• Stars
    star
    286
  • Rank 144,690 (Top 3 %)
  • Language
    Jupyter Notebook
  • License
    BSD 3-Clause "New...
  • Created over 7 years ago
  • Updated over 2 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

Python bindings for the Openpose library

PyOpenPose

Python bindings for the awesome Openpose library.

Openpose github page

Building

Clone and build openpose. If you use cmake then make install will copy all necessary headers and libs to an install forder that you specify (default is /usr/local).

Set an environment variable named OPENPOSE_ROOT pointing to the openpose install folder. For running the example scripts make sure OPENPOSE_ROOT contains a models folder with the openpose models.

Note: Openpose lib is under heavy development and the API changes very often. Some API changes will break PyOpenPose. I try to upgrade as soon as possible but I am usually a few days behind. Openning an issue helps to speed-up the proccess. Current PyOpenPose version is built with openpose commit e382698

Note: PyOpenPose requires opencv3.x. You will have to build openpose with opencv3 as well.

Inside the root folder of PyOpenpose run cmake and build with:

mkdir build
cd build
cmake ..
make

Add the folder containing PyOpenPose.so to your PYTHONPATH.

Building the library for python3 or python2

  • Set WITH_PYTHON3 flag in cmake to True (i.e with cmake-gui).
  • rebuild project

Testing

Check the scripts folder for python examples using PyOpenPose.

More Repositories

1

MocapNET

We present MocapNET, a real-time method that estimates the 3D human pose directly in the popular Bio Vision Hierarchy (BVH) format, given estimations of the 2D body joints originating from monocular color images. Our contributions include: (a) A novel and compact 2D pose NSRM representation. (b) A human body orientation classifier and an ensemble of orientation-tuned neural networks that regress the 3D human pose by also allowing for the decomposition of the body to an upper and lower kinematic hierarchy. This permits the recovery of the human pose even in the case of significant occlusions. (c) An efficient Inverse Kinematics solver that refines the neural-network-based solution providing 3D human pose estimations that are consistent with the limb sizes of a target person (if known). All the above yield a 33% accuracy improvement on the Human 3.6 Million (H3.6M) dataset compared to the baseline method (MocapNET) while maintaining real-time performance
C++
851
star
2

HandTracker

3D Hand Tracking using input from a depth sensor.
Python
311
star
3

MonocularRGB_3D_Handpose_WACV18

Using a single RGB frame for real time 3D hand pose estimation in the wild
Python
160
star
4

MonocularRGB_2D_Handjoints_MVA19

Accurate Hand Keypoint Localization on Mobile Devices
Python
65
star
5

reading_group

Reading group material and links
7
star
6

mocapnet_rosnode

A ROS node for the MocapNET 3D Pose Estimator
C++
5
star
7

wacv_docker

A Dockerfile for our WACV18 paper: Using a single RGB frame for real time 3D hand pose estimation in the wild.
Dockerfile
1
star