Lung Ultrasound Image Classification on POCUS (test)
92.2Overall AccuracyOurs
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| OursBackbone=ResNet-50, Evaluation Protocol=Fine-tuned, Pre-training Dataset=Butterfly dataset2022.07 | 92.2 | 89.2 | 95.1 | 93.1 | |
| USCLBackbone=ResNet-50, Evaluation Protocol=Fine-tuned, Pre-training Dataset=Butterfly dataset2022.07 | 90.7 | 86.1 | 90.3 | 93.5 | |
| SimCLRBackbone=ResNet-50, Evaluation Protocol=Fine-tuned, Pre-training Dataset=Butterfly dataset2022.07 | 86.4 | 83.2 | 89.4 | 87.1 | |
| MoCo v2Backbone=ResNet-50, Evaluation Protocol=Fine-tuned, Pre-training Dataset=Butterfly dataset2022.07 | 84.8 | 79.7 | 81.4 | 88.9 | |
| PretrainedBackbone=ResNet-50, Evaluation Protocol=Fine-tuned, Pre-training Dataset=ImageNet2022.07 | 84.2 | 79.5 | 78.6 | 88.6 |