lakshmi-sathi / mri_reconstruction

MRI Reconstruction. Methodology to score effectiveness of loss metrics. Incorporation of Edge Loss for boosting edges in reconstruction.

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

mri_reconstruction

The stratergy followed is to setup the baseline FastMRI network and then setup a methodology to compare effectiveness of different loss metrics like SSIM, MSE, NMSE, PSNR through assignment of visual ranking for each of them.

The set considered for it (23 ground-truth vs subsampled pairs): MRI_Comp1 mriComp2

For the methodology refer the PAPER

The scoring based on comparison with visual ground-truth ranking, is done using this script

Seperate to this, concept of edge loss is also incorporated. EdgeDetectionFirstResult

For visualization of fastMRI data for the above purposes this script is used script was used.

The paper is available HERE

About

MRI Reconstruction. Methodology to score effectiveness of loss metrics. Incorporation of Edge Loss for boosting edges in reconstruction.


Languages

Language:Python 64.9%Language:Perl 35.1%