MvMM-RegNet: A new image registration framework based on multivariate mixture model and neural network estimation

TL;DR

This paper introduces MvMM-RegNet, a novel image registration framework combining multivariate mixture models and neural networks, capable of groupwise registration and effective in multimodal cardiac image segmentation.

Contribution

It presents a new generative model-based loss function for groupwise registration, extending deep learning registration to multiple images and modalities.

Findings

Achieved high Dice scores in cardiac segmentation tasks.

Demonstrated effectiveness on multiple publicly available datasets.

Outperformed existing methods in multimodal registration accuracy.

Abstract

Current deep-learning-based registration algorithms often exploit intensity-based similarity measures as the loss function, where dense correspondence between a pair of moving and fixed images is optimized through backpropagation during training. However, intensity-based metrics can be misleading when the assumption of intensity class correspondence is violated, especially in cross-modality or contrast-enhanced images. Moreover, existing learning-based registration methods are predominantly applicable to pairwise registration and are rarely extended to groupwise registration or simultaneous registration with multiple images. In this paper, we propose a new image registration framework based on multivariate mixture model (MvMM) and neural network estimation. A generative model consolidating both appearance and anatomical information is established to derive a novel loss function capable of implementing groupwise registration. We highlight the versatility of the proposed framework for various applications on multimodal cardiac images, including single-atlas-based segmentation (SAS) via pairwise registration and multi-atlas segmentation (MAS) unified by groupwise registration. We evaluated performance on two publicly available datasets, i.e. MM-WHS-2017 and MS-CMRSeg-2019. The results show that the proposed framework achieved an average Dice score of $0.871\pm 0.025$ for whole-heart segmentation on MR images and $0.783\pm 0.082$ for myocardium segmentation on LGE MR images.