Research Article
Bio-Inspired Structure Representation Based Cross-View Discriminative Subspace Learning via Simultaneous Local and Global Alignment
Table 3
Classification results (%) of all methods on 20% corrupted CMU-PIE dataset.
| Methods | C1 | C2 | C3 | C4 | C5 | C6 |
| LDA | 9.23 ± 0.10 | 8.66 ± 0.07 | 8.57 ± 0.13 | 6.27 ± 0.11 | 6.82 ± 0.12 | 6.59 ± 0.10 | PCA | 25.98 ± 0.06 | 25.99 ± 0.15 | 26.58 ± 0.14 | 21.97 ± 0.10 | 22.01 ± 0.19 | 20.15 ± 0.17 | LatLRR | 38.26 ± 1.04 | 34.67 ± 1.18 | 35.00 ± 0.84 | 36.11 ± 1.12 | 34.97 ± 0.74 | 35.76 ± 0.78 | LPP | 34.25 ± 0.22 | 30.08 ± 0.67 | 29.89 ± 0.93 | 33.27 ± 0.25 | 30.95 ± 0.51 | 31.07 ± 0.59 | SRRS | 74.30 ± 0.14 | 63.02 ± 0.19 | 72.79 ± 0.15 | 59.21 ± 0.36 | 68.73 ± 0.17 | 54.98 ± 0.27 | RCVL | 70.63 ± 0.08 | 61.39 ± 0.16 | 71.44 ± 0.10 | 57.34 ± 0.16 | 65.02 ± 0.09 | 53.05 ± 0.07 | Ours | 74.58 ± 0.03 | 64.74 ± 0.12 | 73.17 ± 0.07 | 60.86 ± 0.08 | 69.06 ± 0.06 | 56.56 ± 0.10 |
|
|