Table 9.
Subjects | Original data (a) | AO-WGAN (b) | AO-GAN (c) | GAL-WGAN (d) | GAL-GAN (e) |
---|---|---|---|---|---|
MI dataset reconstructed by the gal and ao | |||||
GAN/wgan models | |||||
MI-1 | 60.59 | 61.63 | 59.90 | 58.33 | 67.53 |
MI-2 | 70.31 | 71.70 | 69.10 | 75.00 | 73.26 |
MI-3 | 53.47 | 58.16 | 54.17 | 53.99 | 55.03 |
MI-4 | 59.38 | 68.58 | 68.58 | 64.41 | 66.67 |
MI-5 | 72.22 | 68.92 | 76.22 | 73.09 | 66.67 |
MI-6 | 67.36 | 70.49 | 70.49 | 69.62 | 68.75 |
MI-7 | 56.25 | 58.68 | 73.44 | 58.85 | 60.07 |
MI-8 | 57.81 | 54.34 | 54.51 | 63.54 | 56.60 |
MI-9 | 60.42 | 57.12 | 57.99 | 56.08 | 55.90 |
AVG | 61.98 | 63.29 | 64.93 | 63.66 | 63.39 |
T-test | – | b vs. a | c vs. a | d vs. a | e vs. a |
p-value | – | p = 0.380 | p = 0.215 | p = 0.175 | p = 0.215 |
The bold text is the best performance.