Deep MRI Synthesis Collection
Overview
This is the official collection of deep learnning based MRI synthesis methods. Included a various range of MRI modalities including T1 weighted MRI, diffusion weighted MRI (dMRI), Diffusion Tensor Image (DTI), and Diffusion Fiber Orientation Distribution (FOD).
Our Works
[Publication]
[Code]Zihao Tang*, Xinyi Wang*, Lihaowen Zhu, Mariano Cabezas, Dongnan Liu, Michael Barnett, Weidong Cai, and Chenyu Wang.
Zihao Tang*, Xinyi Wang*, Mariano Cabezas, Arkiev D’Souza, Fernando Calamante, Dongnan Liu, Michael Barnett, Chenyu Wang, and Weidong Cai.
Zihao Tang, Mariano Cabezas, Dongnan Liu, Michael Barnett, Weidong Cai, and Chenyu Wang.
Highlights
OCE-Net (Diffusion MRI Fibre Orientation Distribution Inpainting)
- Present a Order-wise Coefficient Estimation Network to inpaint the disrupted Fibre Orientation Distribution (FOD) coefficients.
- Order-wise Coefficient Decoders to decode the coefficient independently.
- Explore the analysis on connectome.
LG-Net
- Present an end-to-end solution for Multiple Sclerosis (MS) lesion inpainting on T1 MRI images.
- A specifically designed lesion gate consistency.
- State-of-the-art on computer vision similaritiy metrics and brain tissue volumetric metrics.
BibTeX
If you find our data or project useful in your research, please cite:
@inproceedings{tang2021lg,
title="LG-Net: Lesion Gate Network for Multiple Sclerosis Lesion Inpainting",
author="Tang, Zihao and Cabezas, Mariano and Liu, Dongnan and Barnett, Michael and Cai, Weidong and Wang, Chenyu",
booktitle={International Conference on Medical Image Computing and Computer-Assisted Intervention (MICCAI)},
pages="660--669",
year={2021},
organization={Springer}
}