ICON

ICON (Inverse COnsistent RegistratioN) is a non-parametric deep learning registration approach which relies only on inverse consistency for regularity. As the regularization neither involves explicit smoothing or a penality on spatial derivatives no affine pre-registration is required. Hence, ICON is largely parameter free. It also support ITK images as well as ITK transforms so is easy to use. We currently provide a pre-trained ICON registration network for the DESS magnetic resonance knee images of the Osteoarthritis Initiative.

The ICON repository can be found here: https://github.com/uncbiag/ICON