GitHub - sarbjotm/Synthesize-SWI-Type-MRI-Scans: For this project, in a group of five we used Python and trained Adversarial Neural Networks so they could Synthesize SWI-Type MRI Scans The project aimed at reducing the time needed (and the cost) to perform an extensive MRI exam by generating MRI sequences from previously obtained scans of different types through a tailored deep learning model, namely an adversarial neural network. In more detail, the current project will synthesize the SWI sequence from T1w, T2w, and FLAIR. A successful application of the latter would allow medical staff to perform a lower number of sequences (reducing examination time), and to subsequently obtain the remaining needed sequences with software. The libraries used in the project were dominate, fnmatch, getopt, gzip, imageio, nibabel, numpy, os, pandas, pillow, pytorch, scipy, shutil, torchvision version 0.2.0, torchviz, tqdm, unet, and visdom. The deep learning model used for the purpose of this project is known as Multi-Modal Generative Adversarial Network (MM-GAN),

sarbjotm / Synthesize-SWI-Type-MRI-Scans Public

forked from scotchontherocks31/cmpt340project

Notifications You must be signed in to change notification settings
Fork 0
Star 0

For this project, in a group of five we used Python and trained Adversarial Neural Networks so they could Synthesize SWI-Type MRI Scans The project aimed at reducing the time needed (and the cost) to perform an extensive MRI exam by generating MRI sequences from previously obtained scans of different types through a tailored deep learning model,…