SOUP-GAN:

Super-resolution interpolation tool for medical images.

This inference code supports:

1. Thick-slices to thin-slices SR interpolation with arbitrarily user-selected sampling ratios (e.g., from 2x to 6x).
2. Medical imaging mask inerpolation.

SOUP-GAN: Super-Resolution MRI Using Generative Adversarial Networks. Tomography 2022, 8(2), 905-919

Kuan Zhang, Haoji Hu, Kenneth Philbrick, Gian Marco Conte, Joseph D. Sobek, Pouria Rouzrokh, Bradley J. Erickson

Usage:

Install the package KevinSR.

from KevinSR import mask_interpolation, SOUP_GAN

# for SR image interp (prep_type=0: thick-to-thin; 1: thin-to-thin)

thin_slices = SOUP_GAN(thick_slices, factor, prep_type)

# for mask interp

new_masks = mask_interpolation(masks, factor)

Example:

Example_0 (thick_to_thin):

Example_1 (sparse_to_thin):

Mask_interpolation: