This repository contains the accompanying code for "CostDCNet: Cost Volume based Depth Completion for a Single RGB-D Image, ECCV'22"
Successful depth completion from a single RGB-D image requires both extracting plentiful 2D and 3D features and merging these heterogeneous features appropriately. We propose a novel depth completion framework, CostDCNet, based on the cost volume-based depth estimation approach that has been successfully employed for multi-view stereo (MVS). The key to high-quality depth map estimation in the approach is constructing an accurate cost volume. To produce a quality cost volume tailored to single-view depth completion, we present a simple but effective architecture that can fully exploit the 3D information, three options to make an RGB-D feature volume, and a per-plane pixel shuffle for efficient volume upsampling. Our framework consists of lightweight (~1.8M parameters) deep neural networks, running in real time (~30ms). Nevertheless, thanks to our simple but effective design, CostDCNet demonstrates depth completion results comparable to or better than the state-of-the-art (SOTA) methods.
- Ubuntu 18.06 or higher
- CUDA 11.1 or higher
- pytorch 1.8 or higher
- python 3.8 or higher
We recommend using Anaconda
conda create -n costDCNet python==3.8.12
conda activate costDCNet
conda install openblas-devel -c anaconda
conda install pytorch==1.8.0 torchvision==0.9.0 torchaudio==0.8.0 cudatoolkit=11.1 -c pytorch -c conda-forge
pip install -U git+https://github.com/NVIDIA/MinkowskiEngine -v --no-deps --install-option="--blas_include_dirs=${CONDA_PREFIX}/include" --install-option="--blas=openblas"
pip install -r requirements.txt
We used preprocessed NYUv2 dataset like NLSPN.
python eval_nyu.py --data_path PATH_TO_NYUv2
This software is being made available under the terms in the LICENSE file.
Any exemptions to these terms requires a license from the Pohang University of Science and Technology.
@inproceedings{kam2022costdcnet,
title={CostDCNet: Cost Volume based Depth Completion for a Single RGB-D Image},
author={Jaewon Kam, Jungeon Kim, Soongjin Kim, Jaesik Park and Seungyong Lee},
booktitle={ECCV},
year={2022}
}
NOTE : Our implementation is based on the repositories as: