Carles Garcia Cabrera

I originate from Santa Maria de Palautordera, a quaint town in Catalonia, where my love for outdoor pursuits like trail running, rock climbing, and hiking takes precedence. Alongside my passion for sports, I find joy in crafting tools, ranging from simple scripts to intricate models, aimed at aiding people. My diverse professional journey includes roles as a waiter, lighting engineer, and research assistant, coupled with active involvement in local cultural societies and university teaching.

I pursued my Telecommunication Engineering studies at the Universitat Politecnica de Catalunya (UPC), specializing in signal processing and image processing. For my Bachelor’s thesis at the University of Auckland (UoA), I delved into cardiac image analysis under the guidance of Alistair Young and Pau Medrano. Returning from New Zealand, my focus shifted to Advanced Telecommunication Technologies for my Master’s studies at UPC, emphasizing advanced signal processing and data analysis. Once again, I ventured into cardiac image analysis for my Master’s thesis, supervised by Xavier Giró and Kevin McGuinness.

I obtained my PhD at Dublin City University, in the ML-Labs Centre for Research Training in Machine Learning, with a thesis titled “Generalizable Cardiac MRI Analysis with Deep Learning.” Under the supervision of Kevin McGuinness, Kathleen Curran, and Noel O’Connor, my research focused on advancing cardiac magnetic resonance imaging (CMR) analysis using deep learning techniques. This included generating synthetic CMR data for dataset balancing and developing predictive models, contributing to more robust and efficient cardiac MR analysis.

As part of my PhD studies at ML-Labs, I completed my internship at Medis Medical Imaging, under the supervision of Pieter Kitslar and Hans Reiber.

I am currently a Researcher at the Insight Research Centre for Data Analytics at the School of Medicine at University College Dublin. In the context of the SMASH-HCM Project, I contribute in the Data management, biomarker discovery and data-based modelling. Additionally, I supervise Master’s Students on their theses and contribute to a number of research proposals.

Selected publications

2024

Cardiac MRI Reconstruction from Undersampled K-Space Using Double-Stream IFFT and a Denoising GNA-UNET Pipeline

Julia Dietlmeier, Carles Garcia-Cabrera, Anam Hashmi, and 2 more authors

In Statistical Atlases and Computational Models of the Heart. Regular and CMRxRecon Challenge Papers, 2024

Abs

In this work, we approach the problem of cardiac Magnetic Resonance Imaging (MRI) image reconstruction from undersampled k-space. This is an inherently ill-posed problem leading to a variety of noise and aliasing artifacts if not appropriately addressed. We propose a two-step double-stream processing pipeline that first reconstructs a noisy sample from the undersampled k-space (frequency domain) using the inverse Fourier transform. Second, in the spatial domain we train a denoising GNA-UNET (enhanced by Group Normalization and Attention layers) on the noisy aliased and fully sampled image data using the Mean Square Error loss function. We achieve competitive results on the leaderboard and show that the algorithmic combination proposed is effective in high-quality MRI reconstruction from undersampled cardiac long-axis and short-axis complex k-space data.
The state-of-the-art in Cardiac MRI Reconstruction: Results of the CMRxRecon Challenge in MICCAI 2023

Jun Lyu, Chen Qin, Shuo Wang, and 8 more authors

arXiv e-prints, 2024
Synthetic Balancing of Cardiac MRI Datasets

Carles Garcia-Cabrera, Eric Arazo Sánchez, Enric Moreu, and 3 more authors

In Medical Image Understanding and Analysis, 2024

Abs

Possessing a balanced dataset to train the segmentation models used in clinical products is hugely important to enhance the performance in any given subset of patients. Due to demographics and uneven distribution of conditions of the subjects, this is not usually the case. In this work, we propose a novel method that synthetically balances a training dataset by applying deformations to an atlas. In particular, we focus on modelling two diseases that have distinct heart morphologies. Once the atlas is processed, we obtain different slice cuts from it and apply style transfer to make it appear as a real short-axis MRI scan. We then add those synthetic scans to our training set for the segmentation network. Our experiment compares the performance of the distinct trained models in a test set. Additionally, we also present the results on subsets of the test set representing the modelled pathologies. We found that using synthetic scans to balance the dataset resulted in up to a 0.05 increase in the DICE score. Our findings motivate further research on balancing cardiac MRI datasets by using atlas deformation.

2022

Cardiac Segmentation Using Transfer Learning Under Respiratory Motion Artifacts

Carles Garcia-Cabrera, Eric Arazo, Kathleen M. Curran, and 2 more authors

In Statistical Atlases and Computational Models of the Heart. Regular and CMRxMotion Challenge Papers, 2022