Character Recognition on Fonts 109 fonts (test)
72.6AccuracyCROSSGRAD
Evaluation results
| Method | Links | |
|---|---|---|
| CROSSGRADBackbone=Modified LeNet [LeCun et al., 1998]2018.04 | 72.6 | |
| LABELGRADBackbone=Modified LeNet [LeCun et al., 1998], Approach=Label Adversarial Perturbation2018.04 | 71.4 | |
| DANBackbone=Modified LeNet [LeCun et al., 1998], Approach=Domain Adversarial Network2018.04 | 68.9 | |
| BaselineBackbone=Modified LeNet [LeCun et al., 1998], Special Training=None2018.04 | 68.5 |