Swin UNETR

Reported on 6 benchmarks across 3 tasks · 1 paper · 2 SOTA

Note: results are matched by exact model name. Different papers may use the same name for different model variants.

Medical4 results

Medical Image SegmentationonMedical Segmentation Decathlon
Dice (Average)· 2021-11-29
78.68
SOTA
Self-Supervised Pre-Training of Swin Transformers for 3D Medical Image Analysis arXiv:2111.14791
Medical Image SegmentationonMedical Segmentation Decathlon
NSD· 2021-11-29
89.28
SOTA
Self-Supervised Pre-Training of Swin Transformers for 3D Medical Image Analysis arXiv:2111.14791
Semantic SegmentationonLombardia Sentinel-2 Image Time Series for Crop Mapping
Overall Accuracy
79.64
best: 80.77 (UNet3D)
Semantic SegmentationonMunich Sentinel2 Crop Segmentation
Overall Accuracy
95.26

Audio2 results

10-shot image generationonLombardia Sentinel-2 Image Time Series for Crop Mapping
Overall Accuracy
79.64
best: 80.77 (UNet3D)
10-shot image generationonMunich Sentinel2 Crop Segmentation
Overall Accuracy
95.26