Deformable Registration of Brain MR Images via a Hybrid Loss

TL;DR

This paper introduces a hybrid loss function for unsupervised deformable registration of brain MR images, improving alignment accuracy by integrating multiple image characteristics.

Contribution

It proposes a novel hybrid loss that combines intensity, statistical, and boundary information for better registration accuracy.

Findings

Achieves high registration accuracy on the OASIS dataset.

Preserves deformation smoothness during registration.

Outperforms existing intensity-based methods.

Abstract

Unsupervised learning strategy is widely adopted by the deformable registration models due to the lack of ground truth of deformation fields. These models typically depend on the intensity-based similarity loss to obtain the learning convergence. Despite the success, such dependence is insufficient. For the deformable registration of mono-modality image, well-aligned two images not only have indistinguishable intensity differences, but also are close in the statistical distribution and the boundary areas. Considering that well-designed loss functions can facilitate a learning model into a desirable convergence, we learn a deformable registration model for T1-weighted MR images by integrating multiple image characteristics via a hybrid loss. Our method registers the OASIS dataset with high accuracy while preserving deformation smoothness.