Ren, Liu},
    booktitle={2021 International Conference on 3D Vision (3DV)}, 
    title={PanoDepth: A Two-Stage Approach for Monocular Omnidirectional Depth Estimation}, 
    year={2021},
    volume={},
    number={},
    pages={648-658},
    doi={10.1109/3DV53792.2021.00074}
}