MedicalNet3D 醫(yī)療影像預(yù)訓(xùn)練模型
MedicalNet 是一個(gè) 3D 醫(yī)療影像預(yù)訓(xùn)練模型，其將多個(gè) 3D 醫(yī)療數(shù)據(jù)集集合成大數(shù)據(jù)集，基于此數(shù)據(jù)集提供了完整的 3D-ResNet 系列預(yù)訓(xùn)練模型與相應(yīng)的遷移學(xué)習(xí)訓(xùn)練代碼。
適用場(chǎng)景
-
MedicalNet 提供的預(yù)訓(xùn)練網(wǎng)絡(luò)可遷移到任何 3D 醫(yī)療影像的 AI 應(yīng)用中，包括但不限于分割、檢測(cè)與分類等任務(wù)。
-
尤其適用小數(shù)據(jù)醫(yī)療影像 AI 場(chǎng)景，能加快網(wǎng)絡(luò)收斂，提升網(wǎng)絡(luò)性能。
特性
-
支持單卡與多卡訓(xùn)練
-
支持任何 3D 醫(yī)療影像的遷移學(xué)習(xí)任務(wù)
-
支持 3D-ResNet 全系列模型的遷移學(xué)習(xí)任務(wù)
-
支持多種感受野的遷移學(xué)習(xí)
環(huán)境依賴
- Python 3.7.0
- PyTorch-0.4.1
- CUDA Version 9.0
- CUDNN 7.0.5
性能
| Network | Pretrain | LungSeg(Dice) | NoduleCls(accuracy) |
|---|---|---|---|
| 3D-ResNet10 | Train from scratch | 71.30% | 79.80% |
| MedicalNet | 87.16% | 86.87% | |
| 3D-ResNet18 | Train from scratch | 75.22% | 80.80% |
| MedicalNet | 87.26% | 88.89% | |
| 3D-ResNet34 | Train from scratch | 76.82% | 83.84% |
| MedicalNet | 89.31% | 89.90% | |
| 3D-ResNet50 | Train from scratch | 71.75% | 84.85% |
| MedicalNet | 93.31% | 89.90% | |
| 3D-ResNet101 | Train from scratch | 72.10% | 81.82% |
| MedicalNet | 92.79% | 90.91% | |
| 3D-ResNet152 | Train from scratch | 73.29% | 73.74% |
| MedicalNet | 92.33% | 90.91% | |
| 3D-ResNet200 | Train from scratch | 71.29% | 76.77% |
| MedicalNet | 92.06% | 90.91% |
評(píng)論
圖片
表情