Commit 6fb204da authored by cc215's avatar cc215 💬
Browse files

updaye readme

parent c74d1f3a
......@@ -38,14 +38,14 @@ Results will be saved under `test_results` by default.
| | UKBB test (600) | | | ACDC (100) | | | M&Ms (150) | | |
|- |:-: |:-: |:-: |:-: |:-: |:-: |:-: |:-: |:-: |
| configurations | LV | MYO | RV | LV | MYO | RV | LV | MYO | RV |
| bs=1, roi size = 256, z_score | 0.9383 | 0.8780 | 0.8979 | 0.8940 | 0.8034 | 0.8237 | 0.8862 | 0.7889 | 0.8168 |
| bs=1, roi size = 192, z_score | 0.9371 | 0.8775 | 0.8962 | 0.8891 | 0.7981 | 0.8103 | 0.8725 | 0.7716 | 0.7954 |
| bs=-1, roi size = 192, z_score | 0.9139 | 0.8388 | 0.8779 | 0.8790 | 0.7818 | 0.8069 | 0.8679 | 0.7675 | 0.7926 |
| batch_size = 1, roi size = 256, z_score | 0.9383 | 0.8780 | 0.8979 | 0.8940 | 0.8034 | 0.8237 | 0.8862 | 0.7889 | 0.8168 |
| batch_size = 1, roi size = 192, z_score | 0.9371 | 0.8775 | 0.8962 | 0.8891 | 0.7981 | 0.8103 | 0.8725 | 0.7716 | 0.7954 |
| batch_size = -1 (use the whole volume as batch), roi size = 192, z_score | 0.9139 | 0.8388 | 0.8779 | 0.8790 | 0.7818 | 0.8069 | 0.8679 | 0.7675 | 0.7926 |
The model params are stored in `./checkpoints/Unet_LVSA_trained_from_UKBB.pkl`
For optimal performance at deployment, please set it to instance normalization (--batch_size 1) and crop images to (--roi_size 256).
e.g.
- If we want to predict test images from UKBB, we can simply ran
- If we want to predict test images from UKBB, we can simply run the following command:
`python predict.py --sequence LVSA \
--model_path './checkpoints/Unet_LVSA_trained_from_UKBB.pkl' \
--root_dir '/vol/medic02/users/wbai/data/cardiac_atlas/UKBB_2964/sa/test' \
......
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment