Probabilistic dense displacement network for weakly-supervised medical image registration uses approximate min-convolutions and mean field inference to achieve state-of-the-art accuracy on inter-patient abdominal CT, outperforming prior deep learning by 15% Dice overlap.
Medical image analysis 36, 61–78 (2017)
2 Pith papers cite this work. Polarity classification is still indexing.
2
Pith papers citing it
years
2019 2verdicts
UNVERDICTED 2representative citing papers
Nonlinear reflective registration to extract contralateral patches improves Dice scores by 9-13 points when added as features to two CNN architectures for ischemic stroke lesion segmentation on the ISLES 2015 SISS training set.
citing papers explorer
-
Closing the Gap between Deep and Conventional Image Registration using Probabilistic Dense Displacement Networks
Probabilistic dense displacement network for weakly-supervised medical image registration uses approximate min-convolutions and mean field inference to achieve state-of-the-art accuracy on inter-patient abdominal CT, outperforming prior deep learning by 15% Dice overlap.
-
Exploiting bilateral symmetry in brain lesion segmentation
Nonlinear reflective registration to extract contralateral patches improves Dice scores by 9-13 points when added as features to two CNN architectures for ischemic stroke lesion segmentation on the ISLES 2015 SISS training set.