Robust Consistent Video Based on Monodepth Estimation From Single Image

This repository implements Robust Consistent Video Based on Monodepth Estimation From Single Image.

Report

Colab

Installation

Please refer to the colab notebook to install the dependencies.

Execution

All the necessary steps are in the colab for sample demo.

Result

From top to the bottom are initial input video, initial depth estimation, final depth estimation based on our method.

Acknowledgements

Thank the authors for releasing AdelaiDepth, BoostingDepth, Robust Consistent Video Depth Estimation, Consistent Video Depth Estimation, PyTorch, Ceres Solver, OpenCV, Eigen, MiDaS, RAFT, and detectron2.