BRUM: Robust 3D Vechicle Reconstruction from 360° Sparse Images

Installation

1. Clone repository and setup environment. The repository contains submodules, thus please check it out with:

git clone https://github.com/davidedinuc/brum.git --recursive
conda create -n brum python=3.9
conda activate brum

2. Install PyTorch3D.

conda install pytorch==2.0.0 torchvision==0.15.0 torchaudio==2.0.0 pytorch-cuda=11.8 -c pytorch -c nvidia
conda install -c fvcore -c iopath -c conda-forge fvcore iopath
conda install -c bottler nvidiacub
conda install pytorch3d -c pytorch3d
pip install -r requirements.txt

3. Install gs dependencies

pip install gaussians-splatting/submodules/diff-gaussian-rasterization
pip install gaussians-splatting/submodules/simple-knn

4. Install dust3r dependencies.

pip install -r ./dust3r/requirements.txt

Download data

To download out BRUM-dataset, please follow the links below:

Real data
Synthetic data
Blend Files

Usage

#For real data setup ./scripts/run_pipeline_real.sh 
#For synthetic data setup ./scripts/run_pipeline_synthetic.sh 
cd scripts
./run_pipeline_real.sh
./run_pipeline_synthetic.sh