Research Article
Incremental Graph Regulated Nonnegative Matrix Factorization for Face Recognition
Table 4
The recognition rates (%) for the CMU-PIE database during adding 5 batches of images belonging to different classes.
| ā | Supervised methods | Unsupervised methods | ā | BIGNMF-S | GNMF-S | BIGNMF-U | GNMF-U | IOPNMF | NMF |
| Start | 81.37 | 81.37 | 79.15 | 79.64 | 46.32 | 79.01 | Add 1 | 81.46 | 80.78 | 79.21 | 78.77 | 51.23 | 79.14 | Add 2 | 81.26 | 81.98 | 81.29 | 81.75 | 54.85 | 79.85 | Add 3 | 83.28 | 82.45 | 81.36 | 81.03 | 53.86 | 79.38 | Add 4 | 83.3 | 83.45 | 83.61 | 83.59 | 51.96 | 80.28 | Add 5 | 84.75 | 84.49 | 84.73 | 84.58 | 51.47 | 81.97 |
|
|