MOTA: 64.3mAP: 73.5
Papers with Code · CC BY-SA 4.0
Pose Tracking
Pose Tracking is the task of estimating multi-person human poses in videos and assigning unique instance IDs for each keypoint across frames. Accurate estimation of human keypoint-trajectories is useful for human action recognition, human interaction understanding, motion capture and animation.
<span class="description-source">Source: LightTrack: A Generic Framework for Online Top-Down Human Pose Tracking </span>
논문
MOTA: 54.86mAP: 67.81
MOTA: 61.37mAP: 74.03
MOTA: 61.15mAP: 74.04
FID: 0.049Diversity: 9.688Multimodality: 1.119R Precision Top3: 0.800
FID: 0.055Diversity: 9.717Multimodality: 1.687R Precision Top3: 0.814
FID: 0.030Diversity: 9.556Multimodality: 1.374R Precision Top3: 0.809
FID: 31.16MMC: 0.261GenDiv: 10.87PFC: 2.53
MOTA: 57.93mAP: 74.95
MOTA: 51.82mAP: 59.56
FID: 0.032Multimodality: 1.288R Precision Top3: 0.809
MOTA: 63.5IDF1: 73.2IDs: 6760
MOTA: 53.81mAP: 70.28
FID: 0.051Diversity: 9.480R Precision Top3: 0.803
FID: 42.39MMC: 0.25GenDiv: 14.37PFC: 0.54
MOTA: 61.9IDF1: 79.3IDs: 367
MOTA: 58.01mAP: 66.65
FID: 0.053Diversity: 9.585Multimodality: 0.8104R Precision Top3: 0.788
FID: 0.045Multimodality: 1.241R Precision Top3: 0.807
FID: 0.028Diversity: 9.521Multimodality: 2.705R Precision Top3: 0.829
FID: 0.030Diversity: 9.575Multimodality: 1.714R Precision Top3: 0.842
FID: 43.90MMC: 0.250GenDiv: 9.23PFC: 3.05
MOTA: 50.98mAP: 62.95
MOTA: 48.37mAP: 59.22
MOTA: 28.2MOTP: 55.7
FID: 31.01MMC: 0.271GenDiv: 10.98PFC: 2.33
3DMOTA: 94.1