DRMIME: Differentiable Mutual Information and Matrix Exponential for Multi-Resolution Image Registration
Published in
International Conference on Medical Imaging wit...(2020)External Links:
TL
TL;DR
The novelty here is that rather than using traditional ways of approximating MI, this work uses a neural estimator called MINE and supplement it with matrix exponential for transformation matrix computation, which leads to improved results as compared to the standard algorithms available out-of-the-box in state-of the-art image registration toolboxes.
Abstract
In this work, we present a novel unsupervised image registration algorithm. It is differentiable end-to-end and can be used for both multi-modal and mono-modal registration. This is done using mutual information (MI) as a metric. The novelty here is that rather than using traditional ways of approximating MI, we use a neural estimator called MINE and supplement it with matrix exponential for transformation matrix computation. This leads to improved results as compared to the standard algorithms available out-of-the-box in state-of-the-art image registration toolboxes.
References29 items
1
PyTorch: An Imperative Style, High-Performance Deep Learning Library
2
AirLab: Autograd Image Registration Laboratory
3
MINE: Mutual Information Neural Estimation
4
FIRE: Fundus Image Registration dataset
5