Evaluation of Different Types of Filters in Magnetic Resonance Imaging Using Compressive Sensing with Pre-Filtering

Magnetic resonance imaging (MRI) machines allow one to acquire medical images based on static and variable magnetic fields, in such a way as to reveal the interior of human organs. To acquire images with good quality, MRI machines often demand a high number of measurements, which often require long acquisition times. Therefore, an important topic of research consists of developing acquisition methods that reduce the number of measurements and, consequently, the time required to acquire an MRI image. In this work, we use compressive sensing techniques and pre-filtering strategies to reduce the number of MRI measurements. We empirically tested a large set of filter banks to determine which filter settings provide the best image quality. When compared with state-of-the-art filters and to the non-uniform Fourier transform reconstruction, we have been able to increase the quality of the generated images while reducing the required number of radial lines and, therefore, of acquisition samples.