diff --git a/README.md b/README.md index 90e44c2..b2ffd46 100644 --- a/README.md +++ b/README.md @@ -1,4 +1,5 @@ # [ShadowArt-Revisited](https://kaustubh-sadekar.github.io/ShadowArt-Revisited/) + Shadow Art Revisited: A Differentiable Rendering Based Approach ![ShadowArt Teaser](media/teaser.png) @@ -8,51 +9,56 @@ Shadow Art Revisited: A Differentiable Rendering Based Approach **Clone the repository** -```script +```shell git clone https://github.com/kaustubh-sadekar/ShadowArt-Revisited.git cd ShadowArt-Revisited ``` **Install required libraries** -```script +```shell pip install -r requirements.txt -pip plotly +pip install plotly curl -LO https://github.com/NVIDIA/cub/archive/1.10.0.tar.gz tar xzf 1.10.0.tar.gz -python -c "import os; os.environ["CUB_HOME"] = os.getcwd() + "/cub-1.10.0" " +python -c "import os; os.environ["\'CUB_HOME\'"] = os.getcwd() + "\'/cub-1.10.0\'"; " pip install 'git+https://github.com/facebookresearch/pytorch3d.git@stable' +# conda install pytorch3d # alternative method in conda env ``` -We used [PyTorch3D](https://github.com/facebookresearch/pytorch3d) to create the differentiable rendering-based shadowart pipeline. +We used [PyTorch3D](https://github.com/facebookresearch/pytorch3d) to create the differentiable rendering-based shadowart pipeline. ## Examples + ### Create Shadow Art Using Voxel Optimization To create shadow art with two views with files duck.png and mikey.png use the following command. -```script -python val.py cuda:0 output1 600 0.01 -swt 10.0 -l1wt 10.0 -sdlist duck.png mikey.png + +```shell +cd mesh +python val.py cuda:0 output1 600 0.01 0 -swt 10.0 -l1wt 10.0 -sdlist duck.png mikey.png ``` + *For a better understanding of the input arguments type `python val.py -h`* ![Voxel Output](media/voxel_shadow_art.gif) *Gif showing the optimization of the 3D voxel shadow art* - ### Create Shadow Art Using Mesh Optimization To create shadow art with two views with files duck.png and mikey.png use the following command. -```script + +```shell python val.py cuda:0 output1 2000 0.15 0 -swt 1.6 -l1wt 1.6 -mwt 0.0 -i2vwt 0.0 -ewt 1.6 -nwt 0.6 -lwt 1.2 -sdlist duck.png mikey.png ``` + *For a better understanding of the input arguments type `python val.py -h`* ![Mesh Output](media/mesh_shadow_art.gif) *Gif showing the optimization of the 3D shadow art mesh* - ## Citation If you would like to cite us, kindly use the following BibTeX entry. diff --git a/requirements.txt b/requirements.txt index d6520dd..5330775 100644 --- a/requirements.txt +++ b/requirements.txt @@ -4,8 +4,6 @@ numpy==1.19.2 opencv_contrib_python==4.5.1.48 Pillow==9.2.0 scikit_image==0.16.2 -skimage==0.0 torch==1.8.1 torchvision==0.9.1 -tqdm==4.50.2 - +tqdm==4.50.2 \ No newline at end of file