Handwritten Character Classification on EMNIST-Letters (test)
95.88AccuracyKabir et al. (2020)
Evaluation results
| Method | Links | |
|---|---|---|
| Kabir et al. (2020)2022.05 | 95.88 | |
| WaveMix-128/7Number of parameters (Million)=2.42022.03 | 95.78 | |
| LRUWindow size=5, Truncation degree=2, Hanging normalization=true2026.02 | 95.71 | |
| S52026.02 | 95.63 | |
| WaveMix-256/7Number of parameters (Million)=9.62022.03 | 95.56 | |
| ResNet-34Number of parameters (Million)=21.32022.03 | 95.04 | |
| ResNet-18Number of parameters (Million)=11.22022.03 | 94.76 | |
| ResNet-50Number of parameters (Million)=23.62022.03 | 94.64 | |
| µ2NetIteration=2nd, Knowledge Reuse=false2022.05 | 93.68 | |
| µ2NetIteration=1st, Knowledge Reuse=false2022.05 | 93.6 | |
| RFOLHCR2026.02 | 93.21 | |
| Wiyatnoet al.Training set size=Full2019.04 | 91.27 | |
| Log-NCDEUse path signature features=false2026.02 | 90.59 | |
| NCDEUse path signature features=false2026.02 | 89.18 | |
| Cohen et al.Training set size=Full2019.04 | 85.15 |