# Preparatory data analysis for the reconstruction of real-time MRI data

## TL;DR

This paper presents preprocessing methods, coil compression and channel selection, to improve real-time MRI reconstruction by reducing data volume and artifacts, tested on head and heart imaging data.

## Contribution

It introduces and evaluates two novel preprocessing techniques specifically designed for real-time MRI data to address speed and artifact issues.

## Key findings

- Coil compression effectively reduces data volume.
- Channel selection decreases streak artifacts.
- Methods improve image quality in real-time MRI.

## Abstract

Real-time magnetic resonance imaging (MRI) poses unique challenges related to the speed of data acquisition and to the degree of undersampling necessary to achieve this speed. This Master's thesis introduces and evaluates two pre-processing approaches for these problems: Coil compression to reduce the data volume and a channel selection algorithm to reduce streak artifacts which arise as a consequence of undersampling. Both approaches are tested on real data covering anatomical imaging of the head and of the heart.