-
The entire modules can be installed following the instruction below. Note that you may want to install body module only which has fewer dependencies. In this case, you may skip some steps. See below the details.
-
The basic installation
conda create -n venv_frankmocap python=3.7 conda activate venv_frankmocap # Install basic dependencies sudo apt-get install libglu1-mesa libxi-dev libxmu-dev libglu1-mesa-dev freeglut3-dev libosmesa6-dev # Install ffmpeg sudo apt-get install ffmpeg # Install cuda # Choose versions based on your system. For example: # conda install cudatoolkit=10.1 cudnn=7.6.0 # Install pytorch and torchvision conda install -c pytorch pytorch==1.6.0 torchvision cudatoolkit=10.1 # Install other required libraries pip install -r docs/requirements.txt
-
Install Detectron2 (for hand module)
- This is required for hand motion capture. You can skip this if you need only body module
- If you followed the versions mentioned above (pytorch 1.6.0, CUDA 10.1, on Linux), you may try the following:
python -m pip install detectron2 -f \ https://dl.fbaipublicfiles.com/detectron2/wheels/cu101/torch1.6/index.html
- If it doesn't work, follow the instruction of Detectron2
-
Install pytorch3d (optional, for pytorch3d renderering)
- We use pytorch3d for an alternative rendering option. We provide other options (opengl by default) and you may skip this process.
- You may try the following (pytorch 1.6.0, on Linux and Mac).
pip install pytorch3d
- If it doesn't work, follow the instruction of Pytorch3D
-
Install other third-party libraries + download pretrained models and sample data
- Run the following script
sh scripts/install_frankmocap.sh
-
Setting SMPL/SMPL-X Models
-
We use SMPL and SMPL-X model as 3D pose estimation output. You have to download them from the original website.
-
Download SMPL Model (Neutral model: basicModel_neutral_lbs_10_207_0_v1.0.0.pkl):
- Download in the original website. You need to register to download the SMPL data.
- Put the file in: ./extra_data/smpl/basicModel_neutral_lbs_10_207_0_v1.0.0.pkl
- This is only for body module
-
Download SMPLX Model (Neutral model: SMPLX_NEUTRAL.pkl):
- You can use SMPL-X model for body mocap instead of SMPL model.
- Download
SMPLX_NEUTRAL.pkl
in the original SMPL website. You need to register to download the SMPLX data. - Put the ```SMPLX_NEUTRAL.pkl`` file in: ./extra_data/smpl/SMPLX_NEUTRAL.pkl
- This is for hand module and whole body module
-
- Once you sucessfully installed and downloaded all, you should have the following files in your directory:
./extra_data/ ├── hand_module │ └── mean_mano_params.pkl │ └── SMPLX_HAND_INFO.pkl | └── pretrained_weights | | └── pose_shape_best.pth │ └── hand_detector │ └── faster_rcnn_1_8_132028.pth │ └── model_0529999.pth ├── body_module | └──body_pose_estimator | └── checkpoint_iter_370000.pth └── smpl └── basicModel_neutral_lbs_10_207_0_v1.0.0.pkl └── SMPLX_NEUTRAL.pkl ./detectors/ ├── body_pose_estimator ├── hand_object_detector └── hand_only_detector
-
The basic installation
conda create -n venv_frankmocap python=3.7 conda activate venv_frankmocap # Install cuda # Choose versions based on your system. For example: # conda install cudatoolkit=10.1 cudnn=7.6.0 # Install pytorch and torchvision conda install -c pytorch pytorch==1.6.0 torchvision cudatoolkit=10.1 # Install other required libraries pip install -r docs/requirements.txt
-
Install pytorch3d (optional, for pytorch3d renderering)
- We use pytorch3d for an alternative rendering option. We provide other options (opengl by default) and you may skip this process.
- You may try the following (pytorch 1.6.0, on Linux and Mac).
pip install pytorch3d
- If it doesn't work, follow the instruction of Pytorch3D
-
Install 2D pose detector and download pretrained models and sample data
- Install 2D keypoint detector:
sh scripts/install_pose2d.sh
- Download pretrained model and other extra data
sh scripts/download_data_body_module.sh
- Download sample data
sh scripts/download_sample_data.sh
-
Setting SMPL/SMPL-X Models
- You only need SMPL model. See above