Train and evaluate histology models on STHELAR with a leave-one-tissue-out (LOTO) protocol. Demonstrate that a domain-generalization (DG) approach reduces the performance collapse on unseen organs compared to a baseline (CellViT/ViT/UNet).
-
Notifications
You must be signed in to change notification settings - Fork 0
Cross Tissue Generalization using the LOTO (Leave One Tissue Out) approach on hematoxylin and eosin stained tissue samples. Using the STHELAR dataset created by MICS-Lab at Paris-Saclay University.
License
g4nesh/HExt
About
Cross Tissue Generalization using the LOTO (Leave One Tissue Out) approach on hematoxylin and eosin stained tissue samples. Using the STHELAR dataset created by MICS-Lab at Paris-Saclay University.
Topics
Resources
License
Stars
Watchers
Forks
Releases
No releases published
Packages 0
No packages published