FDRN: A Fast Deformable Registration Network for Medical Images

TL;DR

FDRN is a fast, memory-efficient convolutional neural network designed for accurate deformable registration of medical images, leveraging deep supervision, residual learning, and segmentation priors to outperform existing methods.

Contribution

The paper introduces a novel fast registration network with a coarse-to-fine learning strategy, auxiliary segmentation loss, and efficient architecture improvements for medical image registration.

Findings

FDRN outperforms state-of-the-art methods on brain MR images.

The multi-label segmentation loss improves registration accuracy without extra memory cost.

The framework is generalizable to various medical images and anatomies.

Abstract

Deformable image registration is a fundamental task in medical imaging. Due to the large computational complexity of deformable registration of volumetric images, conventional iterative methods usually face the tradeoff between the registration accuracy and the computation time in practice. In order to boost the registration performance in both accuracy and runtime, we propose a fast convolutional neural network. Specially, to efficiently utilize the memory resources and enlarge the model capacity, we adopt additive forwarding instead of channel concatenation and deepen the network in each encoder and decoder stage. To facilitate the learning efficiency, we leverage skip connection within the encoder and decoder stages to enable residual learning and employ an auxiliary loss at the bottom layer with lowest resolution to involve deep supervision. Particularly, the low-resolution auxiliary loss is weighted by an exponentially decayed parameter during the training phase. In conjunction with the main loss in high-resolution grid, a coarse-to-fine learning strategy is achieved. Last but not least, we introduce an auxiliary loss based on the segmentation prior to improve the registration performance in Dice score. Comparing to the auxiliary loss using average Dice score, the proposed multi-label segmentation loss does not induce additional memory cost in the training phase and can be employed on images with arbitrary amount of categories. In the experiments, we show FDRN outperforms the existing state-of-the-art registration methods for brain MR images by resorting to the compact network structure and efficient learning. Besides, FDRN is a generalized framework for image registration which is not confined to a particular type of medical images or anatomy.