Environment

• python 3.7
• cuda 11.5
• cudnn 8.3.2.44
• cmake 3.12 or newer
• OS Windows10
• GPU nVidia GTX1060
• Install opencv-python

python -m pip install --upgrade pip
python -m pip install opencv-python

ref. https://github.com/CMU-Perceptual-Computing-Lab/openpose/blob/master/doc/installation/1_prerequisites.md#windows-prerequisites

Download

git clone https://github.com/CMU-Perceptual-Computing-Lab/openpose
cd openpose/
git submodule update --init --recursive --remote

Build

1. 在openpose下建立新的資料夾，名稱為 build
2. 開啟CMake-gui

• 設定路徑參數，如下圖
  

• 選取BUILD_PYTHON, DOWNLOAD_BODY_COCO_MODE and DOWNLOAD_BODY_MPI_MODEL
  

• 點擊Configure，接著點擊Generate
  

3. 透過openpose/build/OpenPose.sln打開Visual Studio 2019
   

• 選取方案組態 “Release”
  
• 開始Build，點擊建置方案
  

Execution

• 打開命令提示元
  
• 進入openpose/build/examaples/tutorial_api_python
  
• 執行01_body_from_image.py

python 01_body_from_image.py