Research Article
A Constrained Algorithm Based NMFα for Image Representation
Table 1
The comparison of classification accuracy rates on the ORL Database.
| | Classification accuracy Ac() (%) | CNMF | | | | |
| 2 | 93.90 | 92.95 | 93.51 ± 0.09 | 92.14 ± 0.18 | 93.89 ± 0.91 | 3 | 86.20 | 84.33 | 85.71 ± 0.83 | 82.26 ± 1.23 | 86.21 ± 2.07 | 4 | 84.55 | 84.05 | 84.74 ± 0.56 | 83.02 ± 0.47 | 84.78 ± 0.36 | 5 | 80.16 | 78.38 | 80.10 ± 0.22 | 80.15 ± 0.18 | 80.16 ± 0.29 | 6 | 76.72 | 76.73 | 77.13 ± 0.28 | 77.60 ± 0.19 | 78.09 ± 0.40 | 7 | 81.14 | 79.04 | 80.34 ± 0.81 | 78.07 ± 0.88 | 82.15 ± 1.95 | 8 | 78.08 | 77.02 | 78.66 ± 0.69 | 78.77 ± 0.94 | 79.45 ± 1.33 | 9 | 83.19 | 82.23 | 83.57 ± 0.36 | 81.22 ± 0.21 | 85.19 ± 1.21 | 10 | 77.03 | 76.88 | 77.69 ± 0.12 | 77.75 ± 0.19 | 78.96 ± 0.73 |
| Avg. | 82.33 | 81.29 | 82.38 | 81.22 | 83.21 |
|
|