elasticdeform: Elastic deformations for N-dimensional images

This library implements elastic grid-based deformations for N-dimensional images.

The elastic deformation approach is described in

• Ronneberger, Fischer, and Brox, "U-Net: Convolutional Networks for Biomedical Image Segmentation" (https://arxiv.org/abs/1505.04597)
• Çiçek et al., "3D U-Net: Learning Dense Volumetric Segmentation from Sparse Annotation" (https://arxiv.org/abs/1606.06650)

The procedure generates a coarse displacement grid with a random displacement for each grid point. This grid is then interpolated to compute a displacement for each pixel in the input image. The input image is then deformed using the displacement vectors and a spline interpolation.

In addition to the normal, forward deformation, this package also provides a function that can backpropagate the gradient through the deformation. This makes it possible to use the deformation as a layer in a convolutional neural network. For convenience, a TensorFlow wrapper is provided in elasticdeform.tf.

See https://github.com/gvtulder/elasticdeform