Mastering Computer Vision with TensorFlow 2.x

Theory behind OpenPose

OpenPose is the first open source real-time two-dimensional pose estimation system for multiple people in an image or video. It was mainly developed by the students and faculties of Carnegie Mellon University (CMU). The title of the paper is OpenPose: Realtime Multi-Person 2D Pose Estimation Using Part Affinity Fields, and the authors are Zhe Cao, Gines Hidalgo, Tomas Simon, Shih-En-Wei, and Yaser Sheikh. You can find the paper at https://arxiv.org/abs/1812.08008.

Note that OpenPose was first presented in CVPR 2017 in the paper titled Real-Time Multi-Person 2D Pose Estimation Using Part Affinity Fields, which can be found at https://arxiv.org/abs/1611.08050. This is followed by the 2018 paper OpenPose: Realtime Multi-Person ...