Thyroid nodule classification on Hospital Ultrasound dataset (test)
90.19AccuracyDeblurring MAE [Gaussian]
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| Deblurring MAE [Gaussian]Architecture=ViT-B, Pretraining=Ultrasound, Blurring Operation=Gaussian (sigma=1.1)2023.06 | 90.19 | 88.48 | 96.08 | |
| Deblurring MAE [SRAD]Architecture=ViT-B, Pretraining=Ultrasound, Blurring Operation=SRAD (N=40, t=0.1)2023.06 | 90.07 | 88.13 | 95.87 | |
| MAEArchitecture=ViT-B, Pretraining=Ultrasound2023.06 | 89.45 | 87.54 | 95.54 | |
| Zhou et al.Training Strategy=Supervised2023.06 | 88.15 | 86.09 | 94.17 | |
| ConvNextArchitecture=ConvNeXt-L, Pretraining=ImageNet2023.06 | 87.76 | 85.47 | 93.22 | |
| Wang et al.Training Strategy=Supervised2023.06 | 87.44 | 85.16 | 93.11 | |
| Swin TransformerArchitecture=Swin-L, Pretraining=ImageNet2023.06 | 87.43 | 84.92 | 92.83 | |
| MAEArchitecture=ViT-B, Pretraining=ImageNet2023.06 | 87.25 | 85.23 | 93.71 | |
| MoCo v3Architecture=ResNet-50, Pretraining=Ultrasound2023.06 | 87.08 | 84.55 | 92.95 | |
| MoCo v3Architecture=ResNet-50, Pretraining=ImageNet2023.06 | 86.96 | 84.48 | 92.77 | |
| ViTArchitecture=ViT-B, Pretraining=ImageNet2023.06 | 86.38 | 84.17 | 92.69 | |
| SimCLRArchitecture=ResNet-50, Pretraining=ImageNet2023.06 | 86.21 | 83.81 | 92.16 | |
| ResNetArchitecture=ResNet-101, Pretraining=Supervised2023.06 | 86.06 | 83.18 | 91.96 | |
| ViTArchitecture=ViT-B, Pretraining=Supervised2023.06 | 80.6 | 76.98 | 83.89 | |
| Denoising MAEArchitecture=ViT-B, Pretraining=Ultrasound2023.06 | 80.38 | 77.99 | 84.38 |