Working with BibTeX in Python - Biblib
During a PhD, one can read and store hundreds of papers. There is always a need to keep the papers well organized and kept, and also to have a way to easily transfer the citations of the papers to a manuscript.
Gerard Martí Juan
Medical imaging & machine learning
I build AI and machine learning systems for medicine. Right now that means training foundations models at Eden as a Senior AI Researcher, and still collaborating with the UPF BCN MedTech group in Barcelona on fetal Brain MRI. Before that, I was a Ph.D. student at the UPF BCN MedTech group, where I worked on machine learning methods for multimodal, longitudinal analysis of Alzheimer's disease using MRI.
Writing
Notes to myself from the Ph.D. years, published in case they save someone else an afternoon. Nothing new here since 2019.
Removing backgrounds of images in Inkscape
Earlier this week, I was using Freesurfer to generate some nice brain cortex images to add them to a paper. My go-to program to create scientific figures is Inkscape, mainly because it is free, relatively easy to use, and there are lot of nice guides online. However, the problem with this kind of software is that sometimes, simple actions can actually only be done with workarounds or with unintuitive tools. This small blog post is a testament to one of those times.
BrainFortLib - Neuroimaging scripts hub
Processing images and scans of the brain has been a big part of my PhD activity. When I started, I had never worked with brain images before, neither with neuroimaging tools and libraries. I was given a set of scripts that my supervisor had used for processing brain images, but apart from that I had no other support: nobody else in my department worked on brain imaging.
Guide to use Jupyter in an HPC environment
In this small blog post I will detail how to use Jupyter applications (notebooks and JupyterLab) from a computing node in a HPC. This post is a follow-up to the previous post.
Creation of a custom Python environment in the UPF HPC environment
One of the most important aspects of a machine learning related PhD is the large computing power needed for the experiments. Normally, in research centers and universities, this is solved by the availability of a cluster infrastructure with large computing capabilities, accessible by every researcher.