A mask-conditioned 3D latent diffusion model synthesizes breast tumors in DCE-MRI, boosting tumor segmentation Dice by up to 3% on a public dataset.
In: 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
SynBT: High-quality Tumor Synthesis for Breast Tumor Segmentation by 3D Diffusion Model
A mask-conditioned 3D latent diffusion model synthesizes breast tumors in DCE-MRI, boosting tumor segmentation Dice by up to 3% on a public dataset.