FMRI Image Extrapolation

Background

This project was done as part of ASTR 596 (Machine Learning) where we were tasked with taking on any project we want to as long as it implemented some sort of machine learning algorithm. The project I decided to undertake was to see if I could setup some code to take in the output of an FMRI scan of someone’s brain in response to visual stimulus and determine what that stimulus was.

How it Works

The Jupyter notebook describes it better but ultimately I used a from-scratch built genetic algorithm to do the training and testing whhich creates a number of “organisms” which are then tested against the data with a number of the best performing allowed to move on, those are then mutated to create the next population which is tested and so on.

Link

Home
Education
Laboratory Experience
Projects
Publications and Presentations
Work Experience
Journal