Multi-class blob and CC losses via one-vs-rest decomposition and per-component weighting improve foreground Dice, rare-class Dice, and Panoptic Quality on BraTS-METS 2025 compared to baseline.
arXiv preprint arXiv:2504.12527 (2025)
3 Pith papers cite this work. Polarity classification is still indexing.
3
Pith papers citing it
fields
cs.CV 3verdicts
UNVERDICTED 3representative citing papers
Contrast-X benchmark and FlowMI model enable synthesis of contrast-enhanced images from arbitrary non-contrast modality inputs using multi-modal flow matching.
Ensembling inpainting models with median filtering, histogram matching, pixel averaging, and lightweight U-Net refinement yields more anatomically plausible and accurate inpainted MRI regions than individual baseline models.
citing papers explorer
-
Instance Awareness of Multi-class Semantic Segmentation Loss Functions
Multi-class blob and CC losses via one-vs-rest decomposition and per-component weighting improve foreground Dice, rare-class Dice, and Panoptic Quality on BraTS-METS 2025 compared to baseline.
-
Contrast-X: A Multi-Modal Contrast Image Synthesis Benchmark and Universal Modality Flow Matching
Contrast-X benchmark and FlowMI model enable synthesis of contrast-enhanced images from arbitrary non-contrast modality inputs using multi-modal flow matching.
-
Post-Processing Methods for Improving Accuracy in MRI Inpainting
Ensembling inpainting models with median filtering, histogram matching, pixel averaging, and lightweight U-Net refinement yields more anatomically plausible and accurate inpainted MRI regions than individual baseline models.