OpenPose: Realtime Multi-Person 2D Pose Estimation using Part Affinity Fields
Citations
[...]
1,899 citations
Cites background from "OpenPose: Realtime Multi-Person 2D ..."
...Object detection is then a standard keypoint estimation problem [3,39,60]....
[...]
...Object detection powers many vision tasks like instance segmentation [7, 21, 32], pose estimation [3, 15, 39], tracking [24, 27], and action recognition [5]....
[...]
947 citations
687 citations
Cites methods from "OpenPose: Realtime Multi-Person 2D ..."
...PennAction [69] and PoseTrack [3] are the only ground-truth 2D video datasets we use, while InstaVariety [30] and Kinetics-400 [31] are pseudo ground-truth datasets annotated using a 2D keypoint detector [12, 35]....
[...]
566 citations
Cites background or methods from "OpenPose: Realtime Multi-Person 2D ..."
...To obtain 3D pose, shape priors [1, 29] have been incorporated to minimize the shape fitting errors in recent approaches [3, 4, 18, 2]....
[...]
...By detecting the 2D key points, great progress has been made to estimate the 2D poses [32, 4, 47]....
[...]
352 citations
References
123,388 citations
49,914 citations
30,462 citations
"OpenPose: Realtime Multi-Person 2D ..." refers background in this paper
...Each confidence map is a 2D representation of the belief that a particular body part can be located in any given pixel....
[...]
28,888 citations
"OpenPose: Realtime Multi-Person 2D ..." refers methods in this paper
...s 3D realtime single-person keypoint detection, able to predict 3D pose estimation out of multiple synchronized camera views. It performs 3D triangulation with non-linear LevenbergMarquardt refinement [65]. The inference time of OpenPose outperforms all stateof-the-art methods, while preserving high-quality results. It is able to run at about 22 FPS in a machine with a Nvidia GTX 1080 Ti while preservi...
[...]
27,821 citations