Segmentation of Multimodal Myocardial Images Using Shape-Transfer GAN
This addresses the problem of segmenting pathological myocardium in LGE images for clinical evaluation of infarction regions, representing a domain-specific incremental improvement.
The paper tackled myocardium segmentation in late gadolinium enhancement (LGE) cardiac MR images, which is challenging due to pathological brightness and textures, by proposing a shape-transfer GAN that generates realistic LGE images from easily segmented bSSFP images and segments them without using LGE segmentation labels, achieving accurate masks as tested on the Multi-sequence Cardiac MR Segmentation Challenge dataset.
Myocardium segmentation of late gadolinium enhancement (LGE) Cardiac MR images is important for evaluation of infarction regions in clinical practice. The pathological myocardium in LGE images presents distinctive brightness and textures compared with the healthy tissues, making it much more challenging to be segment. Instead, the balanced-Steady State Free Precession (bSSFP) cine images show clearly boundaries and can be easily segmented. Given this fact, we propose a novel shape-transfer GAN for LGE images, which can 1) learn to generate realistic LGE images from bSSFP with the anatomical shape preserved, and 2) learn to segment the myocardium of LGE images from these generated images. It's worth to note that no segmentation label of the LGE images is used during this procedure. We test our model on dataset from the Multi-sequence Cardiac MR Segmentation Challenge. The results show that the proposed Shape-Transfer GAN can achieve accurate myocardium masks of LGE images.