Table 1 Recognition results for the CK+ database
From: Multithreading cascade of SURF for facial expression recognition
Method | Accuracy on CK+ (%) | Â | ||||||
|---|---|---|---|---|---|---|---|---|
| Â | An | Co | Di | Fe | Ha | Sa | Su | mAP. |
CLM [15] | 70.1 | 52.4 | 92.5 | 72.1 | 94.2 | 45.9 | 93.6 | 74.4 |
LBP-TOP [12] | 82.2 | 77.8 | 91.5 | 72.0 | 98.6 | 57.1 | 97.6 | 82.4 |
HOE [13] | 76.4 | 65.4 | 83.6 | 73.3 | 92.1 | 88.6 | 92.8 | 82.3 |
HOG 3D [14] | 84.4 | 77.8 | 94.9 | 68.0 | 100 | 75.0 | 98.8 | 85.6 |
ITBN [35] | 91.1 | 78.6 | 94.0 | 83.3 | 89.8 | 76.0 | 91.3 | 86.3 |
LSH-CORF [36] | 71.3 | – | 90.8 | 79.0 | 92.6 | 90.5 | 96.6 | 86.8 |
3D LUT [6] | 76.3 | 35.1 | 60.5 | 73.8 | 91.0 | 48.2 | 92.8 | 68.2 |
3DCNN-DAP [34] | 91.1 | 66.7 | 96.6 | 80.0 | 98.6 | 85.7 | 96.4 | 87.9 |
PLBP [33] | – | – | – | – | – | – | – | 96.7 |
McSURF (3T) | 91.6 | 70.2 | 84.9 | 86.3 | 94.6 | 88.4 | 96.5 | 87.5 |
McSURF (OD) | 96.1 | 74.6 | 86.2 | 92.5 | 98.2 | 94.1 | 96.4 | 91.2 |