Medical Image Classification on HAM10000 D5 (test)
99.8AccuracyMAIL-RN50
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MAIL-RN50base network=ResNet50, Params=20.8, Flops=2.182026.02 | 99.8 | 99.6 | 99.8 | |
| MAIL-RN18base network=ResNet18, Params=11.7, Flops=1.842026.02 | 99.4 | 99.4 | 99.7 | |
| DRIFA-NetParams=53.8, Flops=4.832026.02 | 98.2 | 97.9 | 98.5 | |
| MFMSAParams=26.9, Flops=1.42026.02 | 97.9 | 97.3 | 97.9 | |
| MTTU-NetParams=38.1, Flops=6.82026.02 | 97.4 | 96.5 | 97.1 | |
| M3AttParams=183, Flops=12.142026.02 | 95.5 | 94.8 | 95.2 | |
| GloriaParams=30.8, Flops=1.542026.02 | 93.7 | 93.7 | 94.5 | |
| HAMLETParams=57.3, Flops=3.522026.02 | 93.5 | 93.4 | 93.2 | |
| NATParams=20, Flops=1.12026.02 | 93.1 | 92.6 | 93.3 | |
| MuMuParams=56.6, Flops=2.972026.02 | 92.8 | 92.3 | 93.1 | |
| DDA-NetParams=12.1, Flops=1.122026.02 | 92.2 | 91.8 | 92.6 | |
| POTTERParams=12, Flops=0.952026.02 | 91.3 | 91.1 | 91.8 |