Table 6 Comparison of our DSwarm-Net model with some recent models on the UTD-MHAD dataset by cross-subject analysis.
From: A union of deep learning and swarm-based optimization for 3D human action recognition
Method | Year | Accuracy (%) | Skeleton data | RGB data | Inertial data |
|---|---|---|---|---|---|
Action machine63 | 2018 | 92.50 | Yes | Yes | No |
PEM64 | 2018 | 94.51 | Yes | Yes | No |
BHDM65 | 2019 | 92.80 | Yes | No | No |
Correlation Congruence66 | 2019 | 94.87 | Yes | Yes | Yes |
Gimme DSE67 | 2020 | 93.30 | Yes | No | Yes |
Fuzzy CNN fusion68 | 2020 | 97.91 | Yes | No | No |
SAKDN69 | 2021 | 98.04 | Yes | Yes | Yes |
Edge Motion70 | 2021 | 95.59 | Yes | No | No |
AMGC71 | 2021 | 95.11 | Yes | No | No |
DSwarm-Net | 2021 | 98.13 | Yes | No | No |
