OpenPose is an open-source library developed by the CMU Perceptual Computing Lab for real-time multi-person 2D pose estimation. It detects keypoints on the human body, face, hands, and feet, enabling applications like gesture recognition, human-computer interaction, and motion analysis.
OpenPose utilizes Part Affinity Fields (PAFs) to associate body parts with individuals in images, achieving high accuracy and real-time performance.
The library supports multiple operating systems and offers pre-trained models for various tasks. For detailed information and access to the codebase, visit the official OpenPose GitHub repository.
« Back to Glossary Index