ngohgia / 3d-convolutional-autoencoder-for-fmri-volumes

Learning spatial and temporal features of fMRI brain images.

3D convolutional autoencoder for fmri volumes

This repository stores the code supporting my article- "3D Convolutional autoencoder for brain volumes: Learning spatial and temporal features of fMRI brain images".

The storeMaskV.py file contains the code to part one of the article, the testfmri.py file contains the code to part two of the article, the DI_andConvertTo_Tf.py file contain code used to download the fMRI files from the cloud and to convert it into a single tfrecords file and the 'data_loss_check_code_snippet.txt' file contain a code snippet presented in the "Theory" section of the article.

Concepts mentioned in the first two parts of the article are implemented sequentially in their respective code files. Comments are provided wherever necessary. Please see the imports for dependencies.