Table 3 Recognition results for the AFEW database
From: Multithreading cascade of SURF for facial expression recognition
Method | Accuracy on AFEW (%) | Â | ||||||
|---|---|---|---|---|---|---|---|---|
| Â | Ne | An | Di | Fe | Ha | Sa | Su | mAP. |
LBP-TOP [12] | 9.0 | 11.7 | 19.6 | 17.9 | 42.3 | 23.8 | 33.6 | 25.1 |
HOE [13] | 6.1 | 11.2 | 16.5 | 9.0 | 33.5 | 15.3 | 28.3 | 19.5 |
HOG 3D [14] | – | – | – | – | – | – | – | 26.9 |
3D LUT [6] | 6.8 | 45.7 | 0 | 0 | 62.0 | 13.2 | 48.6 | 25.2 |
LSH-CORF [36] | – | 23.1 | 12.8 | 38.6 | 9.7 | 21.1 | 10.9 | 21.8 |
STM-ExpLet [29] | – | – | – | – | – | – | – | 31.7 |
McSURF (3T) | 16.3 | 47.3 | 8.2 | 31.6 | 36.5 | 15.4 | 72.8 | 32.6 |
McSURF (OD) | 15.3 | 46.5 | 7.6 | 31.2 | 34.7 | 14.6 | 75.2 | 32.2 |