HOContactopt

Installation

Environment Setup

We first create a new conda environment.

conda create -n hocopt python=3.8
conda activate hocopt

Note Python version ≥3.8 should be applicable. We tested that Python=3.9 is also suitable for our project.

Install Pytorch and Pytorch3d

Choose the corresponding pytorch version to install according to your cuda version.

pip install torch==1.12.1+cu116 torchvision==0.13.1+cu116 torchaudio==0.12.1 --extra-index-url https://download.pytorch.org/whl/cu116

We recommend using source code to install Pytorch3d.

git clone https://github.com/facebookresearch/pytorch3d.git
cd pytorch3d && pip install -e .

Install other dependencies

We recommend installing these packages using .whl file.

The version in brackets is the reference version in our environment. You need to adjust the version according to your configurations. Different versions can be referred to whl-sites.

torch-cluster (1.6.0+pt112cu116)
torch-scatter (2.0.9)
torch-sparse (0.6.16+pt112cu116)
torch-spline-conv (1.2.1+pt112cu116)
pyg-lib (0.1.0+pt112cu116)

Other dependencies play nicely with pip and can be installed with:

pip install git+https://github.com/hassony2/manopth.git open3d tensorboardX pyquaternion trimesh transforms3d chumpy opencv-python

Download MANO Model

Download the Python 3-compatible MANO code from the manopth website. Copy the mano folder from the manopth project to the root of the ContactOpt folder.

Due to license restrictions, the MANO data files must be downloaded from the original project website. Create an account and download 'Models & Code'. Extract the models folder to the recently created mano directory. The directory structure should be arranged so the following files can be found:

mano/webuser/lbs.py
mano/models/MANO_RIGHT.pkl

Running On Datasets

To run HOCOpt on datasets, download the preprocessed datasets and pretrained checkpoints at the website.

python hocopt/run_hocopt.py --split=SELECT_CLASS --checkpoint=PATH-TO-MODEL

The visualization of optimization results.

python hocopt/run_eval.py --split=SELECT_CLASS --vis

SELECT_CLASS=[handoff, fine, im_dexycb, im_ho3d]

Acknowledgements

This repo is built based on Contactopt and CHOI. We sincerely appreciate their contributions to open source.