Deep Group-wise Variational Diffeomorphic Image Registration
This work addresses the need for efficient multi-image registration in medical imaging, such as for intra-patient MRI and CT exams, though it is incremental as it builds upon existing pair-wise methods.
The paper tackles the problem of extending learning-based image registration to handle multiple images simultaneously, achieving competitive performance in terms of image similarity and landmark distances with significantly faster registration compared to existing methods like Elastix and VoxelMorph.
Deep neural networks are increasingly used for pair-wise image registration. We propose to extend current learning-based image registration to allow simultaneous registration of multiple images. To achieve this, we build upon the pair-wise variational and diffeomorphic VoxelMorph approach and present a general mathematical framework that enables both registration of multiple images to their geodesic average and registration in which any of the available images can be used as a fixed image. In addition, we provide a likelihood based on normalized mutual information, a well-known image similarity metric in registration, between multiple images, and a prior that allows for explicit control over the viscous fluid energy to effectively regularize deformations. We trained and evaluated our approach using intra-patient registration of breast MRI and Thoracic 4DCT exams acquired over multiple time points. Comparison with Elastix and VoxelMorph demonstrates competitive quantitative performance of the proposed method in terms of image similarity and reference landmark distances at significantly faster registration.