Table 3 KTH, Weizmann, and UCF Sports Action datasets’ description
From: A novel human activity recognition architecture: using residual inception ConvLSTM layer
| Â | Number of actions | Number of videos | Average video length |
|---|---|---|---|
KTH | 6 | 2391 | 4 s |
Weizmann | 10 | 90 | 1.5 s |
UCF Sports Action | 10 | 150 | 2.20 to 14.40 s |