Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

NIH

Benchmarks

Task NameDataset NameSOTA ResultTrend
Chest X-ray classificationNIH (test)
AUROC87.3
47
Lung segmentationNIH
Recall97.4
16
ClassificationNIH
AUC84.53
9
Pancreas SegmentationNIH
DSC0.8947
8
Out-of-Distribution DetectionNIH
AUF91.5
8
Chest X-ray ClassificationNIH manually re-labelled clean (test)
Pneumothorax AUC89.1
8
Anomaly DetectionNIH clearer (test)
AUC94.6
7
Medical Image ClassificationNIH-8 cross-domain
Macro-AUC72.25
6
Long-context retrievalNIH
Multi-needle Avg Recall100
6
Medical Image ClassificationNIH (100% labeled)
AUC78.7
6
Medical Image ClassificationNIH 10% labeled
AUC71.6
6
Medical Image ClassificationNIH (1% labeled)
AUC0.622
6
Anomaly DetectionNIH AP projection (test)
AUC60.1
6
Anomaly DetectionNIH PA projection (test)
AUC0.708
6
Image ClassificationNIH Radiology X-ray
AUC84.27
5
Long context understandingNIH Multi-needle
Accuracy100
5
Image ClassificationNIH
Accuracy56.4
5
Out-of-Distribution DetectionNIH ID (Xray) vs OOD (Xray)
AUROC0.54
3
Binary ClassificationNIH 10-fold cross-validation local model
Mean F194
2
Showing 19 of 19 rows