TABLE 7:
Aggregated running time comparison (mean±std) of different clustering algorithms on benchmark datasets (in seconds). All experiments are conducted on a PC machine with an Intel(R) Core(TM)-i7–5820, 3.3 GHz CPU and 16G RAM in MATLAB environment. The results of MVEC [14] on CCV, Flower102, Caltech102–20, Caltech102–25 and Caltech102–30 are not reported due to the “out of memory” error.
| Datasets | MKKM+ZF | MKKM+MF | MKKM+KNN | MKKM+AF | MKKM-IK | MIC | MVEC | LF-IMVC |
|---|---|---|---|---|---|---|---|---|
| [25] | [30] | [27] | [14] | Proposed | ||||
| ProteinFold | 1.6 ± 0.2 | 1.5 ± 0.1 | 2.3 ± 0.2 | 1.8 ± 0.2 | 4.1 ± 0.6 | 192.1 ± 2.8 | 163.0 ± 1.4 | 1.1 ± 0.1 |
| CCV | 83.7 ± 10.8 | 85.9 ± 10.7 | 124.1 ± 14.5 | 106.4 ± 11.3 | 130.8 ± 5.2 | 2070.8 ± 15.1 | – | 23.0 ± 4.8 |
| Flower17 | 2.8 ± 0.3 | 2.8 ± 0.3 | 3.9 ± 0.5 | 3.5 ± 0.4 | 5.4 ± 0.6 | 186.4 ± 6.5 | 603.3 ± 13.1 | 1.5 ± 0.2 |
| Flower102 | 230.0 ± 17.4 | 239.3 ± 27.9 | 340.7 ± 17.1 | 279.4 ± 28.1 | 322.1 ± 31.0 | 5161.9 ± 94.1 | – | 117.2 ± 10.0 |
| UCI-Digital | 4.3 ± 0.4 | 4.2 ± 0.5 | 4.4 ± 0.5 | 5.0 ± 0.6 | 7.8 ± 0.9 | 153.4 ± 3.6 | 906.1 ± 30.0 | 1.3 ± 0.2 |
| Caltech102–5 | 4.4 ± 0.2 | 4.4 ± 0.1 | 6.8 ± 0.2 | 4.8 ± 0.2 | 31.6 ± 3.3 | 1204.3 ± 330.4 | 187.4 ± 1.0 | 16.6 ± 0.4 |
| Caltech102–10 | 17.3 ± 0.7 | 17.2 ± 0.7 | 29.5 ± 0.7 | 18.9 ± 0.6 | 74.3 ± 17.0 | 2211.3 ± 668.1 | 809.7 ± 14.0 | 25.8 ± 0.6 |
| Caltech102–15 | 55.6 ± 0.7 | 55.8 ± 0.7 | 84.1 ± 2.3 | 58.1 ± 2.1 | 197.1 ± 28.4 | 3379.8 ± 867.8 | 2226.4 ± 46.1 | 44.9 ± 0.6 |
| Caltech102–20 | 111.0 ± 4.7 | 111.3 ± 4.6 | 199.7 ± 25.3 | 120.6 ± 4.8 | 320.3 ± 37.6 | 5370.2 ± 1753.1 | – | 75.5 ± 0.7 |
| Caltech102–25 | 207.3 ± 14.8 | 209.2 ± 19.7 | 362.7 ± 18.7 | 200.6 ± 4.6 | 566.1 ± 71.0 | 9265.5 ± 1465.2 | – | 32.4 ± 2.4 |
| Caltech102–30 | 357.2 ± 21.5 | 364.5 ± 24.4 | 616.4 ± 36.9 | 360.2 ± 16.8 | 828.2 ± 33.4 | 11896.0 ± 1875.8 | – | 139.2 ± 0.8 |