Robust Consistent Video Depth Estimation

1 Facebook
2Virginia Tech



Virginia Tech

Code (Coming Soon)
Data (Coming Soon)


We present an algorithm for estimating consistent dense depth maps and camera poses from a monocular video. We integrate a learning-based depth prior, in the form of a convolutional neural network trained for single-image depth estimation, with geometric optimization, to estimate a smooth camera trajectory as well as detailed and stable depth reconstruction.

Our algorithm combines two complementary techniques: (1) flexible deformation-splines for low-frequency large-scale alignment and (2) geometry-aware depth filtering for high-frequency alignment of fine depth details. In contrast to prior approaches, our method does not require camera poses as input and achieves robust reconstruction for challenging hand-held cell phone captures that contain a significant amount of noise, shake, motion blur, and rolling shutter deformations. Our method quantitatively outperforms state-of-the-arts on the Sintel benchmark for both depth and pose estimations, and attains favorable qualitative results across diverse wild datasets.

Joint depth and pose optimization

Various configurations of our algorithm: (a-b) Ground truth depth with ground truth and estimated poses, respectively. (c) Misalignments in estimated depth impose jittery errors on the optimized camera trajectories. (d) CVD-style fine-tuning fails in the absence of precise poses. (e) Our flexible deformations resolve depth misalignments, which results in smoother camera trajectories. (f) Using geometry-aware depth filtering we resolve fine depth details (our final result).


 title={Robust Consistent Video Depth Estimation},
 author={Kopf, Johannes and Rong, Xuejian and Huang, Jia-Bin},
 journal={arXiv preprint arXiv:2012.05901},

Related Work

Consistent Video Depth Estimation