Skip to main content
. 2022 Oct 3;18(1):117–125. doi: 10.1007/s11548-022-02761-6

Table 2.

We optimize for the multiple Object Tracking Accuracy (MOTA), each performance metric is averaged across all validation folds

Model MOTA MOTA MOTA MOTA MOTA MOTA MOTA MOTP Prec. Rec. F1 Score
wrist thumb index middle ring pinky total total total total total
Baseline [12] 36.7 45.83 57.35 45.53 34.63 8.49 38.27 85 78.3 59.13 67.37
Our model 30.99 44.74 58.21 48.90 36.46 10.39 39.31 85.28 77.61 62.69 69.35

The bolded numbers represent the best performing scores, in comparison between method