Research Article
Incremental Graph Regulated Nonnegative Matrix Factorization for Face Recognition
Table 2
The recognition rates (%) for the CMU-PIE database during adding 5 batches of images belonging to one class.
| ā | Supervised methods | Unsupervised methods | ā | BIGNMF-S | GNMF-S | CINMF | BIGNMF-U | GNMF-U | IOPNMF | NMF |
| Start | 85.05 | 85.05 | 82.36 | 83.64 | 83.64 | 50.34 | 79.20 | Add 1 | 84.92 | 83.48 | 82.8 | 83.47 | 82.25 | 50.33 | 78.16 | Add 2 | 84.79 | 83.77 | 82.66 | 83.3 | 83.02 | 50.66 | 78.54 | Add 3 | 84.6 | 84.24 | 82.7 | 83.17 | 82.36 | 50.54 | 78.8 | Add 4 | 84.62 | 84.5 | 82.8 | 83.13 | 82.46 | 50.75 | 78.88 | Add 5 | 84.1 | 83.58 | 82.5 | 82.17 | 82.62 | 50.63 | 78.98 |
|
|