Intensity-Based Wasserstein Distance As A Loss Measure For Unsupervised Deformable Deep Registration.
Roozbeh ShamsWilliam LeAdrien WeihsSamuel KadouryPublished in: ISBI (2021)
Keyphrases
- image registration
- distance measure
- medical image registration
- dissimilarity measure
- deformable shapes
- pointwise
- optimal matching
- multi modality
- multiresolution
- kullback leibler divergence
- similarity measure
- distance function
- mutual information
- supervised learning
- pointwise mutual information
- affine transformation
- deep architectures
- weighted sum
- registration errors
- registration accuracy
- nonrigid registration
- kl divergence
- gain ratio
- closest point
- knn
- deformable surface
- registration process
- deformable models
- unsupervised manner
- euclidean distance
- range images
- affine invariant
- image pairs
- point cloud