An Unsupervised Learning Approach to Discontinuity-Preserving Image Registration

Abstract

Most traditional image registration algorithms aimed at aligning a pair of images impose well-established regularizers to guarantee smoothness of unknown deformation fields. Since these methods assume global smoothness within the image domain, they pose issues for scenarios where local discontinuities are expected, such as the sliding motion between the lungs and the chest wall during the respiratory cycle. Furthermore, an objective function must be optimized for each given pair of images, thus registering multiple sets of images become very time-consuming and scale poorly to higher resolution image volumes.Using recent advances in deep learning, we propose an unsupervised learning-based image registration model. The model is trained over a loss function with a custom regularizer that preserves local discontinuities, while simultaneously respecting the smoothness assumption in homogeneous regions of image volumes. Qualitative and quantitative validations on 3D pairs of lung CT datasets will be presented.

A Deep Discontinuity-Preserving Image Registration Network

“…While, VM-Dice(img+seg) is a combination of the previous two methods. We did not compare with the DL-based discontinuity-preserving method proposed in [10], as there is no corresponding source code publicly available. This strategy to register different sub-regions and compose corresponding deformation fields is also applicable to the aforementioned networks.…”

“…In contrast, only one study thus far has proposed a discontinuous DL-based image registration framework. Ng et al [10] proposed a custom discontinuity-preserving regulariser on the deformation fields (used with a typical unsupervised registration network), to preserve discontinuities, while ensuring local smoothness within specific regions. They formulated a regularisation term based on the unsigned area of the parallelogram spanned by two displacement vectors associated with moving image voxels.…”

Evaluating Registrations of Serial Sections With Distortions of the Ground Truths

“…Basically, any kinds of damage can be assessed with masks for the repair, similar to the masks we use here to simulate the damage. Some further recent works assess cracks and discontinuities [154,155].…”

A discontinuity-preserving regularization for deep learning-based cardiac image registration

“…Unfortunately, these traditional methods are slow in speed, and now the mainstream of image registration is based on deep learning. Ng and Ebrahimi (2020) proposed a regularization capable of preserving local discontinuities to cope with sliding motion based on a modified U-Net (Ronneberger et al 2015). The central idea of this regularization is to compute the unsigned area of the parallelogram spanned by two arbitrary vectors.…”