A Deep Learning Based Method for Fast Registration of Cardiac Magnetic Resonance Images

2025-06-23

AI Summary

Cardiac MR image registration faces three key challenges: absence of ground-truth deformation fields, inherent ill-posedness due to non-unique solutions, and high computational cost and hardware demands of existing unsupervised methods. To address these, we propose FLIR—a lightweight, end-to-end unsupervised convolutional network built upon a whole-volume registration architecture with highly efficient convolutional designs to minimize computational overhead. FLIR enables accurate myocardial strain quantification without requiring ground-truth deformation labels. It achieves inference speeds markedly superior to mainstream unsupervised methods on mid- to low-end GPUs, while maintaining registration accuracy comparable to state-of-the-art approaches. Furthermore, FLIR demonstrates high inter-scan consistency in strain measurements, fulfilling clinical requirements for real-time quantitative analysis.

Technology Category

• Computer Vision: Medical and Biological Imaging
• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: Multimodal Perception & Sensor Fusion

Application Category

• Search and Retrieval-Augmented AI: Web evaluation methodologies and metrics
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization

Abstract

Image registration is used in many medical image analysis applications, such as tracking the motion of tissue in cardiac images, where cardiac kinematics can be an indicator of tissue health. Registration is a challenging problem for deep learning algorithms because ground truth transformations are not feasible to create, and because there are potentially multiple transformations that can produce images that appear correlated with the goal. Unsupervised methods have been proposed to learn to predict effective transformations, but these methods take significantly longer to predict than established baseline methods. For a deep learning method to see adoption in wider research and clinical settings, it should be designed to run in a reasonable time on common, mid-level hardware. Fast methods have been proposed for the task of image registration but often use patch-based methods which can affect registration accuracy for a highly dynamic organ such as the heart. In this thesis, a fast, volumetric registration model is proposed for the use of quantifying cardiac strain. The proposed Deep Learning Neural Network (DLNN) is designed to utilize an architecture that can compute convolutions incredibly efficiently, allowing the model to achieve registration fidelity similar to other state-of-the-art models while taking a fraction of the time to perform inference. The proposed fast and lightweight registration (FLIR) model is used to predict tissue motion which is then used to quantify the non-uniform strain experienced by the tissue. For acquisitions taken from the same patient at approximately the same time, it would be expected that strain values measured between the acquisitions would have very small differences. Using this metric, strain values computed using the FLIR method are shown to be very consistent.

Problem

• Fast registration of cardiac MRI for motion tracking
• Unsupervised deep learning with efficient convolution architecture
• Accurate strain quantification in dynamic heart tissues

Innovation

• Fast volumetric registration model for cardiac strain
• Efficient convolution architecture for quick inference
• Lightweight FLIR model ensures consistent strain values