Table 6 Comparative result with the public dataset in Table 4
From: First-person reading activity recognition by deep learning with synthetically generated images
| Â | Baselines | Proposed | Â | ||||
|---|---|---|---|---|---|---|---|
| Â | 1-NN + FV | SVM + FV | DCNN | Â | |||
| Â | Others | Reading | Others | Reading | Others | Reading | Â |
Scores | Precision | 54.2 | 53.1 | 88.5 | 50.7 | 79.0 | 92.8 |
for each class | Recall | 46.1 | 61.1 | 3.0 | 99.6 | 94.2 | 74.9 |
| Â | F-measure | 49.8 | 55.8 | 5.8 | 62.2 | 85.9 | 82.9 |
Averages | Precision | 53.7 | 69.6 | 85.9 | |||
over classes | Recall | 53.6 | 51.3 | 84.6 | |||
| Â | F-measure | 52.8 | 34.0 | 84.4 | |||