Skip to main content
. 2024 Nov 30;24(23):7682. doi: 10.3390/s24237682

Figure 6.

Figure 6

The average precision gain of Pose-Transformer3D with one Transformer decoder layer compared to no self-attention method on the NTU RGB+D 60 dataset. The vertical axis represents the percentage of average precision gain, and the horizontal axis represents 60 action IDs.