Research Article
Car Emotion Labeling Based on Color-SSL Semi-Supervised Learning Algorithm by Color Augmentation
Table 4
Comparison of accuracy between Color-SSL and FlexMatch.
| | Numbers of label sample | Method | Train and test | | 1 | 2 | 3 | 4 | 5 |
| | 10 | Color-SSL | 77.52 (1.22) | 76.26 (1.36) | 75.89 (2.15) | 77.34 (1.39) | 75.95 (1.16) | | FlexMatch | 72.75 (2.09) | 73.69 (1.81) | 71.05 (2.13) | 72.69 (2.11) | 71.36 (2.16) | | 50 | Color-SSL | 81.71 (0.89) | 79.56 (1.12) | 81.73 (0.95) | 82.53 (0.68) | 81.87 (1.69) | | FlexMatch | 73.01 (2.19) | 74.29 (1.25) | 75.09 (0.92) | 74.10 (1.28) | 76.61 (0.93) | | 100 | Color-SSL | 91.15(0.34) | 91.26 (1.01) | 92.53 (0.65) | 91.14 (0.65) | 90.81 (0.85) | | FlexMatch | 79.33 (2.35) | 81.88 (1.65) | 82.73 (1.32) | 81.75 (1.16) | 82.72 (0.95) | | 200 | Color-SSL | 94.77 (0.30) | 91.47 (1.29) | 93.26 (0.74) | 93.65 (0.82) | 92.34 (0.95) | | FlexMatch | 87.91 (0.82) | 90.03 (0.84) | 92.31 (0.55) | 93.15 (1.78) | 91.51 (0.69) |
|
|
