Real-time multi-person pose estimation and tracking with AlphaPose
Real-time multi-person pose estimation and tracking with AlphaPose
AlphaPose: Whole-Body Regional Multi-Person Pose Estimation and Tracking in Real-Time
arXiv paper abstract https://arxiv.org/abs/2211.03375
arXiv PDF paper https://arxiv.org/pdf/2211.03375.pdf
GitHub https://github.com/MVIG-SJTU/AlphaPose
Accurate whole-body multi-person pose estimation and tracking is an important yet challenging topic in computer vision.
To capture the subtle actions of humans for complex behavior analysis, whole-body pose estimation including the face, body, hand and foot is essential over conventional body-only pose estimation.
… present AlphaPose, a system that can perform accurate whole-body pose estimation and tracking jointly while running in realtime.
… propose several new techniques: Symmetric Integral Keypoint Regression (SIKR) for fast and fine localization, Parametric Pose Non-Maximum-Suppression (P-NMS) for eliminating redundant human detections and Pose Aware Identity Embedding for jointly pose estimation and tracking.
… method is able to localize whole-body keypoints accurately and tracks humans simultaneously given inaccurate bounding boxes and redundant detections.
… show a significant improvement over current state-of-the-art methods in both speed and accuracy …
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
