Table 2 Experimental results employing VGG features, before and after fully connected layer (F1-score %)
From: Quantitative comparison of motion history image variants for video-based depression assessment
| Â | Before | After | Feature fusion | |||
|---|---|---|---|---|---|---|
| Â | Max | Mean | 1st | 2nd | 3rd | Â |
MHI | 87.1 | 63.0 | 51.5 | 84.8 | 70.1 | 87.4 |
LMHI | 64.0 | 67.7 | 56.6 | 18.2 | 66.4 | 64.0 |
GMHI | 85.7 | 62.7 | 55.6 | 76.0 | 65.5 | 84.3 |
Combined | 64.6 | 52.1 | 51.8 | 74.7 | 46.3 | 65.1 |