The Topological Confound in Semi-Supervised Medical Image Segmentation: Measurement, Theory, and Two Corrections

Résumé fourni par la source

Code, per-image results, pre-registration record and the topology-aware HD95 reporting utility for the paper "The Topological Confound in Semi-Supervised Medical Image Segmentation: Measurement, Theory, and Two Corrections". Includes the S2BC dual-branch semi-supervised trainer, the CCP training-time correction, the CCF conformal component filter, the arm-table and census evaluators, and the derived-table scripts that regenerate every number in the paper from the released per-image scores.