Breakage-Aware clDice
Cross-source consensus on Breakage-Aware clDice from 1 sources and 4 claims.
1 sources · 4 claims
How it works
Benefits
Highlighted claims
- The breakage-aware clDice term uses group convolution to compute local neighborhood sums for each class in predictions and ground truth. — AG-TAL: Anatomically-Guided Topology-Aware Loss for Multiclass Segmentation of the Circle of Willis Using Large-Scale Multi-Center Datasets
- Breakage-aware clDice avoids expensive per-class skeletonization by using a unified foreground soft skeleton. — AG-TAL: Anatomically-Guided Topology-Aware Loss for Multiclass Segmentation of the Circle of Willis Using Large-Scale Multi-Center Datasets
- The method uses an L1 neighborhood similarity error map to highlight breakages and false positives while reducing emphasis on minor boundary mismatch. — AG-TAL: Anatomically-Guided Topology-Aware Loss for Multiclass Segmentation of the Circle of Willis Using Large-Scale Multi-Center Datasets
- Adding breakage-aware clDice improved small and medium artery performance in the ablation. — AG-TAL: Anatomically-Guided Topology-Aware Loss for Multiclass Segmentation of the Circle of Willis Using Large-Scale Multi-Center Datasets