Synthesis of Late Gadolinium Enhancement Images via Implicit Neural Representations for Cardiac Scar Segmentation

TL;DR

This paper introduces a novel method combining implicit neural representations and diffusion models to generate synthetic LGE cardiac images and masks, enhancing data augmentation for better fibrosis segmentation.

Contribution

It presents a new framework that synthesizes realistic cardiac images and segmentations without annotations, improving segmentation accuracy through data augmentation.

Findings

Synthetic data improved Dice score from 0.509 to 0.524.

The method effectively captures anatomical details in generated images.

Augmentation with 200 synthetic volumes benefits segmentation performance.

Abstract

Late gadolinium enhancement (LGE) imaging is the clinical standard for myocardial scar assessment, but limited annotated datasets hinder the development of automated segmentation methods. We propose a novel framework that synthesises both LGE images and their corresponding segmentation masks using implicit neural representations (INRs) combined with denoising diffusion models. Our approach first trains INRs to capture continuous spatial representations of LGE data and associated myocardium and fibrosis masks. These INRs are then compressed into compact latent embeddings, preserving essential anatomical information. A diffusion model operates on this latent space to generate new representations, which are decoded into synthetic LGE images with anatomically consistent segmentation masks. Experiments on 133 cardiac MRI scans suggest that augmenting training data with 200 synthetic volumes contributes to improved fibrosis segmentation performance, with the Dice score showing an increase from 0.509 to 0.524. Our approach provides an annotation-free method to help mitigate data scarcity.The code for this research is publicly available.