Table 1 Results on view 2 of LFW. MRH approaches used a 1024-component visual dictionary.
From: Face Recognition from Still Images to Video Sequences: A Local-Feature-Based Framework
Method | Mean accuracy | Standard error |
|---|---|---|
| 72.95 | 0.55 |
| 70.38 | 0.48 |
| 67.85 | 0.42 |
PCA (normalised distance) | 59.82 | 0.68 |
PCA (raw distance) | 57.23 | 0.68 |
Randomised Binary Trees (RBT) | 72.45 | 0.40 |
LBP-MSE-80 | 65.27 | 0.47 |
GJD-BC-100 | 67.98 | 0.65 |
SIFT-SIMPLE | 62.95 | 0.71 |
Single LE + holistic | 81.22 | 0.53 |
MRH (probabilistic, normalised distance)
MRH (probabilistic, raw distance)
MRH (probabilistic, normalised distance)