Towards Accurate Multi-person Pose Estimation in the Wild

Benchmark Model Rank Results
keypoint-detection-on-coco-test-challengeG-RMI*AR: 75.1ARM: 69.7AP: 69.1AP50: 85.9AP75: 75.2APL: 82.4
keypoint-detection-on-coco-test-devG-RMIAPM: 62.3APL: 70.0AP50: 85.5AP75: 71.3AR: 69.7AR50: 88.7
multi-person-pose-estimation-on-cocoG-RMIAP: 0.649
multi-person-pose-estimation-on-cocoG-RMI*AP: 0.685
multi-person-pose-estimation-on-coco-test-devG-RMIAP: 64.9APL: 70.0APM: 62.3AP50: 85.5AP75: 71.3
pose-estimation-on-coco-test-devG-RMIAP: 64.9AP50: 85.5AP75: 71.3APL: 70.0AR: 69.7