Skip to main content
. 2021 Jun 29;21(13):4443. doi: 10.3390/s21134443

Figure 3.

Figure 3

Illustrative results for codebook generation and latent space-based representation (HDM05 and CMU subset datasets). Top: Codebook generation for a Mocap video of the throwing high with the right hand while standing class (HDM05). Middle: Codebook generation for a Mocap record of boxing class (CMU subset). Bottom left: PCA-based latent space for HDM05 video. Bottom right: PCA-based latent space for CMU subset video. The first two components are shown for visualization purposes. Black markers represent the original input Mocap frames (time series). Color markers represent the chosen frames (codebook).