Image Classification on EMNIST balanced mixed noise (test)
89.32AccuracyAddUNet–MTL (7–7–7–7–7)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AddUNet–MTL (7–7–7–7–7)Condition=Mixed Noise2026.02 | 89.32 | 1.94 | |
| AddUNet–MTL (9–9–9–9–9)Condition=Mixed Noise2026.02 | 89.19 | 1.32 | |
| CNN (5–3–3)Condition=Mixed Noise2026.02 | 87.62 | 2.83 | |
| MLP (2-layer, 256)Condition=Mixed Noise2026.02 | 81.94 | 3.03 |