Pytorch Implementation of From Image Collections to Point Clouds with Self-supervised Shape and Pose Networks.

We recommend using Anaconda to create a new environment to install Pytorch3d.

conda create -n pytorch3d python=3.9
conda activate pytorch3d
conda install pytorch==1.11.0 torchvision==0.12.0 torchaudio==0.11.0 cudatoolkit=11.3 -c pytorch
conda install -c fvcore -c iopath -c conda-forge fvcore iopath
conda install -c bottler nvidiacub
conda install pytorch3d -c pytorch3d-nightly
pip install wandb pandas opencv-python pillow

We use the code provided by Tulsiani et al. to obtain the rendered images and part segmentation maps. Download links for the ShapeNet point cloud dataset is provided below:
ShapeNet pointclouds (~2.8 GB): https://drive.google.com/open?id=1cfoe521iTgcB_7-g_98GYAqO553W8Y0g

Modify the config dictionary inside run.py and execute the following:

python run.py

Modify the config dictionary inside run_save.py and execute the following:

python run_save.py

@inproceedings{navaneet2020ssl3drecon, author = {Navaneet, K L and Mathew, Ansu and Kashyap, Shashank and Hung, Wei-Chih and Jampani, Varun and Babu, R Venkatesh}, booktitle = {IEEE Conference on Computer Vision and Pattern Recognition (CVPR)}, title = {From Image Collections to Point Clouds with Self-supervised Shape and Pose Networks}, year = {2020} }