Meeting Banner
Abstract #4121

Motion Correction for Myocardial T1 Mapping using Self-supervised Deep Learning Registration with Contrast Separation

Yuze Li1, Chunyan Wu1, Haikun Qi2, Dongyue Si1, Haiyan Ding1, and Huijun Chen1
1Center for Biomedical Imaging Research, School of Medicine, Tsinghua University, Beijing, China, 2School of Biomedical Engineering, ShanghaiTech University, Shanghai, China


A motion correction method for myocardial T1 mapping using Self-supervised Deep learning based Registration with contrAst seParation (SDRAP) was proposed. A sparse coding based method was firstly proposed to separate the contrast component from T1w images. Then, a self-supervised deep neural network was developed to register contrast separated images, followed by the signal fitting to generate motion corrected T1 maps. Models were trained and tested in 47 healthy volunteers using MOLLI sequence, and compared with Free Form Deformation method. Results showed the proposed method can achieve better performance in registration and T1 mapping with higher efficiency (7x acceleration).

This abstract and the presentation materials are available to members only; a login is required.

Join Here