Research Article
Kernel Two-Dimensional Nonnegative Matrix Factorization: A New Method to Target Detection for UUV Vision System
Table 5
Average detection accuracy (standard deviation) (%) versus training number (TN) on testing set 3.
| | n | 300 | 400 | 500 | 600 | 700 | 800 | 900 | 1000 |
| | BDPCA | 71.54 (3.18) | 73.32 (3.46) | 76.51 (3.32) | 77.14 (2.46) | 76.18 (3.63) | 78.37 (4.29) | 81.47 (3.76) | 83.34 (3.07) | | RC2DPCA | 73.16 (2.77) | 74.84 (3.31) | 77.35 (2.45) | 78.71 (3.37) | 79.21 (3.92) | 81.02 (4.13) | 83.56 (3.52) | 85.62 (2.78) | | 2DNMF | 70.63 (4.09) | 71.25 (3.20) | 75.80 (3.64) | 76.49 (2.21) | 77.83 (3.85) | 81.75 (4.49) | 84.42 (4.53) | 86.37 (4.16) | | PNMF | 71.20 (4.24) | 73.46 (4.08) | 78.22 (2.53) | 80.82 (4.65) | 81.42 (4.06) | 82.96 (4.21) | 85.38 (4.17) | 88.09 (3.58) | | MKNMF | 76.28 (3.17) | 78.25 (2.76) | 80.19 (2.62) | 82.04 (2.67) | 83.20 (3.68) | 85.27 (3.93) | 86.84 (3.41) | 89.24 (3.25) | | K2DNMF | 77.46 (2.49) | 79.04 (2.40) | 81.4 (2.12) | 83.28 (2.06) | 84.70 (3.23) | 86.02 (3.28) | 88.51 (3.14) | 90.06 (2.83) |
|
|
