$ git clone https://github.com/lunit-io/mmg-nia
$ pip install -r requirements.txt
$ cd data_preprocessing
and do data-preprocessing here- To train and test 5-fold cross validation, use
$ sh test.sh $GPU_ID $PICKLE_PATH $DATA_ROOT
e.g.$ sh test.sh 0 data_preprocessing/db/shuffled_db.pkl /data/mmg/mg_nia
- If you want to use many GPUs, input multiple numbers:
sh test.sh 0,1,2,3, ...
- If you want to use many GPUs, input multiple numbers:
$ cat resnet34-5fold-result
threshold : 0.1
calculated accuracy is 0.8144577092389047
calculated specificity is 0.8112627121478205
calculated sensitivity is 0.825194007255318
threshold : 0.15
calculated accuracy is 0.8477143885489631
calculated specificity is 0.8706655574469052
calculated sensitivity is 0.768865680293087
threshold : 0.2
calculated accuracy is 0.8648696254049115
calculated specificity is 0.9067113501876425
calculated sensitivity is 0.7214681631914128
calculated auc is 0.9070296037910798
$ cat densenet121-5fold-result
threshold : 0.1
calculated accuracy is 0.8379339405852875
calculated specificity is 0.8391845548624011
calculated sensitivity is 0.8327769440438639
threshold : 0.15
calculated accuracy is 0.8601385974141034
calculated specificity is 0.8777076140580005
calculated sensitivity is 0.7991593433754576
threshold : 0.2
calculated accuracy is 0.872767241617985
calculated specificity is 0.9012399299294076
calculated sensitivity is 0.7745227785230062
calculated auc is 0.9193286916314279