Openpose Cmake

Posted on by

OpenPose represents the first real-time multi-person system to jointly detect human body, hand, and facial keypoints (in total 130 keypoints) on single images.

Apr 23, 2020 Openpose This sample can be put in the folder examples/usercode/ OR preferably, compile and install openpose with the cmake and compile this anywhere The installation process is very easy using cmake. Clone the repository.

  • E.g., Nvidia 30XX GPUs require at least CUDA 11, others (GTX 20XX, V100, Volta or Turing GPUs) require at least CUDA 10. CMake automatically downloads all the OpenPose models. However, some firewall or company networks block these downloads. If so, you might need to download them manually.
  • You can create your custom code on examples/usercode/ and quickly compile it with CMake when compiling the whole OpenPose project. Quickly add your custom code: See examples/usercode/ for further details. OpenPose Python API: Analogously to the C API, find the tutorial for the Python API on examples/tutorialapipython/.

3D-Pose-Baseline: “We provide a strong baseline for 3d human pose estimation that also sheds light on the challenges of current approaches.”

3D-Pose-Baseline Dependencies

  • tensorflow 1.0 or later, follow guide

Openpose Cmake 3

OpenPose Prerequisites

Openpose cmake error
  1. Download and install CMake GUI:
    • Ubuntu: run the command sudo apt-get install cmake-qt-gui
    • Windows: download and install the latest CMake win64-x64 msi installer from the CMake website, called cmake-X.X.X-win64-x64.msi.
  2. CUDA 8:
    • Ubuntu: Run ubuntu/ or alternatively download and install it from their website.
    • Windows: Install CUDA 8.0 after Visual Studio 2015 is installed to assure that the CUDA installation will generate all necessary files for VS. If CUDA was already installed, re-install CUDA after installing VS!
  3. cuDNN 5.1:
    • Ubuntu: Run ubuntu/ or alternatively download and install it from their website.
    • Windows (and Ubuntu if manual installation): In order to manually install it, just unzip it and copy (merge) the contents on the CUDA folder, usually /usr/local/cuda/ in Ubuntu and C:Program FilesNVIDIA GPU Computing ToolkitCUDAv8.0 in Windows.
  4. Ubuntu – Other prerequisites:
    • Caffe prerequisites: By default, OpenPose uses Caffe under the hood. If you have not used Caffe previously, install its dependencies by running bash ./ubuntu/
    • OpenCV must be already installed on your machine. It can be installed with apt-get install libopencv-dev. You can also use your own compiled OpenCV version


Openpose Cmake For Windows

  1. install caffe and opencv, follow guide
  2. follow OpenPose installation ,use custom caffe introduction
  3. clone 3d-pose-baseline (the conversion/mapping from openpose format to 3d-pose-baseline is done here src/
  4. download pre-trained model with 200 epochs

and simply run

I am trying to run the openpose example the COCO model.

However when I compile and run (my executable is 'kpd' for 'Key Point Detection', I get the following error about the file being open As brave as you book pdf.

kpd -i /home/brian/Pictures/MyPic/IMG20180801205312.jpg -m pose_iter_440000.caffemodel -p pose_deploy_linevec.prototxt

OpenCV(3.4.1) Error: Unspecified error (FAILED: fs.is_open(). Can't open 'true') in ReadProtoFromTextFile, file /home/brian/Progams/opencv/modules/dnn/src/caffe/caffe_io.cpp, line 1119 terminate called after throwing an instance of 'cv::Exception' what(): OpenCV(3.4.1) /home/brian/Progams/opencv/modules/dnn/src/caffe/caffe_io.cpp:1119: error: (-2) FAILED: fs.is_open(). Can't open 'true' in function ReadProtoFromTextFile

Aborted (core dumped)


Openpose Cmake Project

It doesn't appear to matter when the prototxt file is in the same folder or not, or whether I use absolute paths or not. I've also tried different models to no avail.

It's probably something simple, but I'm just getting started, so a bit green at OpenCV.

Cmakefiles/openpose_lib.dir/all Error 2

Thanks in Advance

Openpose Cmake Install

editretagflag offensiveclosemergedelete