Learning a Mid-Level Representation for Multiview Action Recognition

Figure 4

Action recognition accuracy with different tree depths on the IXMAS dataset.