Table 4.
The non-subject-specific session to session classification results (accuracy in percentage) on dataset 2a. The best accuracies are highlighted in bold; the lowest accuracies of SCNN-MCD with respect to SCNN are underlined.
Methods | A01 | A02 | A03 | A04 | A05 | A06 | A07 | A08 | A09 | Mean |
---|---|---|---|---|---|---|---|---|---|---|
SCN [31] | 47.06 | 31.22 | 41.02 | 33.19 | 41.57 | 34.71 | 43.09 | 46.01 | 51.78 | 41.07 |
MCNN [34] | 51.91 | 38.06 | 43.34 | 35.81 | 41.50 | 31.11 | 48.09 | 45.01 | 51.29 | 42.09 |
CCNN [34] | 62.07 | 42.44 | 63.12 | 52.09 | 49.96 | 37.16 | 62.54 | 59.32 | 69.43 | 55.34 |
SCNN [29] | 72.29 | 39.26 | 81.59 | 60.90 | 54.03 | 51.58 | 74.70 | 77.11 | 76.86 | 65.37 |
SCNN-MCD | 80.39 | 46.28 | 82.91 | 62.65 | 58.78 | 49.76 | 71.51 | 79.76 | 78.76 | 67.87 |