Nesterov Accelerated ADMM for Fast Diffeomorphic Image Registration

TL;DR

This paper introduces a Nesterov-accelerated ADMM method for diffeomorphic image registration that combines the accuracy of iterative approaches with near real-time speed, outperforming existing methods in accuracy and efficiency.

Contribution

It proposes a novel accelerated ADMM algorithm integrated with Nesterov gradient descent for fast, accurate, and diffeomorphic 3D image registration, especially for cardiac MRI.

Findings

Achieved sub-2 second runtime on GPU for 3D cardiac MRI registration.

Outperformed state-of-the-art deep learning and iterative methods in accuracy.

Produced strictly diffeomorphic deformations with high registration quality.

Abstract

Deterministic approaches using iterative optimisation have been historically successful in diffeomorphic image registration (DiffIR). Although these approaches are highly accurate, they typically carry a significant computational burden. Recent developments in stochastic approaches based on deep learning have achieved sub-second runtimes for DiffIR with competitive registration accuracy, offering a fast alternative to conventional iterative methods. In this paper, we attempt to reduce this difference in speed whilst retaining the performance advantage of iterative approaches in DiffIR. We first propose a simple iterative scheme that functionally composes intermediate non-stationary velocity fields to handle large deformations in images whilst guaranteeing diffeomorphisms in the resultant deformation. We then propose a convex optimisation model that uses a regularisation term of arbitrary order to impose smoothness on these velocity fields and solve this model with a fast algorithm that combines Nesterov gradient descent and the alternating direction method of multipliers (ADMM). Finally, we leverage the computational power of GPU to implement this accelerated ADMM solver on a 3D cardiac MRI dataset, further reducing runtime to less than 2 seconds. In addition to producing strictly diffeomorphic deformations, our methods outperform both state-of-the-art deep learning-based and iterative DiffIR approaches in terms of dice and Hausdorff scores, with speed approaching the inference time of deep learning-based methods.