Human Activity Recognition on HARTH (cross-subject transfer)
58.27Performance S006 -> S032SIGHT
Evaluation results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| SIGHTBackbone=1D-CNN, Learning rate=0.00012026.05 | 58.27 | 65.72 | 79.17 | 61.82 | 97.02 | 98.36 | 76.73 | |
| OATTABackbone=1D-CNN, Learning rate=0.00012026.05 | 48.31 | 56.77 | 75.29 | 62.13 | 96.58 | 87.19 | 71.05 | |
| NOTEBackbone=1D-CNN, Learning rate=0.00012026.05 | 46.23 | 52.24 | 63.7 | 58.37 | 56.87 | 57.04 | 55.74 | |
| TENTBackbone=1D-CNN, Learning rate=0.00012026.05 | 45.92 | 56.78 | 73.09 | 53.4 | 96.28 | 85.39 | 68.48 | |
| OFTTABackbone=1D-CNN, Learning rate=0.00012026.05 | 45.88 | 52.38 | 75.2 | 54.82 | 96.36 | 78.13 | 67.13 | |
| Source-onlyBackbone=1D-CNN2026.05 | 45.87 | 52.37 | 74.83 | 53.71 | 96.52 | 78.25 | 66.93 | |
| RoTTABackbone=1D-CNN, Learning rate=0.00012026.05 | 45.87 | 52.38 | 75.12 | 53.9 | 96.67 | 78.24 | 67.03 | |
| SARBackbone=1D-CNN, Learning rate=0.00012026.05 | 39.02 | 49.05 | 60.06 | 55.79 | 88.37 | 68.53 | 60.14 | |
| COA-HARBackbone=1D-CNN, Learning rate=0.00012026.05 | 36.77 | 50.13 | 47.29 | 59.37 | 46.09 | 43.36 | 47.17 | |
| ACCUPBackbone=1D-CNN, Learning rate=0.00012026.05 | 21.5 | 29.4 | 21.59 | 15.62 | 25.21 | 28.74 | 23.68 |