Table 3 Comparative results of the VTSAMRNN-FARS method with existing models.
Methods | \(\:\varvec{A}\varvec{c}\varvec{c}{\varvec{u}}_{\varvec{y}}\) | \(\:\varvec{P}\varvec{r}\varvec{e}{\varvec{c}}_{\varvec{n}}\) | \(\:\varvec{S}\varvec{e}\varvec{n}{\varvec{s}}_{\varvec{y}}\) | \(\:\varvec{S}\varvec{p}\varvec{e}{\varvec{c}}_{\varvec{y}}\) |
|---|---|---|---|---|
VTSAMRNN-FARS | 99.67 | 99.67 | 99.67 | 99.67 |
Open Pose-LSTM | 92.72 | 94.85 | 97.53 | 95.01 |
2D-ConvNN | 95.88 | 94.35 | 94.93 | 97.70 |
ResNet50 | 96.21 | 95.15 | 95.40 | 91.89 |
2D Pose estimation | 95.60 | 92.00 | 91.90 | 98.16 |
TD_CNN-LSTM | 99.23 | 93.99 | 93.64 | 96.85 |
RetinaNet | 94.69 | 96.75 | 95.03 | 95.61 |
YOLOv7 | 91.51 | 96.68 | 92.34 | 96.10 |
YOLOv5 | 94.83 | 91.94 | 97.16 | 94.34 |
