Go to file
2021-01-19 16:44:46 +01:00
.vscode update setup and add debugging 2021-01-09 16:45:09 +01:00
models add models folder 2021-01-09 14:58:38 +01:00
modules update camera implementation 2021-01-19 16:44:46 +01:00
samples rename sample data 2021-01-12 12:30:55 +01:00
.gitignore update setup and add debugging 2021-01-09 16:45:09 +01:00
config.yaml add models folder 2021-01-09 14:58:38 +01:00
dataset.py cleanup project 2021-01-15 12:04:35 +01:00
Dockerfile add OpenPose docker file 2021-01-12 14:41:54 +01:00
example_fitter.py update camera implementation 2021-01-19 16:44:46 +01:00
example.py WIP: fixing mapping for camera solver 2021-01-15 22:47:15 +01:00
LICENSE Initial commit 2020-12-19 18:26:10 +01:00
model.py WIP: fixing mapping for camera solver 2021-01-15 22:47:15 +01:00
preview.png WIP: fixing mapping for camera solver 2021-01-15 22:47:15 +01:00
README.md setup: fix typo 2021-01-09 16:50:32 +01:00
renderer.py WIP: cleanup scene rendering 2021-01-19 12:31:07 +01:00
requirements.txt update required packages 2021-01-12 14:40:20 +01:00
setup.sh setup: add smpl download script 2021-01-09 16:44:16 +01:00
utils.py cleanup structure 2021-01-19 13:05:02 +01:00

realtime-body-tracking

Installation

This project is using python version 3.6.0. It is also recommended to use anaconda for managing your python environments.

Downloading SMPL models

The SMPL models can be downloaded from this link. Alternatively just use the following script:

./setup.sh

This should copy and rename the SMPL model to the correct folders. Either way the resulting folder structure should look like this (ignore smplx if you don't intend to use it):

└── models
    ├── smpl
    │   ├── SMPL_FEMALE.npz
    │   ├── SMPL_MALE.npz
    │   └── SMPL_NEUTRAL.pkl
    └── smplx
        ├── SMPLX_FEMALE.npz
        ├── SMPLX_FEMALE.pkl
        ├── SMPLX_MALE.npz
        ├── SMPLX_MALE.pkl
        ├── SMPLX_NEUTRAL.npz
        └── SMPLX_NEUTRAL.pkl

Conda Environment

Create a new conda env by typing

conda create --name tum-3d-proj python=3.6

When using VSCode the IDE should automatically use the python interpreter created by conda.

Pip packages

The required pip packages are stored within the requirements.txt file and can be installed via

pip install -r requirements.txt

Exporting imported packages to pip

Since pip is trash and does not save imported packages in a file within the project you have to force it to do so.

pip freeze > requirements.txt

OpenPose