1.下载模型
进入open_model_zoo路径
cd /opt/intel/openvino/deployment_tools/open_model_zoo/tools/downloader
查看所有模型命令：
./downloader.py --print_all
在模型列表中找到要下载的模型并下载：
./downloader.py --name human-pose-estimation-0001
记录xml文件下载路径：
/opt/intel/openvino_2019.3.334/deployment_tools/open_model_zoo/tools/downloader/intel/human-pose-estimation-0001/FP32/human-pose-estimation-0001.xml
2.编译
执行下列命令
cd /opt/intel/openvino/deployment_tools/inference_engine/demos
./build_demos.sh
执行完后，会在home目录生成下面文件目录：
/home/kang/omz_demos_build/intel64/Release
进入multichannel_demo，执行make
cd ~/omz_demos_build/human_pose_estimation_demo
make -j4
注意：也可以直接在omz_demos_build目录中执行编译，则对文件夹下所有模型生效
3.运行
cd ~/omz_demos_build/intel64/Release
./human_pose_estimation_demo -m /opt/intel/openvino_2019.3.334/deployment_tools/open_model_zoo/tools/downloader/intel/human-pose-estimation-0001/FP32/human-pose-estimation-0001.xml -i ~/Downloads/head-pose-face-detection-female.mp4 -d CPU
4.得到结果和图像信息
