Table 2.

Evaluation of pose estimation techniques on the JHMDB and DeepLesion Datasets.

Model	JHMDB dataset				DeepLesion dataset
	Accuracy	Recall	F1 score	AUC	Accuracy	Recall	F1 score	AUC
HRNet (41)	89.54 ± 0.02	87.89 ± 0.03	86.21 ± 0.02	88.13 ± 0.03	90.12 ± 0.02	88.47 ± 0.03	87.65 ± 0.02	88.79 ± 0.03
SimpleBaseline (42)	87.88 ± 0.03	85.91 ± 0.02	84.75 ± 0.03	85.30 ± 0.02	88.44 ± 0.03	86.72 ± 0.03	85.33 ± 0.02	86.02 ± 0.03
DarkPose (43)	90.67 ± 0.03	89.34 ± 0.02	88.78 ± 0.02	89.25 ± 0.02	91.55 ± 0.02	90.11 ± 0.03	89.67 ± 0.02	90.42 ± 0.02
OpenPose (44)	88.11 ± 0.02	86.76 ± 0.03	85.92 ± 0.02	86.25 ± 0.03	89.03 ± 0.02	87.41 ± 0.02	86.67 ± 0.03	86.95 ± 0.02
DEKR (45)	91.34 ± 0.03	89.78 ± 0.03	89.12 ± 0.02	89.97 ± 0.03	92.33 ± 0.03	91.02 ± 0.02	90.34 ± 0.02	91.20 ± 0.03
PRTR (46)	89.22 ± 0.02	88.01 ± 0.03	87.14 ± 0.02	87.80 ± 0.02	90.09 ± 0.03	88.85 ± 0.02	87.92 ± 0.03	89.01 ± 0.02
Ours	93.78 ± 0.02	91.56 ± 0.02	90.89 ± 0.03	91.47 ± 0.03	94.23 ± 0.02	92.68 ± 0.02	91.33 ± 0.03	92.12 ± 0.02

The values in bold are the best values.